Tweet

By Broadcom

Private and public cloud applications, usage models, and scale requirements are significantly influencing network infrastructure design. Broadcom’s StrataXGS® architecture-based Ethernet switches support the SmartScale series of technologies to ensure that such network infrastructure design requirements can be implemented comprehensively, cost-effectively and at scale. This set of innovative and unique technologies, available in current and future StrataXGS Ethernet switch processors, serves as the cornerstone of Ethernet switch systems from leading equipment manufacturers worldwide.
 
This white paper explores the network infrastructure virtualization requirements in private and public cloud networks, and how such requirements affect the design of data center network switches. It also describes features that are enabled by Broadcom’s Smart-NV (Network Virtualization) technology, part of Broadcom’s SmartScale series of technologies, engineered specifically to meet current feature and scale requirements of private and public cloud networks. Smart-NV encompasses comprehensive best practices for today’s high-performance data center switches, and addresses evolving needs of next generation cloud implementations.

Get the White Paper

Tweet

On March 30th I met with Dave Husak, the Founder and CEO of Massachusetts Software-Defined Networking (SDN) start-up Plexxi, along with two other employees. For those who don’t know Dave, he’s intense and driven. Out of this two-hour meeting, Dave provided the inspiration to describe SDN as the third epoch of computer networking. In a nutshell, the first epoch was IBM mainframes and SNA, the second is client-server computing and LAN/WANs, with the third being mobile plus cloud computing and SDN. After attending the second Open Network Summit (ONS) last week to sold-out crowds, the main question I walk away with is this: is SDN the third epoch or a new set of features added to layer 2/3 networking? In other words is SDN a new disruptive market or a high-end networking technology like InfiniBand? In this Lippis Report Research Note, I share the top ten observations at ONS and answer the above question.

First Epoch: The first epoch of computer networking started in the mid 1950s with mainframe computers and the first version of sneaker net, that being batch followed by Binary Synchronous Communication (BSC) and SNA over multipoint lines. The public switched network was based upon circuit switching and the national entertainment network as analog broadcast with all but three or four channels. The Internet was non-existent. IBM states that in the past 30 years, businesses have invested some $20 trillion in labor and money in developing CICS and IMS over SNA applications. But by the mid 1980s, IT business leaders were growing increasingly frustrated with this computing model’s high cost plus long application development times, which were often over budget and feature deficient. In short, SNA was viewed as not flexible or too rigid to support a new computing model. Alas the first epoch’s reign of 35 years was coming to an end.

Second Epoch: During the mid 80s, personal computing was heating up, marked by Apple’s 1984 Super Bowl commercial. At the same time, Ethernet hubs were available and growing as the infrastructure for client-server computing. TCP/IP was standardized in this time frame, which defined the second epoch of computer networking as Routing and Switching aggregated traffic and forwarded layer 2 and 3 (L2/L3) packets. This was the golden internet age of networking. This model was/is so powerful that it’s the core of the public switched telephone network, mobile network and national entertainment network. You can’t say that the second epoch is over but you can hear IT business leaders complain loudly that the network is in the way of their needs and business desires to spin up and down applications, move workloads, etc. The same complaints that were voiced in the 1980s are starting to be said now; network operational cost is too high and that networks are not flexible or too rigid to support a new computing model. So is the second epoch’s reign of 35 years coming to an end?

Third Epoch: Computing drives network architecture and over the past few years, computing has fundamentally changed to mobile and cloud with virtualization providing compute density and efficiency. Networking’s Third Epoch is the era in which we are living today, the era of network programmability, and in particular SDN, which enables the democratization of network application programs and features. With SDN, the notion of layering in the forwarding plane in network switches disappears. Indeed the switches become protocol-ignorant, while layering remains meaningful in hosts and at transition points between networks with different control mechanisms. On the computing side, this is the era of mobility, virtualization and the cloud, with applications finally freed from having to be aware of specific details of network plumbing, like IP addresses and ports.

But established vendors such as Cisco, Juniper, HPN, IBM, Brocade, Dell, Arista, Avaya, Alcatel-Lucent, Extreme and others are dialed into the new network requirements. While most of the above firms have either announced their support of OpenFlow and SDN, many will offer programmable Layer 2/3 networks to address 80% of IT business leader requirements, slowing down SDN deployments. Also remember architectural shifts take a long time to materialize.

So is SDN the third epoch or a new set of features added to layer 2/3 switching/routing networks? My top ten observations from attending this year’s ONS may help answer the question.

1) It’s All about OpEx: Early SDN messages were about cheap switching products from Asia being controlled by sophisticated controllers from the likes of Big Switch Networks, Nicira, NEC, et al. But SDN is all about reducing the cost to operate networks. Consider this: it takes one engineer to manage 8,000 nodes in a mobile network but it takes one engineer to manage 75 nodes (switches or routers) in an enterprise network. The industry has prided its self on a value prop based upon capital cost being 25%, operational cost at 60% and facilities being 15%. Centralizing network control is all about reducing operational cost, and if capital cost gets cheaper too, then so be it, but it’s not the driver.

2) First Virtualize then Customize: It’s becoming apparent that SDN pilots and early implementations are all about virtualizing the network. Yes, we have had network-virtualizing technologies for years with VLANs, MPLS, VRF-Lite and Cisco’s latest Easy Virtual Network, but SDN does it without the huge operational cost of configuring each switch and/or router, and there is complete separation of the address space. There are many requirements for virtualizing the network, including offering unique attributes to common users, segmenting departments and businesses, delivering multi-tenant services, and even offering each employee his/her own virtual network where credentials are checked and IT services allowed; this is huge for BYOD and mobile employees. Once a network is virtualized, these logical networks can be customized, thanks to layer 4-7 services being moved into the virtualization domain. This flexibility is huge in the enterprise campus, data center and service provider markets.

3) Limited Number of Real SDN Implementations: There are only approximately 30 to 60 SDN projects taking place around the world, a small number, so expect limitations and setbacks in SDN promises being made.

4) New WAN: One of the highlights of ONS was the keynote from Urs Hölzle, SVP Technical Infrastructure and Google Fellow. Urs demonstrated SDN in the WAN with a custom built 10GbE switch via merchant silicon equipped with 100s of ports of non-blocking 10GbE, OpenFlow support, Open source BGP, ISIS and scale to Tbps. What was striking about this is the reported improved re-route time, convergence time, increased performance and reliability plus greater control and flexibility afforded by SDN in the WAN than offered by traditional hop-by-hop routing.

5) Can SDN Move Down Market? Early SDN adopters and promoters are the largest of data center owners such as Google, Yahoo, etc., that were represented at ONS. The question is can SDN move down market? Only a few firms can afford to build their own 10GbE switch and experiment over the wide area as Urs did at Google. Yahoo!’s principal architect, Igor Gashinsky, was hopeful to be able to access the Linux kernel of switching and routing devices for greater programmability, but there aren’t too many IT organizations that would enjoy that opportunity. But many may find what Igor builds interesting and potentially useful if it was made available to others.

6) Wither OpenFlow? While OpenFlow has enjoyed much industry discussion, privately, many firms, both large and small, expressed that OpenFlow is all but one approach or mechanisms available to program switches.

7) VMware Has a Big Hand to Play: It’s becoming clear that VMware has a huge hand to play in the emerging SDN market. vSwitch, a VMware innovation, started the SDN and network virtualization journey. Its vDS enabled pooling of network ports across clusters via aggregation of vSwitches. To extend or overlay layer 2 virtual networks over layer 3 boundaries, it jointly developed VXLAN with Arista, Cisco, RedHat, Citrix, Intel, et al, and is now a draft IETF RFC. VXLAN extends large layer 2 VM domains well beyond the 4K VLAN limit to 16 million. It is touted as a key standard that avoids proprietary overlay networks plus allowing VM domains to span virtual and physical networks. In addition, its vCloud Director enables alignment of elastic compute and networking diameters. Via vShield, VMware has added virtual firewalls, load balancing, VPN, IPAM, hybrid cloud extensions, and the ability to logically insert partner services, like IDS/IPS and WOC or WAN Optimization controllers. With this growing SDN stack, VMware is in an excellent position to offer APIs to application developers exposing virtual network topologies and other stateful information. 

8) Network Influence/Control Shifts to Virtualization Domain: There are two tectonic shifts that potentially threaten networking as we know it and networking professionals. The first one is the migration to technologies that recognize the relevance in the hypervisor switch of L2/L3 designs. The second shift is L4-L7 services that are cost optimized and compatible with VMs. Let me explain.

With OpenFlow, Open vSwitch and Quantum providing the base network virtualization tools for KVM, Virtual Box and Xen while VMware providing its own tools mentioned above, a shift in network control or balance of power is occurring into the virtualization domain. When networks are virtualized, they can stay in the virtualized domain, traverse physical layer 2/3 networks or some combination of both.

At the heart of virtual networks is how they traverse physical switches and routers. There are layer 2 tunnels and layer 2 over layer 3 tunnels such as VXLAN and NVGRE. There is OpenFlow in “native” mode, which is OpenFlow supported by all of the switches and routers in a network, and OpenFlow in “overlay” mode where only the hypervisor switches are OpenFlow enabled, and the OpenFlow network is overlaid on the physical L2/L3 design using tunnels. Nicira does overlay mode using its custom STT tunneling technology. Big Switch uses plain vanilla OpenFlow so that it supports overlay and native as well as hybrids of the two.

Both firms envision networking being provisioning and controlled from the virtualization domain where virtual networks are created, managed and layer 4-7 services administered. If and when this model comes to fruition then physical networking becomes less strategic as network services move to the virtualization domain. Firms with large virtualized data centers like this model as they have stranded CPU resources, thanks to memory limitations in virtualized servers. Layer 4-7 network servers are CPU intensive but use little memory, a perfect fit to move these services in to the virtualized domain; a sunk cost that is already being managed. In essence, they view it as getting Layer 4-7 services for free.

At ONS, there were many firms offering virtualized layer 4-7 services such as: vArmour with its distributed firewall, Embrane providing layer 4-7 virtualized services, Radware with its load balancer and firewall, LineRate with its ADC, and don’t forget Cisco, F5, Brocade and many others that have virtualized their Layer 4 -7 appliances.

9) SDN Definition…or What You Thought It Was…Will Be Totally Different Next Year: The definition of SDN as an OpenFlow interface on virtual and physical switches that are controlled by a centralized controller will be totally different next year. As Cisco, IBM, Dell, Juniper, HP, Arista, ALU and others wrap their minds and product lines around SDN, it will take a decisively different shape.

Cisco’s CTO and Chief Architect of the Service Provider Division, David Ward provided a glimpse of things to come as Cisco provides programmatic interfaces in a hope to provide new services and functionality by augmenting existing network control, management and forwarding state. Cisco and all networking firms will offer ways to program L2/3 networks and expose network intelligence to applications. Juniper, for example, offered an SDN over Qfabric option to consider. IBM offered a way to ease control of a converged data center fabric of storage and networking with SDN. Dell offered a multi-tenancy data center solution, thanks to an OpenFlow-based SDN.

10) Pervasive SDN: ONS presentations demonstrated pervasive SDN use cases that spanned the service provider market, cloud computing facilities, enterprise campus networking, wide area networking, data centers and mobile infrastructure. Each use case was driven by the centralization of control to both reduce operational cost and increase functionality. In short, SDN promises to reap more from networking at lower operational cost.

So is SDN the third epoch or a new set of features added to L 2/3 switching/routing networks? From what I learned at ONS, you can make the case for both, but pay close attention to observation number nine, that is “SDN Definition…or What You Thought It Was…Will Be Totally Different Next Year” is key. The networking industry is in hyper-innovation mode, embracing SDN concepts such as exposing network intelligence to applications, exploring programmable networking and SDN-like architectures where more control is centralized to reduce operations. I get the feeling that SDN is the third epoch, but as it’s cranked through the meat grinder of the industry, it will take on a form and shape that is indistinguishable from its current form. What do you think?

Tweet

In one of the most candid discussions with Jayshree Ullal, Arista Network’s CEO, she voices her view of Software-Defined Networking as not a market but a set of features. Jayshree views OpenFlow as but one in many APIs becoming available to program layer 2/3 networks. If you’re looking for SDN hype, then no need to listen to this podcast. But if you want a realistic view of this important industry trend, then listen up.

Listen to the Podcast

Tweet

There are still many manual steps, such as phone calls and emails between IT teams, throughout the life cycle of a virtual machine impeding timely and efficient VM deployment and, ultimately, the expanded deployment of virtual environments and associated benefits. To deliver on the promise of cloud computing’s on demand application availability, tasks need to be automated. Infoblox has launched its Automation Tasks Board tools in the model of a Software-Defined Network to enable IT department members to initiate with the click of a single button multi-step, often-repeated and time-consuming network tasks while providing cross team visibility and auditability. If you want to build a real cloud computing facility, then you need to listen to this podcast.

Listen to the Podcast

Tweet

There are multiple definitions of Software-Defined Networking or SDN. But this is common in a new breakout space for the computer networking industry that’s evolving fast. The most common SDN definition is based upon splitting the data plane or the forwarding hardware of an Ethernet switch from its control plane or the logic that controls how packets flow from ingress to egress. But this definition alone is too limited and needs to be expanded. In this Lippis Report Research Note, we offer the industry a broader SDN definition and view.

First, the SDN definition that is based upon OpenFlow is important but too narrow. OpenFlow offers a standard-based Application Programming Interface or API that links an Ethernet switch and a controller. This offers a model in which layer 2 Ethernet switches are low-cost merchant silicon based devices where flows are directed by a centralized controller(s). While this is innovative and different, in reality it’s not that interesting. There needs to be much more to SDN and that can be found in what resides on top and along side of SDN controller(s) and associated benefits, both in terms of network design and operational models that it affords.

From an architecture point of view, what resides on top and along side of the controller(s) is another API or set of APIs that promise to virtualize networking like VMware did for servers. With a yet-to-be-defined API on top of the controller, a software ecosystem needs to flourish.

SDN Software Ecosystem

Applications such as traffic management, device configuration, network analytics and control, public-private cloud connectivity and security, firewalls, load balancing, etc., are examples of applications that could and should spring up in the virtualization domain, thanks to SDN. Much work is being done now to automate the network layer and virtualization stack into the virtualization domain via SDN applications that may or may not ride on top of an SDN controller(s). The centralization of network provisioning of layer 2 and 3 devices, firewalls, load balancers, VM stacks, etc., will be a huge SDN advantage as it lowers the number of operations staff required to manage a large network. Look toward management of physical switches in the management domain of virtualization engines.

SDN Enabled Cloud Bursting

Enabling burst capability where a corporation can move workload between public and private clouds will be an SDN function. While there is layer 2 functionality available in some controllers today, to enable cloud bursting, this will move to layer 3 over time. But most importantly, SDN controllers are solving the security problem of workload mobility between public and private clouds today, which offers a huge network design and business agility advantage over existing approaches.

SDN Virtualized Network Services

While many firms, such as F5, Brocade, Cisco, Citrix, et al, offer virtualized network appliances, delivering such services within an SDN will offer huge server efficiency. For example, in highly virtualized data centers, memory restriction strands CPU capacity. Network appliances, such as firewalls and load balancers, typically consume little memory but much CPU processing capacity. Commodity servers inside of racks tend to be only 40% CPU utilized, thanks to lack of memory to run more applications upon those servers. These servers are, in essence, stranded, but a low memory, high CPU network application, such load balancing or firewalling, can utilize that un-used resource, increasing data center efficiency. SDN offers this efficiency and it’s a huge win. In an SDN environment, there will be a controller somewhere in the network, and if this runs in the virtualplex as an application then all of this server efficiency just comes to the IT architect, in essence, for free.

Open SDN

The SDN market is evolving in an inclusive open fashion. The OpenFlow interface is open by definition. In addition, components of SDN controllers are being distributed to the open source community, such as Big Switch Network’s FloodLight. Also, FlowScale, a load balancer, RouteFlow which provides virtualized IP routing services over OpenFlow hardware, Open vSwitch and other projects including layer 2 provisioning, VM Migration, etc., are creating an open SDN environment.

Mobile Market Shows the Way

The mobile market may show the way of how SDN will progress. The national mobile infrastructure is well automated to the point where a single network engineer can mange some 8,000 nodes. Most, if not all, large enterprises and cloud providers would welcome such efficiency. In addition, the mobile market, thanks to Apple’s iPhone and iPad plus Google’s android, has shown how a vibrant software ecosystem can add tremendous value and user choice. An SDN software ecosystem would offer IT business leaders with applications that change the nuts and bolts of networking suited to highly-virtualized environments plus solve some of the industries largest problems and opportunities, especially around cloud bursting and workload mobility. If SDN is able to automate network provisioning in enterprise and cloud computing facilities much like mobile networks today would fundamentally change the network operational model.

A Broader SDN View

The definition of SDN needs to be sufficiently broad enough to communicate the above value. To achieve that, SDN will move well beyond an OpenFlow-based definition to an application and capability definition. SDN promises to commoditize network hardware and provide a standard-based application development platform taking much of the features and functionality that exist inside custom proprietary software and driving it into an open SDN space.

But perhaps even more important is how SDN is implemented. In short, SDN promises to be deployed on under-utilized servers that IT organizations already own and operate. SDN promises to completely revolutionize the way in which we do networking. Trends in virtualization and cloud sourcing are only going to get stronger over time. Stranded CPU capacity in virtual engines is a significant previously unavailable resource to tap into and utilize. Running SDN controllers and applications in that domain is, in essence, free to IT organizations.

Think of it this way: IT business leaders will be taking this huge expensive IT infrastructure they currently own and operate to run SDN software and controllers in capacity that they weren’t capable of using anyway. That is a huge win. Add commoditized network hardware to the equation plus network application/service innovation to the mix, and you have a network environment for the new age of cloud computing. This is the promise of SDN and why it’s so important to every corporation, cloud provider and networking vendor.

Tweet

Virtualizing a physical network into multiple logical networks each with unique attributes has grown in popularity. This network design is popular in healthcare, education, travel and other industries. Network virtualization was available only to the largest of enterprises and service providers, thanks to its cost and complexity of MPLS and VRF-Lite. But a new approach called Easy Virtual Network from Cisco changes all of that by reducing cost and eliminating configuration and management complexity opening network virtualization to a much larger segment of the enterprise market. In this Lippis Report podcast, I talk with Sehjung Hah about Cisco’s Easy Virtual Network.

Listen to the Podcast

Tweet

The term “fabric” has nearly as many definitions and permutations as “cloud,” so in this Lippis Report podcast Vikram Mehta, Vice President IBM System Networking joins me to discuss the top 10 attributes that a data center network fabric should possess. With the back half of 2012 kicking-off aggressive data center fabric deployments listening to this podcast is a must to help you with your planning.

Listen to the Podcast

Tweet

Computer networking vendors have been increasing the speed and port density of their Ethernet switches while reducing power draw and price per port. But while Ethernet switching hardware marches on linearly, thanks to 10, 40 and 100GbE, networking software is taking a different historical path as the pace of compute and network technology evolution has diverged, with networking lagging. Highly virtualized server deployment has broken traditional networking approaches on multiple levels, for example. In response, the industry is now developing a “virtualized infrastructure” or “stack” to add network flexibility. To close the technology gap, Software-Defined Networking (SDN) is promoted as the new “organizing principle” to deliver network software and service value. While it will be, likely, years before SDN’s organizing principles take hold, I propose that these two industry activities are inexplicably linked and phased; here’s why…

Software-Defined Networking

There are multiple definitions of SDN. Making it even harder to pin down SDN, the definitions are evolving too. But this is common in a new breakout space for the computer networking industry that’s evolving fast. For this Lippis Report Research Note, we take the SDN definition that is based upon splitting the data plane or the forwarding hardware of an Ethernet switch from its control plane or the logic that controls how packets flow from ingress to egress. This split of data and control planes opens up an innovation injection point into networking that has not been previously available.

During 2011, a market has opened up for controllers. Currently Big Switch Networks, Nicira Networks and NEC are offering standalone centralized controllers. But limited controllers are also available in open source software, OpenStack and VMware’s vSphere/vCloud too. In addition Cisco’s IOS, Juniper’s Junos, Arista’s EOS, etc., are distributed controllers that may interoperate with centralized controllers in the future. In fact, Arista’s EOS already supports OpenFlow, OpenStack and vSphere/vCloud.

The link between the separated data and control plane is an open interface called OpenFlow. Now some end their SDN definition here, but this is just the beginning as the real promise of SDN are the applications that will reside upon the controller to address a wide range of networking issues and opportunities. In fact researchers at Princeton and Cornell are developing the Frenetic programming language that provides high-level network abstraction that gives programmers direct control over the network, allowing them to specify what they want the network to do without worrying about how to implement it.

One can imagine a wide range of applications residing upon a controller such as WAN optimization, traffic engineering optimization, load balancing, security services, etc. In essence, the control plan allows network services that are currently deployed as appliances to be virtualized appliances/applications much like applications that reside on top of a VM. It gets even more interesting, as a centralized control plane can be easily split in to many little control planes, each of which sees its own slice of the data plane topology. In traditional networking where control and data planes are one and the same and in each box, it is much harder to merge control planes and split data planes. It’s possible, but harder to keep complexity and stability in check over the long term. Splitting control plans can have huge value in public cloud multi-tenant or private cloud multi-team networking.

SDN and OpenFlow are at the early stages of its industry matriculation. But one thing is clear: SDN is an organizing principle whereas network software is developed by both network vendors and third parties, and network services are virtualized. SDN thus represents a new industry order and structure as to how value is added to networks. But I digress. The real issue today is solving network inflexibility in the face of highly virtualized data centers.

Enter the “Virtualized Stack” or Virtualized Infrastructure”

Virtualized server deployment has been propelled en masse, thanks to increased data center efficiency, by delivering the same or greater application workload with a reduced number of servers. While this is good, many IT business leaders are now realizing huge consequences to highly virtualized data centers that span from IP address change management to application management.

At the IP address level, networking has become extremely rigid within virtualized environments, slowing down process, limiting moves and changes as well as elongating the time to spin up an application that resides within a VM. Necessary network services to support the virtual cloud infrastructure, such as IP address assignment and management, are still performed largely with manual tools and processes, such as spreadsheets shuffled between various departments or operational groups, which can result in days of delay for something as simple as assigning an IP address to a VM. Contrast that with the virtual server administrator. Virtual instances of servers and machines can be dynamically provisioned, migrated and shut down by a virtual server administrator in minutes.

Moving up the stack, challenges are rooted in application management plus Layer 4-7 services such as WAN optimization, Application Delivery Controllers and security, especially in environments that include multiple hypervisors, a wide variety of workload types and shifting virtual machines.

For example, the new challenges of enterprise application management in virtualized data centers include: what type of and location of network intelligence is required when multiple hypervisors and various workloads exist and shift? Also how do operations groups maintain consistent security policy across both virtualized and non-virtualized environments consistently? And how do operations groups monitor and maintain application flow visibility?

Cisco

Cisco, for example, is addressing these issues via its Virtualization Stack and is now organizing its products around this initiative. Three components define Cisco’s virtualization stack, those being: 1) virtual networking, 2) virtual security and application networking services and 3) orchestration and provisioning. An important part of Cisco’s strategy is the virtualization of appliances such as its VSG or Virtual Security Gateway, the ASA 1000v, the support of VXLAN, the Nexus 1000v, etc.

Brocade, F5, Citrix

But F5, Citrix and Brocade are all virtualizing their appliances, moving away from physical single application appliances to an integrated virtualized suite. One can imagine that these virtualized applications will some time reside upon an SDN controller as their next stage of evolution. In addition each application delivery vendor has a way for programmers to control application network behavior. For example, Brocade recently launched OpenScript, a Perl-based scripting language used to modify the content of and control delivery of packets at Layer 4 through Layer 7 on its ServerIron ADX products. These scripting languages could be standardized and reside within an SDN controller.

Embrane

A good example of what the virtualized Layer 4-7 future may hold is that of a start-up firm called Embrane.

Embrane has virtualized server load balancing, firewalls and VPN termination and placed them upon a distributed software platform called heleos. Heleos runs on x86 servers and any hypervisor. It leverages a distributed virtual architecture that decouples network services functionality from the underlying physical infrastructure and hypervisor technology that it says provides high scalability, flexibility and performance.

IBM & NEC

IBM and NEC offer the best example of a commercial SDN offering with OpenFlow. NEC’s pFlow OpenFlow controller that resides within an IBM server manipulates IBM System Networking G8264 OpenFlow switch’s flow table. The link between the two is OpenFlow 1.0.0. The NEC pFlow controls traffic, discovers topology, gathers stats and other functions while the G8264 forwards traffic based upon these flow commands.

What’s impressive about the IBM/NEC SDN solution is that it has customers such as: Tervela validated the IBM and NEC OpenFlow solution ensures predictable performance of Big Data for complex and demanding business environments. Selerity’s IBM and NEC’s OpenFlow solution improved real-time
decision-making for global financial markets. Stanford’s IT Department chose IBM and NEC’s OpenFlow solution to deliver network capacity on-demand to its academic community. What’s important about these use cases is that IBM is communicating SDN via OpenFlow’s value in business terms, which will only increase as industry adoption accelerates.

In essence the SDN market has started, and as its technology underpinnings solidify, many of today’s network services will fall under the SDN umbrella. In fact, nearly all network vendors are launching SDN programs as a new way to communicate existing product value and their evolution into a SDN. Just like the Appian Way where all roads lead to Rome, all network services may very well lead to an SDN.

Tweet

WAN bandwidth, or the lack thereof, is the single largest contributor to poor application performance in branch offices resulting in reduced productivity and the inability of IT to implement their entire strategy. This is especially acute as applications are increasingly centralized in data centers and video traffic starts to dominate flows all of which put more pressure on the WAN. A new approach to wide area networking takes advantage of virtualization concepts. WAN virtualization promises to reduce WAN cost and increase performance. If your WAN utilizes T1 MPLS links at $300/month or higher and ADSL between $10-$15/ month, then you can benefit significantly from WAN virtualization as it offers 3 to 4 times the bandwidth at lower cost and in the process eliminates WAN design constraints limiting IT business leaders. Keith Morris of Talari Networks joins me to discuss WAN Virtualization. I have first-hand knowledge of how good this solution works; if you fit the above profile, then listen up.

Listen to the Podcast

Tweet

By Jim Metzler, Ashton Metzler & Associates

This white paper proposes a new approach to architecting and designing data center networks for current dynamic and highly virtualized data centers.

Get the White Paper

Tweet

The Inflobox plug-in module for VMware’s vCenter Orchestrator promises to bridge network and virtual IT team silos by enabling automated assignment of IP addresses to VMs. The Infloblox plug-in enables IP addresses to be automatically assigned to virtual machines in less than a minute and then constantly monitored and managed, which simplifies troubleshooting, accelerates time to value and offers greater flexibility for the virtual team. Steve Garrison, VP Marketing for Infoblox and I discuss the problems of networking in virtualized environments and how Infoblox’s vCenter plug-in solves them.

Listen to the Podcast

Tweet

By Zeus Kerravala, Senior Vice President and Distinguished Research Fellow

The enterprise data center has undergone several major transformations since the introduction of computing as a corporate resource. The computing platform has evolved from main frame computing, to client server, to Internet-based computing, and now we sit on the precipice of the next major data center transition—the evolution to a fully virtualized data center. Each transition saw the cost of computing driven down and the importance of the network elevated. Each phase allowed organizations to increase the efficiency of their data center operations and improve asset utilization, ultimately leading to a better experience for end-users. This white paper provides a perspective on the rise of network value as it acts as the backplane for the virtualized data center and provides design recommendations.

Get the White Paper

Tweet

By Avaya

This White Paper discusses the benefits and applicability of the IEEE 802.1aq Shortest Path Bridging (SPB) protocol, which is augmented with sophisticated Layer 3 routing capabilities. The use of SPB and the value to solve virtualization of today’s network connectivity in the enterprise campus as well as the data center are covered.

This document is intended for technically savvy network managers as well as network architects who are faced with:
• Reducing time to service requirements
• Less tolerance for network down time
• Network Virtualization requirements for Layer 2 (VLAN-extensions) and Layer 3 (VRF-extensions)
• Server Virtualization needs in data center deployments requiring a large set of Layer 2 connections (VLANs)
• Traffic separation requirements in campus deployments for security purposes as well as robustness considerations (i.e., contractors for maintenance reasons needing access to their equipment or guest access needs)
• Multi-tenant applications such as airports, governments or any other network with multiple discrete (legal) entities that require traffic separation

Get the White Paper

Tweet

Just as 10 Gigabit Ethernet (GbE) is going through widespread
deployment in the data center, the discussion has now shifted to even
higher speed interconnects—namely 40 GbE and 100 GbE
By IBM

In July 2006, the Institute of Electrical and Electronics Engineers (IEEE) Higher Speed Study Group was formed to look into the next evolutionary step after 10 Gigabit Ethernet (GbE). In the past, Ethernet speeds would increase by a factor of 10. However, the next generation jump from 10 GbE to 100 GbE has proven to be a technological challenge. Some within the IEEE group felt that 100 GbE made sense for communication service providers and other backbone network providers, but not as a next step for servers—it was simply more speed and expense than would be needed for the near future. While the IEEE initially planned to standardize only on 100 GbE as the next step after 10 GbE, server vendors initiated a push in early 2007 to include 40 GbE in the standard, with the rationale that the effort used to develop 40 GbE would be used for the development of 100 GbE.

In July 2007, the IEEE 802.3ba study group was named, and it is the first standard to include two different Ethernet speeds—the 40 Gbps rate for local server applications, and the 100 Gbps rate for internet backbone—to serve both market needs. In June 2010, the official 802.3ba standard was ratified, opening the field to higher performance in server systems and components, data centers, network storage and systems, high-performance computing (HPC) clusters, data centers, carriers, and the like. This paper provides perspective on the placement and use of 40 and 100GbE.

Get the White Paper

Tweet

Simplified and optimized service orchestration maximizes the return from a virtualized computing environment
By Avaya

This white paper discusses Avaya’s approach to Data center Networking from a fabric perspective. Virtualization within the data center is now taken for granted, with some declaring that ‘Cloud Computing’ will be the choice of most enterprises and that applications and information will become commodities. Experience has proved one thing; the data center of the future cannot be built on the technology of the past. General-purpose products, outmoded techniques, and legacy designs cannot be re-packaged as ‘data center-ready’. Ethernet is readily available, cost-effective, extensible, and – as the 40/100 Gigabit developments prove – scalable. Find out Avaya’s approach to data center networking fabric by downloading this white paper.

Get the White Paper

Tweet

By all counts, Cisco’s upgrade of the Catalyst 6K via its new Supervisor 2T, or Sup2T, is its most ambitious and thoughtful yet for the venerable platform. The Sup2T is a 2 Terabit (Tb) platform that triples the previous Sup720 performance. Thanks to the support of Virtual Switching System (VSS), the platform allows two 2 Tbps switches to combine into a single 4 Tbps virtual switch. The Sup2T is a major upgrade to the most widely-deployed switching platform in campus and data center networking in the industry. But while these performance numbers are impressive, it’s the new Cat6K’s network services and pricing that deliver most of the value. From a services’ point of view, the Cat6K stands alone.

Cisco’s Cat6K is the firm’s most successful product with over 700,000 systems and 110 million ports installed, worth some $42 billion. This product’s success increases the stakes for Cisco as it introduces a major upgrade. Cisco had to consider backward and forward customer migration, increased competition and pricing pressure especially as many firms are starting to offer core switches based upon merchant silicon. In short, Cisco had to eliminate the trade-off of innovation versus investment protection and find a way to deliver both simultaneously. A detailed review of the new Cat6K with Sup2T finds that Cisco has navigated well by incorporating customer feedback from multiple theaters and industry segments in the form of some 200 features, most of which are incorporated into ASICs, something with which merchant silicon based switching firms cannot compete.

Merchant Silicon versus Custom ASIC

There will be an increase in the number of core switches offered from various vendors during 2012 thanks to the availability of merchant silicon, but these products, for the most part, will be focused on primarily performance while falling short on network services. Network services are hardware and software features that provide the tools, customization and design options for IT architects to optimize their networks and applications to either run faster and maintain secure, reliable, high-quality user experiences whether it’s for video traffic, virtualized desktops, general purpose office productivity or client facing web traffic.

For example, consider something as mundane as counters. In the Cat6K Sup2T and new modules, there are more than two million counters, enough to have separate counters for every protocol, including IPv4, IPv6, multicast, unicast, MPLS, etc. What this says is that Network Operations engineers will be afforded a level of granularity and visibility into the network well beyond anything they previously could gather. But I digress; let’s focus on the big picture of the new Cat6K.

The New Cat6K by the Numbers

The last major upgrade for the Cat 6K was the Sup720-10G in 2007, which was the first management module with 10GbE uplinks. The Sup2T enables 40GbE interoperability and interface speed transition as the Cat6K will support 100MbE, 1GbE, 10GbE and now 40GbE in a modular chassis platform. The performance leap on the 2 Tb portfolio is complemented by a quadrupling, or more, of the NetFlow, Access Control List and Quality of Service capacities of the platform to meet the increasing manageability, security and service demands of enterprise networks. The platform now offers 720 Mpps of IPv4 and 360 Mpps of IPv6 performance, roughly a twofold increase over the previous generation. In a word, the Cat6K scales logically.

What Cisco engineering has done is tripled the performance, quadrupled the platform scalability and added new network services—several of which are industry firsts and all of which protect investment by being backward compatible with these forward innovations. For example, central forwarding line cards that started shipping in 2003 are supported in the Sup2T. The E-series chassis and power supplies that started shipping in 2004 are supported with the Sup2T. For a large segment of the Cat6K installed base, all that is required is the install of the new Sup2T to gain increased performance, scale and network services. This is perhaps one of the easiest refresh offers Cisco has ever made.

Network Services Rich

As for network services, the Cat6K supports some 2,600 features that the market has demanded. Most of these features were developed over time with many firms depending upon them to run their networks. In addition to hardware backward compatibility, Cisco had to be software backward compatible too by supporting these 2,600 features, which are supported in the Sup720 and the wiring closet Sup32, in the Sup2T. Some of these features include IPv6, multicast, NetFlow, MPLS, etc. But clearly the market does not stand still, and Cisco engineering has added some 200 new innovations to the Sup2T, some of which will also be supported on previous versions of supervisor engines.

Interestingly enough is that with backward support of new network services supported on the Sup720, IT architects can choose to move these Cat6Ks down a network layer and place the Sup2T Cat6Ks in the distribution and core, extending the entire portfolio of network services from access, distribution and core. Some of these new innovations are Flexible NetFlow, Role-based Access Control, Virtual Private LAN Service (VPLS), Bridged Domain Technology, etc. Following are a few of the next generation innovations introduced with the Sup2T.

NetFlow: NetFlow scalability in the Cat6K Sup2T has increased fourfold with larger tables being supported in the ASICs. Up to 13 million NetFlow entries are possible in a single system. That is up to eight times the visibility afforded by the previous generation of NetFlow hardware. Over time, most networks will have a mix of 1GbE, 10GbE and 40GbE; this new version of NetFlow introduced sample NetFlow so NetOps does not have to export all traffic to collector, a huge complexity and time reduction. Also NetFlow visibility is now protocol independent, meaning that it does not matter if a network is running IPv4, IPv6, MPLS, Unicast, Multicast, etc. In addition, select modules, rather than the central supervisor, are able to export NetFlow to the NetFlow collector offering yet another way to scale.

MACsec: From a security perspective, the Cat6K Sup2T natively supports MACsec, or IEEE 802.1AE, embedding it within line cards offering line-rate, hop-by-hop encryption and decryption. In addition to the new Cat6K, the Nexus 7K, Cat 3K and Cat 4K currently support MACsec, thereby enabling end-to-end secure communications much like IPSec and SSL but over the LAN.

Role-Based Access Control List (RBACL): Access Control Lists, or ACLs, can now be programmed in role-based scenarios controlling user access to IT resources. Roles can be finance, human resources, marketing, engineering, sales, executive management, etc. Role-based access control allows NetOps to configure which IT resources each user is allowed to access for each type of job role, thereby controlling their access to servers, applications, WAN connections, etc. Role-based access control is an addition to the Sup2T’s ACL Dry Run, which first tests if ACL changes will fit in the ACL Ternary Content-Addressable Memory or TCAM before they go live with the configuration. Using ACL Dry Run will help avoid potential network disruption since NetOps engineers will know whether the ACL changes will be supported in hardware before implementing them. If an ACL change does not pass the Dry Run, then the system will indicate which resources are being exhausted, allowing the NetOps staff to adjust the ACL accordingly.

Network Virtualization: The new Cat6K Sup2T boosts its network virtualization capabilities that enables physical infrastructure to be logically divided. For example, airports, such as Zurich, Munich, Toronto, etc., use network virtualization to change gate attributes as an airline carrier completes the boarding process and transitions the gate to another carrier. They also use network virtualization to separate out kiosk vendors from operations from WLAN AP guest access to airline carrier support, etc. Governments network virtualization to logically segment departments while they share the same physical building/floors/office spaces. Universities use network virtualization to logically segment administration, research, faculty and student interests. Just as with other previously-mentioned capabilities, Sup2T increases the scalability for network virtualization up to fourfold with support for up to 4K MPLS VPNs, 32 instances of (VPN Routing and Forwarding) VRF-lite, native VPLS in hardware, allowing for VPLS-facing interfaces to be any interface in the system, and more.

New Service Modules

Admittedly, the Cat6K with the Sup2T is not the fastest Ethernet switch on the market with 2 Tbps of switching capacity. Cat6K doesn’t need to be the fastest given its place in campus networking and mid-range data centers. However, it does need more than enough performance to never be the bottleneck in IT delivery while providing a wide range of software options to control traffic and optimally design enterprise IP networks. Cisco engineering has done this with 2 Tbps, and 4Tbps with VSS, far greater capacity of most, if not all, campus and mid-range data center networks operating at a range of 10/100/100, 10GbE and soon 40GbE. For higher performance, Cisco offers the Nexus 7K with 9 Tbps of switching capacity for data center switching designs.

To increase performance in the Cat6K, it’s not just the supervisor engine that’s been upgraded. New service modules, such as the new Wireless Service Module 2 (WiSM-2), Adaptive Security Appliance Service Module (ASA-SM) firewall, Network Analysis Module 3 (NAM-3) and Application Control Engine 30 (ACE30) load balancing were introduced to take the Cat6K with Sup2T to the next level of hardware-based services processing. Remember, service modules allow IT business leaders to reduce the number of devices in their network they need to manage, improving energy efficiency and reducing carbon footprint. These new service modules have been upgraded for performance and scalability, as services performance has to scale with network performance. For example, the ASA-SM offers a threefold increase in performance with 15-20 Gbps of stateful application firewalling. NAM-3 has been upgraded in performance by a factor of fifteen, allowing application visibility and analysis at 15 Gbps. The WiSM-2 scales up to 20 Gbps of throughput and support for up to1,000 centrally-managed access points, a threefold increase in performance and scalability.

Integrated and Virtualized Network Services

Unique to a Cisco environment is that service modules and appliances basically share the same operating system, meaning that there is operational consistency between the two platforms. For example, if an IT architect implements an ASA appliance and ASA-SM, NetOps will experience the same operating system, management and look and feel between the appliance and service module. This consistency allows NetOps to best utilize and manage network services independent of physical packaging and network location, thereby increasing operational efficiency and innovation injection. Thanks to network services being integrated into the Cat6K, and the ability to virtualize services, IT architects are afforded design choices where they can regulate the number of appliances versus service modules in their network by choosing to utilize service modules more over time and obtain their green benefits too. Note that the ASA-SM and ACE-30 can be virtualized or divided between users/groups, thereby extending their reach throughout a corporate network and reducing the number of appliances in the process.

Cat6K with Sup2T Pays to Upgrade to 10GbE

From a pricing point of view, it’s best to think of the Cat6K with Sup2T as the device to transition a campus and mid-range data center network from 1GbE to 10GbE. With 1GbE in the access layer, via upgraded Cat4K with Sup7-E and/or Cat3K / 3750X, connected to a Cat6K with Sup2T in the distribution layer providing 10GbE to the core, Cisco estimates that this configuration will be 20% less costly than a similar configuration utilizing the Sup720 and older versions of the Cat4K and 3K. This design provides for 10GbE between access, distribution and core. In essence, Cisco is paying IT leaders 20% to upgrade to 10GbE with a new generation of switching.

Economics plays a large role in network design. From an economics perspective, Cisco is responding to competitive pressure with new pricing and design options with this Cat6K upgrade. While the Cisco Cat6K Sup2T represents increased performance, what IT business leaders will find is that for typical configurations independent of data center or campus, 1GbE, or 10GbE, the overall cost of a Cat6K network is actually reduced by 20 to 25%. For example, the 48 port 10/100/1000 copper line cards were sold in two versions: centralized and distributed forwarding modes. The centralized forwarding mode is priced at $15K and comes with 256MB of memory, while distributed forwarding is $22.5K. New Ethernet line cards (6800 Series) have Distributed Forwarding Card 4 (DFC4) daughtercards by default and come with 1GB of memory that are priced at the same $15K as the centralized forwarding mode cards, closing the price gap between centralized and distributed forwarding mode to the lower cost centralized pricing. IT architects are offered distributed forwarding performing line cards, which are higher performance throughout the system, at a third of previous generation cards. This is but one important example that demonstrates that the Sup2T is a price reduction over Sup720 around 10GbE.

New Network Design Options and Economics

Campus networking traffic patterns are dominated by north-to-south flows, thanks to the centralization of IT application delivery within data centers. While over time, an increase in east-to-west flows may occur thanks to peer-to-peer applications, north-to-south flows are getting thicker and denser especially as the industry adopts virtualized desktop computing and real time video communications. These thicker north-to-south flows are being accentuated as more applications are being hosted in corporate data centers and private cloud facilities for IT complexity and cost reduction. At the same time, enterprise mobile computing has skyrocketed with the adoption of iPhones, Android-based devices and iPads. For example, Gartner predicts that 55 million tablets will be sold worldwide by the end of 2011. Thanks to lower power output antennas on these new mobile devices, the density of WLAN APs are also increasing to provide coverage. This is creating a challenge to roam seamlessly without user experience interruption.

Mobile and cloud computing economics and increasing traffic volume are driving a new model for campus networking. It’s a model that seeks to increase wired and wireless network bandwidth, scale logical networking and extend network services such as security throughout the enterprise network via centralized management control methods. It’s a model that also seeks greater visibility and control of flows to optimize performance and apply resources where needed. Network virtualization, where physical network infrastructure is logically segmented to assign different network attributes to various groups/departments/entities, has become a mandatory requirement in some industry segments. And from a design point of view, high reliability needs to be systemic as all corporate productivity is flowing across this IT asset.

For those with Cat6K-based networks, installing the Sup2T offers a range of new network design options and economics. For example, encryption is now embedded and integrated. Network services are increasingly becoming virtualized, offering greater reach, cost effectiveness and lower carbon footprint. 10GbE and 40GbE speeds can be strategically placed where bandwidth is needed. NetOps is offered a common look and feel between appliances and service modules, reducing operational cost and increasing efficiency. Logical networking can scale to support more IPv6, more WLAN APs and users, greater visibility into the network via NetFlow, greater stateful application firewalling, etc. It’s clear that Cisco engineering has made tremendous efforts on security with TrustSec, taking ACLs to the next level, NetFlow’s deeper visibility, network virtualization via MPLS or VPLS for segmentation and bringing parity to IPv6 and IPv4.

Cisco is paying customers to upgrade to both the Cat6K Sup2T and 10GbE. Obviously, there’s additional capital cost to spend to gain the return, but from a historic perspective, the upgrade cost is a fraction of previous switch generations. With the Cat6K Sup2T upgrade, IT business leaders gain a wide range of network services, some of which are mentioned above, that will prove to be invaluable as IT marches on toward an IT delivery model dominated by mobile and cloud computing with nearly everything becoming virtualized.

Tweet

By Arista Networks

VMWARE VXLAN is a new network technology developed by VMware that enables stateful VM mobility across traditional L3 routed boundaries. This enables more freedom and flexibility in matching workloads to computing power. By enabling a larger, and essentially flatter network while building on top of proven models for stable scaling of networks such as routing and equal-cost multipath forwarding, VXLAN enables any workload to be provisioned on any virtualized host, anywhere in the network that is IP reachable. No longer do routed topology decisions restrict workload mobility.

If you are a VMware and network administrator who is building virtualized networks with more than 250 VMs or want to stretch a virtual machine farm across two data centers or two or more routed domains with full workload portability, then you need to read this white paper.

Get the White Paper

Tweet

By Extreme Networks

Several technology inflection points are coming together that are fundamentally changing the way networks are architected, deployed and operated, both in the public and private cloud. From performance, to scale, to virtualization support and automation to simplified orchestration, the requirements are rapidly changing and driving new approaches to building data center networks. This white paper does an excellent job at articulating cloud-scale network architecture via an open fabric that accounts for all major industry trends.

Get the White Paper

Tweet

By IBM

Data centers are undergoing monumental paradigm shifts. As demand for greater processing continues to outstrip available floor space, rack space, power and air-conditioning, the market has turned to virtualization to use the resources available efficiently. Most networking switches are not aware of VMs. This creates security and availability issues for both server and network administrators as they try to exploit the value of virtualization and manage this new environment. IBM® System Networking offers VMready®, switches. Find out how these switches solve the most difficult VM mobility and visibility problems.

Get the White Paper

Tweet

Back in November of 2009, I wrote Lippis Report Research Note 136 titled “HP Plans to Acquire 3Com Accelerating a New IT Convergence Era.” In that Research Note, I wrote

“When 3Com is fully integrated into HP what kind of networking revenue and market share can HP gain? ProCurve + 3Com is approximately $2B of revenue now. With the existing product lines can HP generate $5B, $10B or more of network revenue over five years? Time will tell.”

Well after nearly two years, HP Networking or HPN’s North America (NA) layer 2/3 Ethernet switch market share by revenue is nearly the same, bouncing between 5% and 6.1%, according Dell’Oro, with HPN’s Q2CY11 NA switch revenue share being down to 6%. Considering HPN’s limited results after significant investments in sales, channels and marketing, including its “proof-of-concept” plus “A Catalyst for Change” Cisco Trade-in program, not to mention engineering investment, the question is can HP make it in networking? We attempt to answer that question in this Lippis Report Research Note.

Market Share Analysis: 2% Growth Comes from Asia and RoW

HP had approximately 6% WW (Worldwide) layer 2/3 Ethernet switch market revenue share with its ProCurve product line before the 3Com acquisition, according to Dell’Oro. Post 3Com acquisition, HPN’s WW Ethernet switch revenue market share rose to approximately 10%, thanks to 3Com’s 4% share contribution, and stayed that way for three quarters until Q1CY11 where an additional 2% was gained thanks to increases in APR (Asia and Pacific Rim) and RoW (Rest of the World) theaters, according to Dell’Oro. In short, HPN’s NA switch market share has been flat since it acquired 3Com. From a WW switching perspective, HPN’s share of ports has also been flat with 20% share in Q1CY10 to 20.2% share in Q1CY11, according to Dell’Oro. In this same period, NA share of ports has been on a steady decline but with HPN maintaining share thanks to gains in APR and RoW.

In short, in nearly two years, HP gained 2% of WW layer 2/3 Ethernet switch revenue market share, all of which came in Q1CY11 and held during Q2CY11, according to Dell’Oro, and is directly attributed to APR and RoW markets. Its bright spots are in routing and WLANs, which increased 2.5% and 2.2% in revenue share, respectively, between Q1CY10 to Q1CY11. Its IPS/IDS revenue share has been steadily declining, losing .3% share over the same period.

Yes, it’s very difficult to gain share in an established market as HPN has discovered. HPN’s value proposition has been grounded as a lower cost alternative to Cisco, a firm that’s greater than 20 times HPN but sells architected solutions.

Huawei Could Shut Down APR and RoW

HPN’s growth is coming from APR and RoW theaters, which is understandable considering that HP obtained H3C, the once Huawei/3Com joint venture (JV) when HP acquired 3Com. Remember that Huawei and 3Com entered into a JV back in the early 2000s called H3C with the hope that H3C could produce lower cost networking products that 3Com would sell in NA while opening up the Chinese market. In Lippis Report Research Note 16, Bruce Claflin, 3Com’s then President and CEO, had hoped that H3C would deliver success much like Amdahl did over IBM in the 1980s and 1990s when Amdahl gained huge market share from IBM in the Front End Processor (FEP) business by offering similar products priced well below IBM.

Fast forward to late 2006 when Huawei agreed to sell its stake in H3C to 3Com. Huawie had a non-compete agreement with 3Com post the sale of its stake in H3C, which has since expired, allowing Huawie to more aggressively and organically pursue the Ethernet switch market. And it has, as in early 2011, Huawie announced a new Enterprise Business Division.

Surprisingly H3C’s massive product portfolio has not made it into the HPN NA channel, partly explaining HPN’s flat NA share growth. H3C’s products were to be HPN’s competitive advantage. More alarming for HP, however, is the prospect that Huawie’s Enterprise Business Division will bring its enterprise product portfolio right to H3Cs Asian customers, cutting off HPN from this bright spot. Also when H3C was partly owned by Huawei, the Chinese government was tremendously supportive of H3C, but since H3C is 100% owned by HP, the Chinese government has no incentive to support H3C and will more than likely shift its support to Huawie when its Enterprise portfolio is ready. The danger here is that in the quarters to come, HPN’s APR and RoW market could start to dry up. Much of the future growth for H3C had been pinned on continuing its China dominance. But wait it gets worse.

Huawie is threatening to hijack Bruce Claflin’s and now HPN’s low cost networking value proposition and use it for its own advantage. First Huawie will more than likely go after the H3C installed base in Asia then onward to NA and Europe. One possible scenario has HPN competing with Huawie as to who is the lowest cost provider of networking. This would push HPN up market and force it to change its value proposition to an architected solution, where it will find Cisco. HPN has started to move in this direction with its recently announced FlexNetwork Architecture. This scenario would, in essence, squeeze HPN between Huawie on the low end and Cisco on the high end. If networking gets into a price war game, Huawei could out low price HPN and that should be the major concern to HPN as it represents an estimated $800 million a year in revenue.

But Huawie will face stiff headwinds in NA as Huawei has a credibility problem with most North American buyers. IT business leaders know it as a low cost provider and that Cisco did a good job of raising the visibility of how Huawei tried to steal intellectual property source code. Therefore, while Huawei could have some impact in NA, the most immediate opportunity for Huawei enterprise is in China, specifically the install base that H3C had built.

Lacking Data Center Network Strategy and Products

HP certainly has product to support one of the most comprehensive data center visions in the industry. HP has servers, storage, a huge services group and network products. HPN’s FlexNetwork architecture is an interesting vision if an IT architect wishes to extend a fabric across an entire campus, branch and data center but the underlying architectural detail and products are missing. The A12500 series has been available for two years, but not in NA in any great numbers. HPN recently said that it will be available in the 2H2011. The new A10500 data center switch was announced in May but is scheduled to ship some time in the second half of 2012. HP’s networking strategy in highly virtualized data centers is limited to its Virtual Connect product. HPN’s data center networking share according to Infonetics, and UBS is estimated at 6% versus Cisco’s 81%. This is where the networking market is at its hottest versus HPN’s strong hold in education and low cost networking.

For a company with the portfolio size of HP and its strength in data centers, it’s curious that HP is the only mainstream network vendor that doesn’t have a good data center fabric story. Cisco clearly does, as does Brocade, Juniper, Extreme, Dell/Force10, Arista Networks, Alcatel Lucent, IBM, Mellanox, etc. HP doesn’t, and it’s surprising, considering its large position in the data center market. It would be refreshing to hear HP communicate what a unique HP data center architecture looks like tied into mainstream industry pain points.

How Can HPN Win?

How can HPN turn this around and participate in an effective way, utilizing its deep assets of broad product line, services, software, support, brand, financial strength and low price points to bring value to both customer and shareholders? Certainly HPN has product but it needs to bring the H3C products to NA and wrap the services group around them. HPN needs high performance and low latency 10GbE and 40GbE data center switching products since 10GbE represents some 25% of the total Ethernet switch market and growing, according to Infonetics. HPN recently announced a family of Top of Rack (ToR) switches called the 5830-switch family targeted for 2H2011 availability, but few details are available. HPN should consider acquiring Arista Networks, which may cost it two quarters of switching revenue but would add between 5 and 10% to its switch revenue and plug a major hole in its networking product line.

In addition, HPN needs leadership consistency as HPN has transitioned leadership from Marius Haas, previous HPN GM who left HP for KKR in May, to now Bethany Meyer, a marketing executive who is interim SVP and GM of HPN. Bottom line: HPN needs to create leadership stability. The first order of business for whomever is to lead HPN should be to communicate what the unique HPN vision is as it’s still not clear to the market. In short, what is it about the HP data center and HPN that’s going to create a competitive advantage over Cisco, IBM, Dell and Oracle other than low cost. For example, consider Cisco’s data center vision, which is very clear. Cisco’s data center business advantage architecture is a system’s approach that bundles products together to deliver business outcomes.

The above is a straight-line approach to winning an established game, but HP needs to do something big and radical that is out of the box but meets market needs. It could consider acquiring Xsigo, a firm that recently released its server-based fabric as an alternative to processing at the network layer. This could be an approach that disrupts what networking actually is in the data center. HP would best be served to develop a compute centric view of the world. Clearly some IT business leaders will buy into this model while others may not, but one thing is certain and that is data center computing buyers tend to be closer to the CIO, offering HP a potential competitive advantage.

HPN needs to develop a new vision for computing and networking, and deliver it via a bold strategy and vision that’s disruptive rather than “we sell cheaper than everybody else.” HP has the brainpower and financials to develop a disruptive approach to data center networking; they just need the thought and executive leadership. In short, HPN needs to lead this industry and not just be a fast follower.

Tweet

By IBM

Use of Ethernet as a switching fabric provides servers with a single connection and can greatly reduce the amount of equipment required in the data center. Companies with storage networks are switching from Fiber Channel to Ethernet-based storage solutions that use 10GbE. This trend is accelerating now with lossless DCB or Data Center Bridging Ethernet products such as IBM BNT RackSwitch G8124. With the adoption of the new DCB Ethernet protocols, Ethernet switching fabric can offer the technical features and the economic value necessary to become the switching fabric of choice for data center networking, storage and clustering. Find out how by downloading this white paper.

Get the White Paper

Tweet

By Ken Won, Force10 Networks

Most IT organizations have implemented server, storage and/
or network virtualization to some degree, but to fully realize
the benefits of virtualization, IT organizations need to consider
adopting all three in an integrated manner. Server virtualization for
x86 servers has been around for about a decade, and is relatively
mature. Storage virtualization is not quite as mature, but numerous
storage virtualization products are available today. And while
network virtualization in the form of VLANs and VPN has been
around for many years, new types of network virtualization are just
now being introduced to the industry. In this white paper, we’ll
take a look at virtualized servers, storage, and networking, and see
how automated network switching helps unify these environments
into a cohesive whole.

Get the White Paper

Tweet

By IBM

This white paper provides a summary of how IBM® Tivoli®
Storage Productivity Center supports the IBM System Storage®
SAN Volume Controller. This information is intended to describe
the management capabilities of Tivoli Storage Productivity Center
in a virtualized storage configuration.

Get the White Paper

Tweet

Avaya has expanded its cloud networking product portfolio with the Virtual Services Platform or VSP 7000, a 10GbE Top of Rack (ToR) Switch and its new Virtualization Provisioning Service or VPS management software.
The VSP 7000 boasts a multi-Terabit Fabric capable of horizontally stacking many 7000s, offering a new approach to server connections. The VPS virtualization management application provides visibility and automated control of virtualized infrastructure. VPS is designed to provide NetOps and SysOps shared visibility of physical, logical and now virtual IT assets. Dan DeBacker, Director of Data Solutions Architecture at Avaya, discusses the new VSP 7000 plus VPS and the new data center design options they afford.

Listen to the Podcast

Tweet

In Cisco’s Data Center Fabric, it has delivered a set of features and innovations that solve some of the most difficult networking challenges found in virtualized infrastructure. IP address and VM mobility plus adapter and VM Fabric EXtenders (FEX) offer increased support for virtualized data center infrastructure, offering designers flexibility to move virtualized assets independent of location. These innovations are proposed by Cisco that promises virtualization aware networking, lower cost and increased performance. Omar Sultan, Senior Manager, Data Center Architecture at Cisco Systems, and I discuss Cisco’s new data center virtualization tools.

Listen to the Podcast

Tweet

Two years after Cisco launched its Unified Computing System it has 5,400 customers, holds the #3 market share ranking for x86 blade servers WW, behind only HP and IBM, according to IDG and recently broke numerous world computing performance benchmark records. While UCS has leaped frog competitors with performance plus memory and I/O capacity the most important aspect of UCS is the business value it drives. I explore this topic with Todd Brannon, Senior Manager for UCS marketing at Cisco Systems about the vision and strategy of Cisco’s Fabric Compute and the value its customers are gaining from its use. Todd brings great customer examples to this podcast, which is a must for any IT leader evaluating a data center fabric.

Listen to the Podcast

Tweet

The tech sector is at a crossroads. In just 18 short months, mobile and cloud computing has fundamentally changed business assumptions and technical underpinnings of IT delivery. And in the process IT business leaders are fundamentally changing their buying requirements and corporate IT investments challenging existing vendor relationships. The tech sector served up corporate IT along technical lines of computing, networking, storage and applications, but these lines are blurring as every major multi-billion dollar IT firm now seeks to deliver vertical offerings comprised of a single rack of compute, storage and networking to address scale and simplicity associated with the new mobile and cloud computing models. Cisco, IBM, HP, Dell and Oracle all are repositioning their data center offers to address the market opportunity and shift to assist IT leaders building iBusinesses. In this Lippis Report Research Note, we dive into Cisco’s Data Center Fabric as it’s the furthest along at integrating compute, networking and storage access for corporate advantage offering a glimpse of IT’s future.

What’s driving a new fabric or structure of data centers is rooted in the interplay between technology and business opportunity. The efficiency of server virtualization to reduce energy consumption and increase server utilization drove its massive deployment that was boosted by an economic cycle starving for efficiency. At nearly the same time, mobile computing, thanks in large part to Apple’s iPhone and iPad plus Google’s android- based devices, introduced a new tier of computing that unleashed increased corporate productivity, evident in today’s productivity boom. Equipped with a new IT delivery model that is both more flexible and centralized, IT business leaders have begun en masse to build private cloud facilities.

The iBusiness

The end result is the construction of iBusinesses that possess simultaneously lower IT cost and the ability to quickly address market dynamics, thanks to faster application deployments plus a nimbler and mobile workforce. While it’s too early to aggregate the benefits of iBusiness in terms of productivity improvements, market share gains, IT expense as a percentage of corporate revenue and other metrics, early adopters are experiencing improvements that span IT departments and most importantly, corporate operations.

In short, a Data Center Fabric of compute, networking and storage reduce IT operational cost, the largest budget component of IT Total Cost of Ownership (TCO) and provide the foundation for a faster responding business that is able to exploit the value of mobile and cloud computing to corporate advantage.

Data Center Fabric Requirements

A core set of data center fabric requirements is emerging, thanks to early adopter deployments that possess the following attributes fundamental to iBusinesses.

Scale: Computational density is increasing at a fast pace with the ability to support hundreds to hundreds of thousands of servers per data center. This increased density of computing is also driving higher virtualization ratios as the ratio of virtual to physical servers is increasing from 10:1 to soon 60:1, which taxes the logical network of MAC address, /32 IP host route table size and ARP entry size. The ability to support both east-west and north-south traffic flows over an increasingly 10GbE and 40GbE low latency, non-blocking, high performance network fabric has become paramount as small queries from mobile devices drive a tsunami of east-west plus north-south data center traffic flows, all of which must be combined and transmitted back to the mobile device at millisecond speeds.

Mobility: As virtual machines (VMs) are moved within and between racks of computing and between data centers plus between private and public cloud facilities, the ability of the Data Center Fabric to support such moves is fundamental. VM aware Data Center Fabrics support VM mobility, allowing IT business leaders to maximize efficiency while enabling a degree of freedom to move containers of IT workloads (data, applications, VMs) as business requirements demand.

Consolidated IO: A significant cost reduction strategy and performance enhancement is the deployment of a single physical 10GbE and soon 40GbE network that supports both storage and network traffic. Cost savings is found in reduced cabling requirements, storage and network switches as well as server network and storage interface cards.

Consolidated Management: As compute, storage and networking converge into a single virtualized Data Center Fabric, the ability to manage these resources across operational groups become increasingly important. Not only is the technology converging, but IT organizational design is under review to focus this human resource into a services organization rather than siloed technology departments. The ability to manage the Data Center Fabric as a centralized resource that is partitioned to unique IT departments is an aid to organization re-design. It’s very helpful that a common look and feel for all resources be available so as to hasten a learning curve and accelerate cross-discipline service delivery.

Cloud Enabled: The combination of the above attributes results in a Data Center Fabric that is cloud-enabled, meaning that containers of workload are movable not only within a data center but also between them and into private and public cloud facilities. The ability to move workloads provides IT leaders with the tools to expand and contract their IT resources and shop their IT needs from a wide range of cloud providers, assuring executive management that their IT cost is competitive.

iBusiness Outcomes

Those who have deployed a Data Center Fabric are rewarded with favorable business outcome results. Cisco’s Data Center Fabric unifies network services, networking and storage plus computing through its Unified Network Services (UNS), Unified Fabric (UF) and Unified Computing System (UCS), respectively. Early adopters have benefited by viewing and procuring their data center assets from this unified holistic perspective versus compute, network and storage separately. For example, Kindred Healthcare saved approximately $6.6M on just cabling cost for a 1,000-server data center, thanks to its deployment of a Data Center Fabric. Additional operational savings was gained by a reduction in the number of management points the operations group has to manage too. To Kindred’s surprise and delight they noticed that the Data Center Fabric enabled different groups—the virtualization team, the network team, and the storage team—to work together as one on a common platform versus in silos; a huge help to hasten deployments especially as Kindred has been growing through acquisitions.

Other early adopters are Almaviva wine producers that saw its revenue increase 2 to 3%, thanks to its data center fabric deployment that also reduced its cabling and power consumption cost by 70% and 60%, respectively. Tutor Perini Corporation was able to reduce its device count and power consumption by 60% and 38%, respectively. Coca Cola was able to consolidate 80 servers down to four, plus reduced cabling 30 to 60%. Terremark saw a 30% improvement in application performance and server density increased by a factor of four. The Apollo Group, owner of the University of Phoenix and other educational properties, doubled the size of its network without an increase in IT staff, lowered per-port switching cost while increasing port volume and freed up several rows of space in its data centers. Avago Technologies, a manufacturer, accelerated batch processing by 30 to 40%, increased business flexibility and decreased operational cost by 40% while adding a third data center. CareCore National, a health benefit management concern, increased business agility by being able to launch new lines of business in just two weeks, down from six months. These iBusinesses’ benefits were gained, in large part through the insight and leadership of IT executives and their deployment of Cisco’s Data Center Fabric architecture.

Cisco has been investing heavily in its Data Center Fabric portfolio. It owns some 80% of the data center switching market and in just two short years, possesses the number three-market share ranking for x86 blade servers worldwide, behind HP and IBM, according to an IDC report released in May. Over the past quarter, Cisco has added to its UF portfolio with the new Nexus 3000, 5548 and 5596 switches. It has expanded its Fabric Extender (FEX) offering to include the adapter and VM FEX, a key technology in converged IO plus virtualization aware networking. To increase mobility of workloads, it has added IP address location independence with its OTV (Overlay Transport Virtualization) and LISP (Location ID/Separation Protocol) features to its Nexus Operating System. Fiber Channel over Ethernet (FCoE) can traverse more devices, thanks to a new director-class multihop FCoE feature available on the Nexus 7000 and MDS 9500. Data Center LANs, SANs and virtualization infrastructure can now be managed via a single pane of glass, thanks to the Cisco Data Center Network Manager. On the computing side, Cisco has expanded the UCS server portfolio with multiple form factors, including Blade and Rack-Mounted, and in the process, has broke three world performance records. Cisco has followed up that with a new set of I/O components for UCS, which was just announced on July 13th.

At the crossroads of the tech industry are two paths; one is a legacy approach of building data centers by acquiring compute, storage and networking gear separately with IT professionals integrating these components. The other road is one of vertically-integrated offerings of compute, storage and networking where IT professionals focus on automating business processes turning their corporation into an agile iBusiness. I advise choosing the latter.

Tweet

Cisco’s Nexus Operating System that runs on the Nexus switches and evolved from Cisco’s MDS SAN-OS is a fundamental building block of its Data Center Fabric. Berna Devrim, Senior Manager of Data Center and Virtualization Marketing at Cisco Systems, discusses the next generation of Nexus OS designed to address the biggest data center issues. These issues are virtualization scale and mobility, cloud spec scale, LAN and SAN convergence plus operational efficiency. This is one of the best audio podcast we’ve produced, so sit back listen, learn and enjoy.

Listen to the Podcast

Tweet

In November of 2010 Avaya launched its Virtual Enterprise Network Architecture or VENA, which describes its computer networking investment strategy and technology direction. VENA was designed for virtualized data center infrastructure but has been expanded to include campus and branch office networking. Dan DeBacker, Director of Data Solutions Architecture at Avaya discusses the unique attributes VENA possesses to meet today’s enterprise networking needs and the business outcomes its customers are gaining by its deployment.

Listen to the Podcast

Tweet

By Infoblox

Organizations of all sizes are transitioning to virtualization technology at a dramatically increasing rate. Conducted between September and December 2010, Gartner’s annual survey of 2,014 CIOs collectively responsible for $160 billion in IT spending in 50 countries across 38 industries found that “Currently, 3% of CIOs have the majority of IT running in the cloud or on SaaS technologies, but over the next four years CIOs expect this number to increase to 43%.” The motivation for adopting the new technology is enticing: increased flexibility, decreased workload for internal IT departments, faster access to on-the-fly infrastructure needs, increased worker productivity, and significant capex cost reductions. Despite this worldwide trend across all industries, most organizations will fail in the conversion process to these new technologies, thanks to complexity. This white paper explains how network automation calms complexity allowing IT business leaders to reap the rewards of dense virtualization deployment.

Get the White Paper

Tweet

by Force10 Networks

As data centers scale to support thousands of servers, IT managers are seeking better ways to network those servers while reducing costs and power consumption. Moreover, in large-scale data center cluster environments, inter-node communication bandwidth is increasingly becoming the main bottleneck. Compute nodes located across different physical switches may not have full bandwidth in a conventional hierarchical network design of interconnected switches. The solution is a distributed core architecture based on low-cost, high-capacity switches. This paper describes the use of Force10 Network’s® Z9000™ core switching system in a distributed core architecture to address these issues.

Get the White Paper

Tweet

By IBM

How are technology leaders helping their organizations adapt to the accelerating change and complexity that mark today’s competitive and
economic landscape? To find out, we spoke in person with 3,018 CIOs, spanning 71 countries and 18 industries. They shared how they are innovating with technology for organizational success. CIOs increasingly help their organizations cope with complexity by simplifying operations, business processes, products and services. To increase competitiveness, 83 percent of CIOs have visionary plans that include business intelligence and analytics, followed by mobility solutions (74 percent) and virtualization (68 percent). Since our 2009 Global CIO Study, cloud computing shot up in priority, selected by 45 percent more CIOs than before and leaping into a tie for fourth place with business process management (60 percent each). In this report, CIOs provide you insight in to both their challenges and opportunities from increasing complexity.

Get the White Paper

Tweet

In the quest to build a data center network that is flat, fast and fault tolerant, the networking industry has served up a new two-tier architecture made up of Top-of-Rack or ToR server connecting switches and highly dense Core switches. Core switches connect ToR devices. The key question is how best to provide switch connectivity to deliver a lossless, high performance, low latency fabric that supports virtualization mobility. The answer is found in Multi-Chassis Link Aggregation Group or MC-LAG, Transparent Interconnection of Lots of Links or TRILL and/or Shortest Path Bridging (SPB). Dhritiman Dasgupta, Director of Product Marketing for Fabric and Switching Technologies at Juniper Networks, discusses its QFabric approach and why TRILL’s hair pinning method to inter-VLAN routing is a poor approach. It’s a fascinating discussion that you have to listen to. You can also download a white paper on this topic here.

Listen to the Podcast

Tweet

Cisco is the only large vertical IT supplier with compute, storage and deep networking capabilities making its data center offering unique. It’s Unified Network Services or UNS, Unified Computing System or UCS, Unified Fabric and policy based management make up the Cisco Data Center Fabric. In this Lippis Report podcast I talk with Shashi Kiran; Director of Market Management for Data Center/Virtualization at Cisco Systems about the vision and strategy of Cisco’s Data Center Fabric and the value its customers are gaining from its use.

Watch the Video

Tweet

IBM’s System Networking is focused upon creating an interconnect fabric/System Network within data centers upon which servers and storage rely upon to deliver IT applications. IBM recently formed IBM System Networking that includes BNT along with partnerships of Cisco Brocade, Juniper and Mellanox plus its management solutions of Tivoli. Vikram Mehta, Vice President IBM System Networking is my guest as we discuss IBM’s expanding role in System Networking.

Watch the Video

Tweet

by Force10 Networks

Cloud computing is an emerging computing model that promises a new era of flexibility and control in providing data center resources. In the cloud model, data center managers can mix and match computing, storage, and networking resources to provide an agile and highly flexible resource for customer applications. To realize the full potential, this paradigm requires open, standardized interfaces between data center layers of compute resources, the network, and storage elements. While the industry has moved toward open computing and storage layers over the past few years, networking has remained largely proprietary. Force10’s Open Cloud NetworkingSM framework is intended to unlock the network layer so data center operators can get the most out of their data center architectures and, in turn, get the most out of their cloud deployments.

Get the White Paper

Tweet

Force10 introduced a disruptive set of products to address the new cloud networking market driven by massive virtualization and cloud spec scale data centers. First its new Z9000 Distributed Core System promises to be 1/10th the footprint, at 1/5th the price and 1/20th the power consumption of other core switches. The Z9000 is 32x40GbE or 128x10GbE platform where multiple Z9000s are connected via an open distributed core supporting up to 24,000 10GbE server links. In addition, Force10 introduced a next generation ToR switch, the S7000 that supports 40x10GbE, 12×2/4/8G FC plus 4x40GbE combining networking, storage and Open Cloud Applications. Arpit Joshipura, Chief Marketing Officer at Force10 Networks, discusses Force10’s Open Cloud Networking products and strategy, and the new design options they afford to IT architects.

Listen to the Podcast

Tweet

IBM’s System Networking is focused upon creating an interconnect fabric/System Network within data centers upon which servers and storage rely upon to deliver IT applications. IBM recently formed IBM System Networking that includes BNT along with partnerships of Cisco, Brocade, Juniper and Mellanox plus its management solutions of Tivoli. Vikram Mehta, Vice President, IBM System Networking, is my guest as we discuss IBM’s expanding role in System Networking.

Listen to the Podcast

Tweet

In Lippis Report 172, I mentioned three huge trends that are starting to interact with each other creating a perfect storm that is gripping the tech industry. One of those trends is the creation of a software ecosystem in the networking market, thanks to the Clean Slate program out of Stanford University that has spawned the Software Defined Network (SDN) initiative and open controller protocol called OpenFlow. I spent a week in the Valley talking to people at Stanford and many industry executives from Cisco, Juniper, Marvell, Big Switch, Nicira, Arista, IBM and others. In this Lippis Report Research Note, I share with you what I learned. OpenFlow-based SDN is being both hyped and in its current state, limited, but it does represent a new paradigm that has the industry abuzz, filled with possibilities.

Centralized Controller Model

OpenFlow is a protocol, or API, that modifies forwarding tables in network switches. It sits between a switch and controller. The controller can run on a centralized computer/server that has an Über view of the network and its topology. When a packet enters a switch and the forwarding table does not contain a path for the packet, it’s passed to the controller. The controller then searches the packet’s destination address and defines a table entry with associated attributes to create a path through the network, which the packet and subsequent packets are to follow. The controller then sends a message to each switch in the path the packet will traverse via the switch’s OpenFlow API, which modifies the switch’s forwarding table. Every subsequent packet with the same destination address will then be forwarded based upon this table in cut-through mode. The first store-and-forward stage takes about 50ms; yes, a long time, but it can be significantly shortened. Subsequent packets being forwarded in cut-through mode travel at switch latency, which for 10GbE Top-of-Rack (ToR) switches is between 500ns and a few microseconds.

Now this search method is a bit controversial as some claim that all that the controller needs is a large TCAM to compute the table flow. Some worry that a Cartesian explosion may occur, corrupting the calculation, but this is an engineering problem with an engineering solution, perhaps via multi-staging the flow tables.

This centralized controller model can scale as has been proven in distributed computing models used by all the major cloud providers. An example at Stanford demonstrated that a network of 35,000 PCs with approximately 2,000 switches generated 15 to 20k flows/sec. A controller can support 2M flows/sec at half a 2007 PC processor capacity. Further, modern 48-port ToR switches can request 100s of flows/sec with controllers supporting 2M flows/sec, which means that a single controller can support 10s of thousands of ToR switches. In short, a centralized controller-based OpenFlow SDN can theoretically scale.

How an OpenFlow SDN Is Different Than Today’s Network Architecture

The above model departs significantly from today’s network architecture in a few key ways. First there is the concept of a centralized controller(s) versus a distributed packet forwarding architecture based upon topology discovery. There may be separate links for control and data plane communications, which would also be a significant departure from today’s single physical network that supports both control information and data forwarding. There is no layer 2 and 3 construct in an OpenFlow SDN, which has been the semantics of computer networking over the past twenty plus years.

Software Defined Network Ecosystem

Further, on top of the controller is another API, yet to be fully defined, that enables application developers to write network applications without knowledge of the underlying network structure. In short, the API abstracts the network, allowing the programmer to focus on what she/he needs to accomplish versus how to configure the network to comply. The creation of a software ecosystem creates the possibility of a new network paradigm where low cost Asian switches populated with SDN software force an economic collapse of the existing network market. While this is highly unlikely, it does warrant careful observation and mitigation planning on the part of established vendors.

An OpenFlow SDN offers significant differences, which is why there is such excitement surrounding OpenFlow. The genius of the approach is the separation of data and control plain so that SOA-based application developers and researchers can layer applications onto the network, injecting innovation at speed via a software ecosystem. Further centralized controller-based networks such as the national cellular network plus dense compute management have proven to reduce operational cost and increase control in complex systems.

There is an industry group called the Open Network Foundation, or ONF, that is promoting the use and interoperability of OpenFlow SDN enabled switches. The above OpenFlow SDN example is primarily an academic description as OpenFlow is well regarded as the leading open implementation to date for providing SDNs within the research community. But there will be many networking concerns introducing controllers that reside in the switch. Further, the definition of a controller is a bit vague as some define it as a network operating system, such as Cisco’s IOS or NX-OS, Juniper’s JUNOs, Arista’s EOS, etc., while others define it as a management entity, performing configuration changes. But before we dive into this, let me explain a few problems that an OpenFlow SDN may solve.

Innovation at Speed: The institutions that were created to assure interoperability and inject innovation into our industry have become too cumbersome and slow such that networking has fallen behind compute and storage advances. The way innovation is injected into networking today is that a proposal is made to a standards group, such as the IETF, IEEE, etc., and all interested parties compete for the best ideas or technical advantage. This process can take a few years just to modify a few bits in the header of a packet. Then, once the standard is completed, companies build to it, which can take another eighteen to twenty-four months. This approach is not serving the industry any longer, and there needs to a more rapid way to inject innovation. An OpenFlow SDN promises such an approach where applications can be added to the network rapidly, thanks to the abstraction of layer 2 and 3 forwarding.

Traffic Engineering: Fine-grained traffic engineering utilizing a variety of forwarding actions is an application that service providers and enterprises seek to optimize application performance.

Tagging vs. Table Manipulation: There is much agreement in the industry that the network has become too ridged in virtualized data centers, restricting the movement of VMs between racks, data centers, etc. Further, as appliances such as firewalls, load balancers, IPS, etc., have become virtualized, there needs to be a method to steer traffic to them to service an application. The industry has responded to this by proposing the placing of tags on packets to guide its path to the right VM, appliance. An OpenFlow SDN implementation could simply modify switch-forwarding tables to guide the application through a chain of appliances mitigating tagging and offering applications appliance servicing within highly virtualized infrastructures.

The Real World

An OpenFlow SDN is new, and it’s unrealistic to think that it’s without challenges; here are some OpenFlow challenges.

Trust: The single largest issue an OpenFlow SDN has is trust. Will IT business leaders trust it within their networks, especially their data center? If a controller is sourced from a new company, how comfortable will the IT team be that it’s modifying switch-forwarding tables? How many controllers are needed for a particular load? What will the support model be? How complicated will it be to manage multiple controllers?

Interoperability: The current construct of OpenFlow requires knowledge of the switch’s hardware semantics of L2/L3/VLAN architecture; therefore, each controller implementation may be different and thus unclear how controller interoperability is achieved. Further, it’s unclear how applications written for one controller will work on another.

Network Stability: This issue may be linked with trust, but it’s unclear why a third-party controller should search packets to define a path through the network topology. Rather, why not use existing network operating systems for what they are good at– topology discovery, etc.–so that IT business leaders are more comfortable running OpenFlow-based SDN applications on top of a stable network. In short, will OpenFlow controllers introduce instability?

Controller Placement: If we take the definition of a controller to include existing network operating systems, then there will be both distributed and centralized controllers within a network. From a design point of view, how does an IT architect approach distributed versus centralized controllers and what are the trade-offs?

It’s unfair to expect that a new approach to networking would have the above issues all sorted out before deployment. These are not barriers to entry but rather challenges that the OpenFlow SDN community will work on over the next one to two business cycles. Let me be clear…OpenFlow-based SDN is a very big deal and is being embraced by all vendors including established firms and start-ups. What is driving most companies is the promise of a software ecosystem to inject innovation and value into their network products.

Established firms will support OpenFlow SDN via OpenFlow client reference implementation within their switches but will add proprietary extensions that differentiate their OpenFlow version from others. Cisco, Juniper, Arista, et al, will differentiate based upon how much of their network operating system they expose. Established firms should have an advantage over smaller ones in attracting software developers as their installed base is much larger.

New companies such as Big Switch Networks and Nicira will focus on solving particular problems in the data center, service provider and enterprise network that existing layer 2/3 networks either don’t solve or don’t solve easily. Virtualization of both servers and desktop are two prime areas, and I expect a suite of SDN Virtualized Applications to emerge from these firms and others.

The service provider market is perhaps the biggest OpenFlow SDN winner as early experiments have shown that the existing three-tier service provider architecture of packet switching, optical core and edge may shrink over time to just two, thanks to traffic management applications.

OpenFlow SDN has successfully introduced the concept of controller-based networking and the controller market. OpenFlow 1.1 is in standardization process and once completed, will be the first defined open controller API to communicate between network and controller, offering greater control of cloud network resources and management. But perhaps the greatest contribution an OpenFlow SDN will offer is the potential to usher in a wave of fast-paced innovation not seen before in the networking industry.

Tweet

In the world of High-Frequency Trading (HFT), opportunities exist only fleetingly and therefore trading solutions must run at the lowest latency to be competitive. Low-latency 10 Gigabit Ethernet has become the interconnect of choice for HFT solutions. IBM® and Mellanox® have demonstrated a solution that performs at high throughput rates and low latency to facilitate High-Frequency Trading solutions.

Get the White Paper

Tweet

Three strong trends are taking shape that are so powerful they threaten the status quo of the networking industry. These trends are more like storms than new markets; in fact they represent a major industry discontinuity. The first storm is happening now and is represented by merchant silicon for 10 and 40 GbE chips lowering the barrier of entry for new entrants in the Ethernet switch market. The second storm is much weaker but promises to be just as big, or bigger, than the first. This second storm is the creation of a software ecosystem in the networking industry, thanks to initiatives such as Software Defined Networks (SDN), OpenFlow, Arista Network’s EOS Central, etc. The third storm is the paradigm shift in enterprise IT spending thanks to mobile and cloud computing. These three storms are starting to interact and feed upon each other, forming a perfect storm in the networking industry. The perfect storm is already doing damage, as all major IT firms position product portfolios to navigate through it and prepare for its aftermath of making existing networking legacy.

IBM, for example, sees the perfect storm as an opportunity to optimize performance of IBM systems for new and emerging workloads like cloud computing and analytics that require instant access to information by investing in networking. In this Lippis Report Research Note, we focus on IBM’s networking strategy and analyze its potential impact.

IBM created the System Networking group to organize its network resources and execute its strategy. It’s a strategy to implement a data center fabric that ushers in a smart compute model that federates storage, compute, memory and I/O into pools of resources that are brought together to meet business requirements. It recently acquired BLADE Network Technologies (BNT), which produces blade and Top of Rack (ToR) data center switching gear, network-aware virtualization technology, load-balancing and management software. From an organizational point of view, IBM System Networking includes BNT and an IBM group that used to be called Data Center Networking that possesses Fiber Channel and InfiniBand assets. System Networking also maintains working relationships with networking leaders such as Brocade, Cisco, Juniper Systems and Mellanox.

IBM has been selling system networking solutions with its servers and storage offerings for decades. Systems and networking are now more interconnected, making it important to continue partnering with core networking providers like Brocade, Cisco and Juniper while enabling closer connections with IBM servers by increased investment in systems networking technology.

But why did IBM enter the System Networking business and why now? In short, IBM executives saw an opportunity to gain control of a critical data center asset, address customer needs, and add a key component to deliver on its vision of Smarter Computing. From discussions with IBM executives, they stress common concerns of their largest data center customers, which have propelled IBM into the System Networking business. Clearly, Cisco’s launch of Unified Computing System or UCS and the forecasted perfect storm also factor heavily into IBM’s calculus. IBM is hearing demand and seeing a shift in the networking industry that has opened a door for it to be a leader in data center enterprise networking, or System Networking, as IBM now calls it.

Cloud Spec Scale

The largest data center customers are implementing cloud spec facilities that are boosting up their infrastructure spend and deployment by an order of magnitude in many cases. Yes, that’s ten times the size of their normal data centers. This scale has created unique problems that challenge linear approaches and are focusing IT business managers to seek alternative solutions to scale.

The old model of increasing capacity of memory, compute, I/O, and storage, etc., by acquiring more servers does not work any longer. IBM seeks to solve this scale problem with Smarter Computing that delivers elastic services to federate a pool of resources that are brought together to meet business needs for Big Data analytics and private and public clouds. Resources could be memory, I/O, compute or storage. The goal is to bring together the right proportion of resources to solve a particular workload.

Why Networking Is Important to IBM

To deliver on Smarter Computing, IBM realized that to offer a federated pool of resources, it needs a network fabric that connects these assets, and thus this is what System Networking is all about. IBM let other industry players connect high-density blade and rack system with their network gear. This left IBM out of the innovation loop and control allowing others to set the rate and pace of network innovation.

The need to own the network and provide IT business leaders with vertical IT expertise has become apparent. If the data center rack is the new computer, and multiple racks are the new super pod, how does a supplier make this system look and feel like one large computer? It all starts with connecting these elements together in a very smart fashion using physical connections and software to orchestrate resources and infrastructure simpler than today’s approach.

How can IBM make dense IT infrastructure simpler to deploy and manage as its largest customers deploy ten times more infrastructure? Most IT business leaders translate this into the need for rack infrastructure management, configuration management, and database technology to keep track of IT assets, etc. While IBM has director and utility tools, System Networking is a critical component of Smarter Computing. IBM executive management figured that System Networking will play an even more important role in solving new IT business leader requirements that include simplifying massive amounts of IT infrastructure installation and orchestration, be it physical or virtual.

At the high end of the enterprise computing market, IT business leaders are acquiring IT assets like airlines buy airplanes and hotel builders buy property. Both airlines and hospitality concerns worry about the same thing: use or occupancy rate management. Airlines want to ensure that they have the right size aircraft for a particular flight route so that few, if any, seats are left empty.

As IT business leaders scale up their data centers to cloud spec, thanks to IT service demand, how do they ensure that the capacity acquired is effectively utilized and not over or under designed? Most, if not all, IT business leaders have embraced server virtualization as the key technology affording efficiency gain.

Without System Networking, IBM management realized that it was unable to address IT business leaders’ full virtualization requirements. The data center network needs to be virtual machine aware. In fact, this is one of the biggest reasons why IBM acquired BNT as IBM needed BNT’s network virtualization expertise.

More Business Goes Online

The reason why IT business leaders are deploying so much more infrastructure is that more of their business is going online. Just think about your average day. When communicating to each other we text, email, VoIP and videoconference. When you want to go see a movie, you book it online. You bank online. You pay your bills online. You trade stock online. You make airline reservations online. You read news online, your photos are stored online, office productivity tools are online, etc. As more and more business goes online, the scale of IT infrastructure needed increases.

In addition to more business going online, IBM’s big analytics business needed networking too to be first class. IT business leaders are putting in place more analytic systems, decision support systems and data warehousing systems so they can mine their depositories of vast amount of information that they have about customers, business, products, competitors and supply chain, etc., so they can make smart important business decisions.

This is why data warehousing, data mining, smart analytics or solving the big data equation is so important to IBM. This is why IBM acquired Netezza. Now, what is the difference between a good data warehousing engine and a great one? The answer: how fast data can be transported to and from the analytic engine, or how fast is the network. For IBM to be a successful player in smart analytics and be recognized as the clear leader in this large and very important market, it realized that it needed to be in the networking business.

Controlling TCO at Scale

As data centers have been scaling up, so too has Total Cost of Ownership or TCO. For every dollar that CFOs spend on servers and storage, they spend between 15 and 25 cents on networking. IBM is not able to control a customer’s TCO as it has no control over 15 to 25% of the IT budget. Therefore, how could IBM profess to solve the TCO equation when it can’t provide a credible solution to 15 to 25% of the TCO problem? IBM needed to have a voice and solution for TCO, thus this too factored into its thinking of re-entering the networking industry.

The change in IT buying requirements is the first of three storms that IBM saw as IT business leaders are building private clouds and experiencing scale issues associated with them. Data center buying criteria is changing as scale, density, deployment, orchestration management, efficiency and utilization, security, being able to extract meaningful decision support information out of information repositories, as well as cost of ownership become high priority items. The merchant silicon storm stirred up by companies such as Broadcom and Fulcrum Microsystems got IBM’s attention. IBM got a close up look at this storm, as BNT built its new ToR switches with Broadcom’s Trident-1 10GbE and 40GbE chips and decided to invest by acquisition. It was these two storms and its forecast of a third in the creation of a network software ecosystem emerging that in the end tipped IBM’s hand and led it into the data center system networking industry, or System Networking, as IBM now calls it.

The New IBM

IBM realized that not having System Networking was a competitive disadvantage especially in its analytical systems business. There was an underlying reliance on the network that IBM didn’t control. IBM realized that System Networking is a strategic asset, and it needed to invest.

IBM is now a three-stack business with its platform business including compute, storage and networking, then software and lastly, services. Software is the biggest business followed by services, and then its platform business. Without networking, IBM’s business model was incomplete. How can you drive innovation in software and smart analytics, etc., and all the services to go around it, if you have one or two missing pieces in the platform equation?

Others to Follow

IBM is not the only large vertical IT player to beef up its networking business. Clearly there is HP, Oracle, IBM, Dell and Cisco. Cisco possesses a different portfolio mix than the others with its dense networking portfolio. HP, on the other hand, possesses approximately $2.5 billion worth of networking products/revenue, but lacks data center networking.

Consider Oracle and IBM—they are both focused on the data center. With Oracle’s recent acquisition of Sun, it too is viewing the perfect storm as an opportunity to enter the networking market. But the fundamental thing that is different about IBM is that it is singularly focused on the data center. This contrasts with Cisco’s network focus while HP strives to be the low cost alternative to Cisco, plus its huge consumer line of products, such as printing and personal computing. Dell, on the other hand, is focused on transitioning away from the personal computing market into higher margin businesses, networking being one of them.

What all of these firms are searching for is a new networking model to emerge, and the perfect storm may very well provide it. With low cost merchant silicon that competes with custom ASICs, network switching is fast, low latency, low power consuming and low cost. With software defined networking (SDN), a new software ecosystem could emerge that challenges established network services and in the process, starts an innovation race between established vendors and a new software industry. SDN is critical if a new networking model is to emerge as it could enable innovation that differentiates common merchant silicon-based network switches. In short, the perfect storm could enable the large IT vendors to leapfrog into a new system networking paradigm.

IBM has its work cut out for itself. BNT has expanded from Ethernet embedded blade server switches to ToR switches. IBM will enter the aggregation space with the implementation of technologies such as TRILL (Transparent Interconnection of Lots of Links) and 802.1Qbg, the Edge Virtual Bridging (EVB) standard that will seek to break the model of large centralized mainframe like modular switches. And, through partners such as Brocade, Cisco, Juniper and Mellanox, IBM System Networking offers a portfolio of Fibre Channel and Infiniband as well as Ethernet solutions, for servers and storage from network edge to core. IBM’s point is that if servers and storage can scale out then why can’t networking?

IBM is developing new networking products that it hopes would enable it to change the networking landscape and how people think of networking. It seems that IBM System Networking is working on a scale out networking model that allows IT business leaders to start smaller and expand as needed without large upfront capital outlays. It is looking to make networking a bit smarter.

IBM System Networking is focused on building what it calls “a scalable fabric,” which connects servers, storage and networking. Thus IBM advocates to keeping network intelligence close to servers and storage making its fabric fast, low cost, virtual and reliable.

Time will tell how successful IBM is in System Networking, but one thing is for sure, cloud computing has kicked up quite a perfect storm for it.

Tweet

By Extreme Networks

This white paper is targeted to enterprise and service provider IT managers who may be evaluating the ability of iSCSI SANs to satisfy the performance requirements of their most demanding storage applications. Storage application performance is dependent on more than just the performance of the storage array. The full data path and the integration with the other layers of the data center architecture must be taken into consideration. Optimizing the full data path from the server, through the network to the storage can contribute to significant improvements in performance and service levels.

This white paper and the accompanying configuration guide details the work of Intel, Extreme Networks® and NetApp to demonstrate the impact on iSCSI performance in real-world environments, using Extreme Networks CLEAR-Flow technology to help you achieve optimal iSCSI performance.

The output of the collaborative testing is presented in two separate documents:

• White Paper: provides an overview of the latest advancements around iSCSI SAN capabilities and performance, including real world, end-to-end performance results with CLEAR-Flow.

• Configuration Guide: provides a sample reference architecture with step-by-step configuration details for all the components and describes the end-to-end solution performance across a number of configurations.

Get the White Paper

Tweet

By BLADE Network Technologies, an IBM Company

Private clouds describe an architecture in which servers and networks in the data center can rapidly respond to changing demands, by quickly scaling compute capacity and connecting that server capacity where it is needed. The technology underlying this fluidity is server virtualization, which by now has been established as a reliable and essential core technology for most data centers. Five critical networking technologies are needed to build a private cloud network:

• Up to 1.28Tbps of line-rate, loss-less bandwidth
• Single-wire networks for data and storage with DCB support
• Support for thousands of virtual ports
• Virtual machine awareness
• Extremely low power requirements

This white paper details a cloud network architecture that incorporates these five success factors.

Get the White Paper

Tweet

In the Lippis Report, we have discussed the fundamental changes shaping a new data center network architecture. These drivers are massive virtualization, a sea change in traffic patterns that are now dominated with east-west flows on top of existing north-south traffic, ultra low latency, the emergence of cloud spec data centers, etc. As a result, data center networking attributes are changing with requirements of traffic, steering in virtualized infrastructure, avoiding manual network changes as VMs move, removing oversubscription (thanks to spanning tree), streamlining network tiers to hasten east-west traffic flows, etc. The industry is responding to these changes and requirements with new approaches to data center networking, such as the Open Networking Foundation, Cisco’s FabricPath, Juniper’s QFabric, Brocade’s VCS, Avaya’s VENA, Nicira Networks’ network virtualization software, etc. In this Lippis Report Research Note, we explore a key technology to enabling two-tier network fabrics, and that’s link aggregation and its various approaches, including Multi-Chassis Link Aggregation Group, Transparent Interconnection of Lots of Links (TRILL) and Shortest Path Bridging (SPB).

Over the past year, firms such as BLADE Network Technologies, an IBM Company, Force10 Networks, Juniper Networks and Voltaire/Mellanox have introduced 48 port10GbE top-of-rack (ToR) switches. Before Interop in May, there will be six more companies making similar announcements. With 10GbE priced at $300 per port and below for server connections, the transition from 1 to 10GbE is on its way in the data center. Now most, if not all, of these switches possess two 40GbE uplinks. Also by Interop, at least two firms will announce Core switches with dense 40GbE capability. So the question is how are these ToR products being connected so as to address the changes mentioned above?

One Thousand Plus Servers Connected at 10GbE
Consider a 1,024-server data center where all servers are dual home connected into the fabric via 10GbE. This example could be a Global 2000 company data center, but many Global 2000 companies and service provider hosting companies have larger scale requirements in the tens of thousands of servers to over one hundred thousand. In this example, approximately 2,048 10GbE connections are needed. Consider this requirement using traditional approaches.

If designing this data center fabric with traditional spanning tree protocol (STP)-based networking, there would be blocked links between access and distribution. The IT architect would rely upon a three-tier structure that forces an oversubscription of nearly 8:1 between access and aggregation, and 2:1 between aggregation and core, or a total of 16:1 oversubscription. There would be 64 access switches, 8 aggregation switches and two Core switches required and four pods to house access and aggregation switches. In addition, east-west traffic flows are forced to traverse these network tiers, incurring delay with every passage of a switch.

To eliminate the oversubscription and reduce latency, a two-tier network architecture can be utilized. One approach is to use 43 of the new 10GbE ToR switches to connect servers. Connecting ToR switches would be some number of Core switches with enough capacity to support 512 40GbE or 2,048 10GbE connections, if non-blocking is a requirement. The Core switches would need to be connected together too at very high speeds and densities. Yet another approach would be to use Core switches to connect servers. Assuming a Core switch capable of supporting 256 10GbE ports, then eight Core switches would connect servers. Now, if the IT architect required non-blocking, then a Core switch would need to terminate 48 10GbE for each ToR switch or 256 10GbE links for each server facing Core switch. There lies the rub; with such large numbers of parallel 10 or 40 and eventually 100GbE links, there needs to be a way to aggregate and route between ToR and Core switches.

Enter link aggregation. The two-tier architecture allows the level of oversubscription and blocking to be designed and managed by choosing the number of links to be aggregated.

Key to this design is the elimination of STP with some number of multi-links between ToR and Core that eliminate oversubscription, and enable a non-blocking fabric, assuming the switches are designed with enough backplane capacity to support packet forwarding equal to the sum of leaf ingress bandwidth. High spine switch performance is fundamental in the two-tier leaf-spine architecture as it collapses the aggregation layer in the traditional three-tier network. Further, by connecting every switch together in a full mesh via link aggregation connections, every server is then one hop away from each other, reducing latency and providing VM mobility service.

There are multiple approaches for connecting ToR/leaf and Core/spine switches at high bandwidth via some type of link aggregation.

Multi-Chassis Link Aggregation Group or MC-LAG covered in project IEEE 802.3ad allows one or more links to be aggregated together to form a Link Aggregation Group. MC-LAG is a method of inverse multiplexing over multiple Ethernet links as if it were a single link. This layer 2 transparency is achieved by the LAG using a single MAC address for all the device’s ports in the LAG group. LAG can be configured as either static or dynamic. Dynamic LAG uses a peer-to-peer protocol for control, called the Link Aggregate Control Protocol LACP.

TRILL or Transparent Interconnection of Lots of Links is an emerging IETF protocol based upon a link state routing algorithm IS-IS that broadcast routes available to all TRILL connected devices for pair-wise optimal unicast paths. TRILL is invisible to routers as it runs over layer 2 links such as Ethernet and PPP.

Shortest Path Bridging or SPB is an IEEE 802.1aq standard solution for shortest path frame routing in multi-hop Ethernet networks with arbitrary topologies. SBP, like TRILL uses IS-IS link-state routing protocol to advertise both topology and logical network membership. SPB packets are encapsulated at the edge either in mac-in-mac 802.1ah or tagged 802.1Q/802.1ad frames and transported only to other members of the logical network. Unicast and multicast are supported, and all routing is on symmetric shortest paths.

MC-LAG vs. TRILL vs. SPB

As you would expect, there is debate over which approach is best, MC-LAG vs. TRILL vs. SPB. It doesn’t help that TRILL is an IETF standard, while SPB and MC-LAG are IEEE. Picking a winner is complex as there are pros and cons to each, and all protocols have their supporters in the vendor community. MC-LAG may be the most widely supported protocol but lacks link state routing to define paths. Some even question if you need IS-IS at this level of the network.

From an implementation point of view, many firms are betting on SPB, such as Brocade in its VCS, Avaya in the VENA, Alcatel-Lucent in its OmniSwitches. These firms like SPB for its following advantages:
SPB scales to support 100s to 1000s of multi Terra bit switching enabling a non-blocking two-tier network fabric;
SPB creates logical trees, which can be extended out of the data center and into the campus increasing SPB’s usefulness.

SPB service provider deployments are planned for 2011 and they believe SPB offers increased scalability over TRILL. Further, SPB will interoperate with carrier infrastructure to allow seamless data center-data center connections in the near future. This is an interesting and compelling option in that SPB could be the link that connects private and public clouds via a single data center fabric.

SPB advocates boast that for network architects/designers and operations, there is a quick learning curve as SPB uses the existing IS-IS protocol, and for service providers, SPB is already available through OAM (Operations, Administration and Maintenance), enabling it to be managed through existing management services.

Perhaps the biggest proponents of TRILL is IBM and Cisco, which has its FabricPath offering based upon it, and Data Center Bridging before that. Its proponents point to TRILL’s advantages of multi-pathing that delivers higher throughput between leaf and spine connections. TRILL too can be extended out of the data center into the campus and cloud as service providers offer TRILL connections. It’s also backward compatible with classic bridges, and was developed by Radia Perlman, the inventor of Spanning Tree Protocol.

Juniper’s QFabric

Then, in addition to the above, Juniper recently announced its QFabric architecture, which disaggregates the data, control and management planes. Its QFNodes are ToR switches, which are connected to its QF Interconnect chassis and managed via QF Director management platform. There are two separate connections for data and control plane traffic, with control traffic on a 10GbE link while data traffic runs at 40GbE. It’s not clear if the QFabric is a cell based data gram architecture, or if it uses Ethernet data grams. If QFabric is a cell based architecture, then it would not utilize TRILL, SPB or MC-LAG for inter-switch high-speed aggregated connections and routing.

Enter the Open Networking Foundation

Then, there’s Open Networking Foundation (ONF) started by Deutsche Telekom, Facebook, Google, Microsoft, Verizon and Yahoo with 17-member companies, including major equipment vendors, networking and virtualization software suppliers, and chip technology providers. ONF is proposing a new approach to data center networking it calls Software-Defined Networking (SDN).

SDN comprises of two basic components: a software interface (called OpenFlow) for controlling how packets are forwarded through network switches, and a set of global management interfaces upon which more advanced management tools can be built. The first task of ONF will be to adopt and then lead the ongoing development of the OpenFlow standard (www.openflow.org), and encourage its adoption by freely licensing it to all member companies. ONF will then begin the process of defining global management interfaces. The hope is that SDN will help networks become both more secure and more reliable.

Nicira Networks is a real world provider of SDN type solutions. Nicira proposes splitting control and data planes so that the data center network can be completely virtualized, like VMware did for servers. That is the right operational model for networking, where you treat the physical infrastructure as a generalized resource pool of switching capacity, and all of the services intelligence is done at edge in software, and the physical network does one thing and one thing only…forwards IP packets. 

MC-LAG, TRILL and SPB offer a linear approach to scaling data center networking while Juniper’s QFabric and ONF’s SDN offer new departures in the design and architecting of data center and cloud computing networking. While QFabric and SDN are interesting, they need to be developed and understood, but represent a new approach to networking that our industry has not seen. Over the next several years most IT architects will choose the linear approach as QFabric and SDN become fleshed out and their pros and cons articulated.

Tweet

By BLADE Networking Technologies, an IBM Company

Today, virtualization, cloud computing and high frequency trading place new demands on the system network fabric to deliver non-stop, ultra low-latency traffic flows. This traffic is increasingly “east-west” in nature to enable machine-to-machine communications versus the “north-south” traffic that characterizes conventional client/server and Web-based application environments. To deliver this east-west traffic using the most efficient flows, large flat networks are becoming increasingly popular. These flat Layer 2 networks eliminate extra hops to decrease latency, do not block any paths across the network, and are simple to configure. Such flat networks are built with large numbers of inexpensive top-of-rack switches, scale horizontally by simply adding more switches, and enable VLANs to span across a data center to provide larger server pools for virtualization. TRILL or TRansparent Interconnection of Lots of Links is fundamental to IBM’s approach and is explained in this white paper.

Get the White Paper

Tweet

The data center switching market is heating up. To address the scale issues posed by mobile and cloud computing nearly every network vendor is launching its own version of a 10/40/100 GbE fabric to connect servers and storage to the internet. At the heart of this fabric is a two-tier (Fat-Tree) network made up of leaf/ToR and spine/Core switches. Here leafs connect servers and spines connect leafs while also being interconnected in a logical mesh. The protocols to create this logical mesh are based upon IS-IS link state routing, but each vendor is taking a unique approach with Cisco using its FastPath, Alcatel-Lucent and Avaya using SPB (802.1aq Shortest Path Bridging) while Brocade VDX is based upon TRILL (Transparent Interconnection of Lots of Links). Juniper recently announced QFabric but has not detailed what it’s using for logical meshing. At the center of new data center design are leaf and spine switches. In Lippis Report Research Note 166, we detailed the latest ToR switches. In this Lippis Report Research Note 167, we dive into performance and power consumption measurements plus the use of SPB of Alcatel-Lucent’s OmniSwitch 10K, a new entry into spine/core data center switching market.

During December 6-10, 2010, the Lippis Report and Ixia conducted the industry’s first 10GbE data center switching evaluation of Top-of-Rack and Core Ethernet switches at the modern iSimCity lab in Santa Clara, CA. We evaluated Alcatel-Lucent’s OmniSwitch 10K, Arista’s 7504 Series Data Center Switch, BLADE Network Technologies’, an IBM Company, IBM BNT RackSwitch G8124 and IBM BNT RackSwitch G8264, Force10 Network’s S-Series S4810, Hitachi Cable’s Apresia 15000-64XL-PSR, Juniper Network’s EX Series EX8216 Ethernet Switch and Voltaire®’s Vantage™ 6048. We are conducting a second round of test scheduled for the week of April 4-8 at iSimCity, and it is open to all suppliers of 10 and 40 GbE data center switching.

There were three Core/Spine Switches evaluated for performance and power consumption in the Lippis/Ixia test. These participating vendors were:

Alcatel-Lucent OmniSwitch 10K
Arista 7504 Series Data Center Switch
Juniper Network EX Series EX8216 Ethernet Switch

These switches represent the state-of-the-art of computer network hardware and software engineering, and are central to private/public data center cloud computing infrastructure.

If not for this category of Ethernet switching, cloud computing would not exist. The Lippis/Ixia public test was the first evaluation for every Core switch tested. Each supplier’s Core switch was evaluated for its fundamental performance and power consumption features. The Lippis/Ixia test results demonstrate that these new Core switches provide state-of-the-art performance at efficient power consumption levels not seen before. The port density tested for these Core switches ranged from 128 10GbE ports to a high of 256 10GbE.

IT business leaders are responding favorably to Core switches equipped with a value proposition of high performance, high port density, competitive acquisition cost, virtualization aware services, high reliability and low power consumption. These Core switches currently are in high demand with quarterly revenues for mid-size firms in the $20 to $40M plus range. The combined market run rate for both ToR and Core 10GbE switching is measured in the multibillion-dollar range. Further, Core switch price points on a 10GbE per port basis are a low of $1,200 to a high of $6,093.

Their list price varies from $230,000 to $780,000 with an average order usually being in the million plus dollar range. While there is a large difference in list price as well as price per port between vendors, the reason is found in the number of network services supported by the various suppliers and 10GbE port density.

We compare each of the above firms in terms of their ability to forward packets: quickly (i.e., latency), without loss or their throughput at full line rate, when ports are oversubscribed with network traffic by 150%, in IP multicast mode and in cloud simulation. We also measure their power consumption.

Alcatel-Lucent launched its new entry into the enterprise data center market on December 17, 2010, with the OmniSwitch ™ 10K. The OmniSwitch was the most densely-populated device tested with 256 ports of 10GbE. The test numbers below represent the first public performance and power consumption measurements for the OmniSwitch™ 10K running software version 7.1.1.R01.1638. The Alcatel-Lucent OmniSwitch™ 10K Modular Ethernet LAN Chassis is the first of a new generation of network adaptable LAN switches. It exemplifies Alcatel-Lucent’s approach to enabling what it calls Application Fluent Networks, which are designed to deliver a high-quality user experience while optimizing the performance of legacy, real-time and multimedia applications. So how did the OmniSwitch 10K do?

RFC 2544 Layer 2 and 3 Latency Test

The OmniSwitch 10K was tested across all 256 ports of 10GbE. Its average latency ranged from a low of 20,561 ns or 20 μs to a high of 36,823 ns or 36 μs at jumbo size 9216 Byte size frames for layer 2 traffic. Its average delay variation ranged between 5 and 10 ns, providing consistent latency across all packet sizes at full line rate. What this means is that the OmniSwitch 10K can be counted on to forward packets at these latencies without much variation which is extremely important for predictable performance.

For layer 3 traffic, the OmniSwitch 10K’s measured average latency ranged from a low of 20,128 ns or 20μs at 64Bytes to a high of 45,933 ns or 45μs at jumbo size 9216 Byte size frames. Its average delay variation for layer 3 traffic ranged between 4 and 10 ns, providing consistent latency across all packet sizes at full line rate.

RFC 2544 Layer 2 and 3 Throughput Test

The OmniSwitch 10K demonstrated 100% throughput as a percentage of line rate across all 256 10GbE ports. In other words, not a single packet was dropped while the OmniSwitch 10K was presented with enough traffic to populate all of its 256 10GbE ports at line rate simultaneously for both L2 and L3 traffic flows. Not a single packet was dropped while 2.5 Tbps of traffic passed through its line cards and backplane.

RFC 2889 Congestion Test

The OmniSwitch 10K demonstrated nearly 80% of aggregated forwarding rate as percentage of line rate during congestion conditions. A single
10GbE port was flooded at 150% of line rate. The OmniSwitch did not use HOL blocking, which means that as the 10GbE port on the OmniSwitch became congested, it did not impact the performance of other ports. There was no back pressure detected as the Ixia test gear did not receive flow control frames. This was not the same for the Arista 7504. See the full test report here.

RFC 3918 IP Multicast

The OmniSwitch 10K demonstrated 100% aggregated throughput for IP multicast traffic with latencies ranging from 9,596 ns at 64 Byte size packets to 28,059 ns at 9216 Byte size packets. The OmniSwitch 10K demonstrated the lowest multicast latencies of all vendors.

Cloud Simulation Test

The one test that was not RFC based is a cloud simulation that was developed by the Lippis Report and Ixia. This test determines the traffic delivery performance of the DUT (device under test) in forwarding a variety of north-south and east-west traffic in cloud-computing applications. This test measures the throughput, latency, jitter and loss on a per application traffic type basis across M sets of 8-port topologies. The following traffic types are used: web (HTTP), database-server, server-database, iSCSI storage-server, iSCSI server-storage, client-server plus server-client. The north-south client-server traffic simulates Internet browsing; the database traffic simulates server-server lookup and data retrieval, while the storage traffic simulates IP-based storage requests and retrieval. When all traffic is transmitted, the throughput, latency, jitter and loss performance are measured on a per traffic type basis.

The OmniSwitch 10K performed extremely well under cloud simulation conditions by delivering 100% aggregated throughput while processing a large combination of east-west and north-south traffic flows. Zero packet loss was observed as its latency stayed under 28μs.

Power Consumption Test

The OmniSwitch 10K represents a new breed of cloud network spine switches with power efficiency being a core value. The OmniSwitch consumes 13.3 Watts/10GbE port with a TEER (Telecommunications Energy Efficiency Ratio) value of 71. TEER is a measure of network-element efficiency quantifying a network component’s ratio of “work performed” to energy consumed. Larger TEER values are better and the OmniSwitch is second only to Arista in TEER value while Juniper’s EX8216 measured a 44 TEER. You can download the OmniSwitch 10K test report here:

The OmniSwitch 10K power cost per 10GbE is estimated at $16.26 per year. The three-year cost to power the OmniSwitch is estimated at $12,485.46 and represents less than 3% of its list price. Keeping with data center best practices, its cooling fans flow air front to back, which is the norm except for Juniper’s EX8216 which pushes air from side to side unless a third party cabinet from vendors, such as Chatsworth, enclose the EX8216 to support hot-aisle and cold-aisle deployments.

Discussion:

The OmniSwitch™ 10K seeks to improve application performance and user experience with deep packet buffers, lossless virtual output queuing (VOQ) fabric and extensive traffic management capabilities. This architecture proved its value during the RFC2889 layer 2 and layer 3 congestion test with a 78% aggregated forwarding rate when a single 10GbE port was oversubscribed at 150% of line rate. The OmniSwitch™ 10K did not use HOL blocking, back pressure or signal back to the Ixia test equipment with Aggregated Flow Control Frames to slow down traffic flow. Not tested but notable features are its security and high availability design for uninterrupted uptime. The OmniSwitch™ 10K was found to have low power consumption, front-to-back cooling, front-accessible components and a compact form factor. The OmniSwitch™ 10K is designed to meet the requirements for mid- to large-sized enterprises data centers.

To demonstrate how the OmniSwitch™ 10K operates as a lossless fabric plus its ability to deliver carrier class quality of service (QoS), Alcatel-Lucent conducted two separate sets of tests; its data is available here. The lossless fabric test configured 256 x 10GbE ports connected with fully meshed traffic running at wire-speed via Ixia test equipment. The objective of this test was to demonstrate that as fabric and management modules were pulled and inserted into the OmniSwitch™ 10K chassis, zero loss at 100% load would result, and the fabric would be lossless. With fully meshed traffic running through all 256 GbE ports, the following modules were changed.

1. Fabric module was pulled out.
2. Fabric module was inserted back.
3. Management module (a fabric resides on this module) was pulled out.
4. Management module was inserted back.
5. Management module was pulled out causing a management failover in addition to fabric failover.
6. Management module was inserted back.

The result of the lossless fabric was that the fabric was lossless as the above modules were pulled and inserted.

The carrier class QoS objective was to demonstrate no packet loss at wire-speed with P0-P7 (priority) traffic running in fully meshed scenario as in the test above. The carrier class QoS test configured 256 x 10GbE ports connected with fully meshed traffic, priority 0 to 7, running at wire-speed via Ixia test equipment. In this scenario, the OmniSwitch™ 10K delivered zero loss with consistent store-and-forward average latency in range of 132,357 ns to 139,448 ns. See this test report details.

Cloud Network Architecture

There are three approaches to connect spine switches together to create a network fabric. MC-LAG (Multi-Chassis Link Aggregation Group), which allows one or more links to be aggregated together to form a Link Aggregation Group, TRILL and/or SPB are emerging standards that design a solution for shortest path frame routing in multi-hop Ethernet networks with arbitrary topologies, using an existing link-state routing protocol technology.

While there is debate over which approach is best, SPB has the following advantages. SPB deployments are planned for 2011 and offer increased scalability than TRILL. Further, SPB will interoperate with carrier infrastructure to allow private-public or private-private or public-public data center-to-data center connections. For network architects/designers and operations, there is a quick learning curve as SPB uses the existing IS-IS protocol, and for service providers, SPB is already available through OAM (Operations, Administration and Maintenance), enabling it to be managed through existing management services.

Paramount in the two-tier leaf-spine architecture is high-spine switch performance, which collapses the aggregation layer in the traditional three-tier network connecting spine switches together. The above captures the major trends and demands that IT business leaders are requiring from the networking industry. The underpinnings of private and public data center cloud network fabric are 10GbE switching with 40GbE and 100GbE ports/modules. 40GbE and 100GbE are in limited availability now but will be increasingly offered and adopted during 2011. Network performance including throughput performance and latency are fundamental switch attributes to understand and review across suppliers, because if the 10GbE switches an IT leader selects cannot scale performance to support increasing traffic volume plus shifts in traffic profile, not only will the network fail to be a fabric unable to support converge storage traffic, but business processes, application performance and user experience will suffer too.

During 2011, an increasing number of servers will be equipped with 10GbE LAN on Motherboard (LOM) driving 10GbE network requirements, and in 2012, high-end servers will be equipped with 40GbE LOM starting 40GbE’s growth curve. In addition, with nearly 80% of IT spend being consumed in data center infrastructure with all IT assets eventually running over 10GbE switching, the stakes could not be higher to select the right product upon which to build this fundamental corporate asset. Further, data center network equipment has the longest life span of all IT equipment; therefore, networking is a long-term investment and vendor commitment.

We review the Alcatel-Lucent OmniSwitch 10K from a perspective of performance and power measurement, mesh protocol support and key product features. Alcatel-Lucent has entered the data center switching market with a very competitive Core/spine switch. Clearly there are differences between Core switch vendors, and it’s advised to conduct a detailed review. For starters Click here for a copy of Alcatel-Lucent’s OmniSwitch 10K plus cross-vendor test results report.

Tweet

By Cisco Systems

Deploying network services in virtual data centers is extremely challenging. Traditionally, such Layer 4 through 7 services relied on intrusive, inline deployment and static network topologies. They were thus completely at odds with highly scalable virtual data center designs with mobile workloads, on-demand virtual machine (VM) provisioning, and strict service-level agreements (SLAs).

Cisco® Unified Network Services (UNS) addresses all of these problems by creating a framework for multiple services that can be configured and provisioned on demand, dynamically, to suit the service needs of enterprise applications and cloud users. This dramatically reduces network management overhead, allowing for a much more agile data center and business while providing improved application performance and a secure infrastructure. Cisco UNS comprises Cisco’s industry-leading solutions for virtual data centers that deliver.

● Load balancing and application controllers
● WAN acceleration
● Network security
● Network analysis and monitoring

Get the White Paper

Tweet

During December 6-10, 2010, the Lippis Report and Ixia conducted the industry’s first 10GbE data center switching evaluation of Top-of-Rack and Core Ethernet switches at the modern iSimCity lab in Santa Clara, CA. We evaluated Alcatel-Lucent’s OmniSwitch 10K, Arista’s 7504 Series Data Center Switch, BLADE Network Technologies’, an IBM Company, IBM BNT RackSwitch G8124 and IBM BNT RackSwitch G8264, Force10 Network’s S-Series S4810, Hitachi Cable’s Apresia 15000-64XL-PSR, Juniper Network’s EX Series EX8216 Ethernet Switch and Voltaire®’s Vantage™ 6048. We are conducting a second round of test scheduled for the week of April 4-8 at iSimCity, and it is open to all suppliers of 10GbE data center switching. We learned a lot about these products, both in the lab and out. In this Lippis Report Research Note, we dive into the Top-of-Rack 10GbE switches we tested as they represent a new generation of products that exhibit low power consumption, low latency, high performance and are all based upon new single chip designs from Broadcom, Marvell or Fulcrum Micro.

The Top-of-Rack (ToR) switches tested at iSimCity were the:

BLADE Network Technologies, an IBM Company, IBM BNT RackSwitch G8124 and IBM BNT RackSwitch G8264;
Force10 Network’s S-Series S4810;
Hitachi Cable’s Apresia 15000-64XL-PSR;
Voltaire®’s Vantage™ 6048.

All of these ToR switches utilize a new single chip design, but mostly from different silicon suppliers. With a single chip provided by chip manufacturers Broadcom, Marvell or Fulcrum Micro, vendors are free to invest resources other than ASIC development, which can consume much of a company’s engineering and financial resources. With merchant silicon providing a forwarding engine for their switches, these vendors are free to choose where to innovate, be it in buffer architecture, network services such as virtualization support, 40GbE uplink or fan-in support, etc.

The Lippis/Ixia test results demonstrate that these new chip designs provide state-of-the-art performance at efficient power consumption levels not seen before. In addition, price points on a 10GbE per port basis for ToR switches are a low of $351 to a high of $520.

IT business leaders are responding favorably to ToR switches equipped with a value proposition of high performance, low acquisition price and low power consumption. These ToR switches currently are the hot boxes in the industry, with quarterly revenues for mid-size firms in the $10 to $15M plus. We compared each of the above firms in terms of their ability to forward packets: quickly (i.e., latency) without loss or their throughput at full line rate, when ports are oversubscribed with network traffic by 150 percent, in IP multicast mode and in cloud simulation. We also measured their power consumption. Click Here for a copy of BLADE’s G8124 and G8264 plus cross-vendor test results report and Click Here for a copy of Force10’s S4810 plus cross-vendor specific report.

Latency Measurement Anomalies

When evaluating five products from four companies, there are bound to be anomalies. One anomaly was found during latency measurement. As both BLADE and Force 10 use the same Broadcom chip in their G8264 and S4810 ToR switches, respectively, one would expect their latency measurements would be close, but the S4810 showed lower latency values. As it turns out, the Broadcom chip allows switches to forward in cut-through and/or store-and-forward mode. The G8264 was configured and tested in cut-through mode while the S4810 and all other switches were configured and tested in store-and-forward. Test equipment, such as Ixia and others, measure latency very differently in these two forwarding modes.

During store-and-forward testing, test equipment subtract packet transmission latency, decreasing actual latency measurements by the time it takes to transmit a packet from input to output port. This makes comparisons between the two-latency measurement testing methodologies difficult. Also other potential device specific factors can impact latency too. But looking at the bigger picture, latency is being measured in the hundreds to thousands of nanoseconds across various packet sizes, making these switches the fastest forwarding engines in the market.

One of the biggest surprises was Voltaire’s Vantage 6048 ToR latency results, which were the highest of the group by nearly a factor of 2. Voltaire, now owned by Mellanox, used the Marvell 10GbE single chip code named Lion. The Hitachi Apresia 15000-64XL-PSR showed low latency results but it had other difficulties. For example, the largest frame size supported is 9044, excluding it from the 9216 byte size packet tests. Further, there is no latency data for the Apresia 15000-64XL-PSR at 64 bytes due to configuration difficulties during testing. The 15000-64XL-PSR could not be configured to maintain a VLAN at 64 bytes which eliminated packet signature to measure latency at this packet size.

A big surprise and delight found was how low the average delay variation was for all suppliers. Average delay variation was in the 5 to 10ns range, meaning that all of the above ToR switches deliver their latency results reliably.

Throughput

The results of RFC 2544 throughput testing should be boring with all ToR switches showing 100% throughput at line rate. The only anomaly here was the Apresia 15000-64XL-PSR during layer 2 forwarding, dropping packets at between 128 to 2176 packet sizes.

Congestion Testing

RFC 2889 congestion testing was telling too. Here the IBM BNT RackSwitch G8124 and IBM BNT RackSwitch G8264, Force10 Network’s S-Series S4810 and Voltaire®’s Vantage™ 6048 performed as expected, that is, offering 100% line rate under congestion conditions without head of line blocking and using back pressure or pause messages to control the flow of traffic. Here again, Hitachi Cable’s Apresia 15000-64XL-PSR showed head of line blocking and low throughput especially at the higher packet sizes of 2176 bytes.

IP Multicast

For RFC 3918 IP Multicast Throughput No Drop Rate testing, the IBM BNT RackSwitch G8124 and IBM BNT RackSwitch G8264 and Force10 Network’s S-Series S4810 performed flawlessly, exhibiting 100% line rate throughput and nanosecond latency with the G8124’s average latency 700ns and below. The IBM BNT RackSwitch G8264 and Force10 Network’s S-Series S4810 IP multicast performed as expected as they are both based upon the same Broadcom chip. The G8264 demonstrated a slight advantage of 100ns at the higher packet sizes while Force10 showed approximately 100ns advantage at the lower packet sizes. The Apresia 15000-64XL-PSR and Vantage 6048 do not support IP Multicast at this time.

Cloud Simulation

The one test that was not RFC based is a cloud simulation that was developed by the Lippis Report and Ixia. This test determines the traffic delivery performance of the DUT (device under test) in forwarding a variety of north-south and east-west traffic in cloud-computing applications. This test measures the throughput, latency, jitter and loss on a per application traffic type basis across M sets of 8-port topologies. The following traffic types are used: web (HTTP), database-server, server-database, iSCSI storage-server, iSCSI server-storage, client-server plus server-client. The north-south client-server traffic simulates Internet browsing, the database traffic simulates server-server lookup and data retrieval, while the storage traffic simulates IP-based storage requests and retrieval. When all traffic is transmitted, the throughput, latency, jitter and loss performance are measured on a per traffic type basis.

This test is telling too as it’s designed to be a simulation of real-world cloud-computing traffic. The results here show that the IBM BNT RackSwitch G8124 and G8264 delivered the lowest latency consistently across all protocol types. The Apresia 15000-64XL-PSR performed very well in this test too, followed by Force10’s S4810 followed by Voltaire’s Vantage 6048. Anomalistically, both the Force10 S4810 and Vantage 6048 spiked in terms of latency for east-west database-server, HTTP and iSCSI-Storage traffic flows. Both IBM BNT RackSwitches and Force10’s S4810 tested in cut-through mode.

Power Consumption

Power consumption or energy efficiency has become a paramount concern in data centers as the cost of power and cooling start to dominate TCO (total cost of ownership) over a three-year period. The ToR switches tested offer the lowest power consumption of switching products evaluated in public industry test. Their power consumption measured in WATTS per 10GbE via ATIS methodology ranged from 3.6 to 5.5. We then projected annual cost per 10GbE to be between $4.36 to $6.70, with the Apresia 15000-64XL-PSR offering the lowest power consumption. The IBM BNT RackSwitch G8264 and Force10’s S4810 were very close at $4.78 and $4.91, respectively, with the G8264 having a slight advantage. Of the 48-port 10GbE ToR switches, Voltaire’s Vantage 6048 consumed the most energy at 5.5Watts/10GbE.

While not confirmed, the IBM BNT RackSwitch G8124 may be based upon the Fulcrum single chip set, code named Bali, as well as Arista’s 7124 and Force10’s S2410. The Apresia 15000-64XL-PSR may be based upon the Broadcom Trident single chip. There are rumors in the industry too that large networking firms may start to utilize merchant silicon rather than build their own, as these chips offer a quicker path to market and are delivering solid performance, latency and power efficiency results.

While I detail ten recommendations in the test report, here I’ll focus on one. 10GbE ToR switches are ready for mass deployment, delivering full line rate
throughput at zero packet loss and nanosecond latency plus single- to double-digit delay variation. In addition, these ToR switches offer low power consumption with energy cost over a three-year period estimated between 3 and 4% of acquisition cost. Clearly there are differences between vendors, and it’s advised to conduct a detailed review. For starters Click here for a copy of BLADE’s G8124 and G8264 plus cross-vendor test results report and Click Here for a copy of Force10’s S4810 plus cross-vendor specific report.

Tweet

By Ken Won, Director of Product Marketing at Force10 Networks

Server and storage environments have seen a lot of changes in the past ten years, while developments in networking have remained fairly static. Now, the demands of virtualization and network convergence are driving significant changes in the data center network. Networks have always been considered as plumbing that connect servers and storage, but new, dynamic switches are changing the network’s role in the overall data center. It’s not your father’s network anymore, and savvy data center managers need to understand and plan for the changes that are coming.

This white paper discusses new network technologies, explains what they are, and suggests how to plan for them in future data center architectures.

Get the White Paper

Tweet

By Arista Networks

While servers and applications have fully embraced the concepts of automation, sadly the network infrastructure, on which they all rely, is still mired in legacy technologies. Current methodology, such as CLI, requires extensive hands-on provisioning and configuration by knowledgeable personnel. In modern cloud infrastructure, network managers must be able to centralize provisioning and configuration roles to improve reliability, minimize bring-up costs, and contain the expenses of creating a cloud data center service. Whether you are looking to maximize the efficiency and reliability of your existing operations, or you are looking to take advantage of a cloud-based infrastructure, fully automated provisioning is an essential capability, and Arista’s Zero Touch provisioning offers the first approach to automated network configuration.

Get the White Paper

Tweet

by Lippis Report Testing

This report details the test results of the Force10 Networks S-Series S4810 enterprise Top-of-Rack (ToR) Ethernet switch. All 48 10GbE ports of the Force10 Networks S-Series S4810 are populated and tested for performance, throughput, latency and power consumption. The Force10 Networks S-Series S4810 test results are then compared against IBM BNT RackSwitch G8264, Voltaire Vantage™ 6048 and Apresia 15K ToR switches.

To download the test report please CLICK HERE

Visit the Link

Tweet

by Lippis Report Testing

This report details the test results of the IBM BNT RackSwitch G8124 & G8264 enterprise Top-of-Rack (ToR) Ethernet switches. All 24 and 64 10GbE ports of the IBM BNT RackSwitch G8124 & G8264, respectively, are populated and tested for performance, throughput, latency and power consumption. The IBM BNT RackSwitch G8124 & G8264’s test results are then compared against Force10 Networks S4810, Voltaire Vantage™ 6048 and Apresia 15K ToR switches.

To download the test report please CLICK HERE

Visit the Link

Tweet

by Lippis Report Testing

This report details the test results of the Alcatel-Lucent OmniSwitch™ 10K enterprise core Ethernet switch. All 256 10GbE ports are populated and tested for performance, throughput, latency and power consumption. The OmniSwitch™ 10K’s test results are then compared against Juniper’s EX8216 and Arista Networks 7504 core switches.

To download the test report, please click here

Visit the Link

Tweet

A cross vendor comparative test report conducted at Ixia’s iSmiCity defined by The Lippis Report. This report provides detailed test information on the following new products that have not been previously tested in public. The report details test results of the following products:

Alcatel-Lucent OmniSwitch 10K,
Arista 7504 Series Data Center Switch,
Arista 7124SX 10G SFP Data Center Switch,
Arista 7050S-64 10/40G Data Center Switch,
Brocade VDXTM 6720-24 Data Center Switch,
IBM BNT RackSwitch G8124,
IBM BNT RackSwitch G8264,
Force10 S-Series S4810,
Hitachi Cable, Apresia15000-64XL-PSR,
Juniper Network EX Series EX8216 Ethernet Switch,
Voltaire® VantageTM 6048.

This 93-page report is a must for those evaluating 10/40 GbE data center switching equipment for private or public cloud infrastructure. You don’t want to buy data center switching gear until you read this report.

To download the report click here.

Get the White Paper

Tweet

Any IT business leader knows that the single most important technology driving data center design change is server virtualization to the point that a virtual machine (VM) is now the data center building block. As server virtualization marches on until nearly every physical server has been virtualized, networking in a virtualized environment is being forced to fundamentally change too. By networking, I mean not only layer 2 and 3 forwarding but network services too, such as application controllers, WAN optimizes, firewalls, etc., which are fundamental for mission critical application performance, cost reduction and high application availability especially where service level agreements are required.

Adding new applications to a data center has become highly complex, thanks to all the routing paths that need to be set-up to provide connectivity and reach of network services plus the configuration and policy set-up for network services specific to the application. Then, once the application is operational, it’s hard to virtualize it and move it via v-motion, et al, while keeping set-up and policies intact, especially routing paths. The current state of rigid networking consumes time and cost, but most importantly limiting the speed and agility in which new applications can be delivered and businesses react to market dynamics. This is a nasty problem, riddled with complexity and associated cross-administrative operational cost limiting the number of applications that can be virtualized until this problem is solved.

An entirely new approach to deploying, provisioning and managing data center network services in a virtualized environment is needed, and Cisco is addressing this need with its Unified Network Services or UNS. Cisco’s UNS is not just a suite of its layer 4-7 network service offerings such as ACE, WAAS, etc., but a framework for transparently inserting network services into a virtual server environment for steering traffic to network services on a per-VM basis plus an extensible and integrated policy management architecture. The key word in UNS is “unified,” as UNS makes network services available to both physical and virtual servers and their associated applications via steering traffic to network services hosted in appliances/modules/blades or within a VM. UNS promises to help reduce the costs to deploy new applications plus to enable more applications to be virtualized. In short, UNS offers an approach to deploy, provision and manage new applications without the network set-up complexity mentioned above. In addition, it also promises to remove network complexity associated with virtualizing applications and their moves. UNS is a main pillar of Cisco’s Data Center Business Advantage architecture, along with Cisco’s Unified Fabric and Unified Computing Services. These pillars combine to form the tightly-integrated next generation data center components including the network, storage, application services, virtualization layers and network services.

Cisco’s UNS is addressing mobile (v-motion) applications and their associated changing or dynamic network topology requirements by steering traffic to appropriate network services that are centrally controlled via policy. These network services such as firewalls, application controllers, WAN acceleration, load balancing, etc., can be packaged in appliances, modules, server blades and/or other form factors and/or increasingly as a virtualized service. UNS is a modern approach to applying layer 4-7 network services to both non-virtualized applications and VMs, while in the process solving some of the most complex problems associated with virtualized infrastructure.

Dedicated Hardware Services to Virtualized Network Services

Traditional network services are frequently placed in-line or in the flow of traffic, that is firewall, IPS, load balancing, application controllers, WAN acceleration, etc., forming a line of layer 4-7 network services. But as applications are virtualized, their movement may take them out of the path of traffic flow, thus creating difficulty to maintain network services to VMs and their applications. In most data centers, a mix of physical and virtual network services is emerging as well as a mix of virtual servers and physical servers based upon old and new investment. What IT business leaders demand is that their investment in physical and/or virtual network services support both virtualized and non-virtualized applications so they may extract the highest value from their IT dollars. This is a hard problem to solve and requires new thinking in networking which is what UNS is focused upon delivering. In short, UNS allows a mix and matching of physical and virtual network services to support either virtualized or non-virtualized applications through a more flexible approach to networking and policy management. So how do IT architects create this level of flexibility?

In a UNS environment, the physical placement of network services in appliance/modules/server blades, etc., or virtualized form is moot, offering IT architects a new degree of freedom to access these services anywhere in a virtualized infrastructure. A network service can be offered to a VM and its associated traffic, independent upon its form factor, be it a physical appliance, dedicated module or virtualized network service as long as the VM and softswitch send traffic to the appropriate service as the application moves around the data center.

That’s important as traffic patterns have shifted from primarily north-south to a mix of east-west and north-south, resulting in the need for network services to offer far greater flexibility in their reach to service VMs and the applications they contain. And as network services are logically wrapped around a VM via policy, they receive the benefit of all moving together, solving one of the biggest virtualization problems in the industry, manually intensive change management. Parallel to making network services accessible independent upon location and its packaging is the added benefit of virtualizing network services as this will decrease the number of hardware appliances in a data center, reducing complexity, total cost of ownership and energy consumption.

Unified Network Services Is a Platform for Inter-Cloud Mobility and On-Demand Provisioning

But perhaps even more important than solving the immediate change management problem is that unified network services deliver a set of attributes that put in place the tools and ability to deliver elastic IT services between clouds—the holy grail of cloud computing. With core network services unified, a degree of flexibility is gained far beyond current technology and offers a platform in which service advertising and registry can occur so that a “provision proxy” can automate network service configuration to meet new IT service delivery needs in near real time; but this is a topic for another day. The important point is that a unified network service is a platform that all large IT firms, cloud providers and enterprises will be investing in over the next business cycle.

Cisco’s Unified Network Services or UNS

In this Research Note, we review Cisco’s UNS, the most comprehensive approach to data center and cloud network service deployments in the industry thus far. UNS addresses the on-demand provisioning problem so sought after in virtualized infrastructure. That is when IT leaders need to allocate resources from within or between a private or public cloud on demand and quickly, UNS will respond to a capacity request so that network services are provisioned in the right order, at the right capabilities and within minutes rather than months. In short, UNS’s vision is to enable on-demand network service delivery and on-demand provisioning to accommodate VM container workload mobility within the construct of an Enterprise’s IT model or service architecture.

The Virtual Security Gateway

UNS is both a vision of on-demand service provisioning and the products that enable its construct. Within UNS, Cisco has launched its data center firewall called VSG or Virtual Security Gateway, and is on a path of virtualizing its data center service products including the Wide Area Application Services or WAAS, et al, and providing them with consistent policies via its VNMC or Virtualized Network Management Control software. VSG is an example of a virtual service node, as compared to physical ASA security appliances. The key underpinning technology to VSG is the Nexus 1000v and vPATH, which enable traffic to be re-routed or steered to the virtual firewall nodes; more on this below.

Cisco’s VSG offers a model of how network services are virtualized and in the process, solves some of the biggest server virtualization problems while delivering added flexibility value. VSG is a proof-point of Cisco’s ability to solve the firewall problem within virtualized infrastructure; that is how to provide firewall services to flows destined to and between various VMs. vPATH, a software module within the Nexus 1000v softswitch, steers traffic to VSG, the firewall, which blocks or allows traffic flow to its destination. Further, VSG assures that the correct network security service is applied and a VM’s policies follow it as it moves between physical servers. VSG policy is centrally managed through the VNMC umbrella management platform.

Central to UNS is vPATH technology that confers the same VSG benefits discussed above to Cisco’s new Virtual WAAS or vWAAS WAN acceleration offering. vPATH is fundamental to UNS as it delivers unification by being the same underlying infrastructure for both VSG and vWAAS. Therefore, by inserting vPATH technology/software into the virtual switch, hypervisors and VM’s traffic is re-directed as needed to deliver network services, such as firewall, WAN acceleration, etc.

vPATH

In the case of VSG, through VNMC, policy is created to define what type of traffic needs to be redirected, and then what action to take upon that traffic once it arrives at the firewall. As traffic reaches a server or Nexus 1000v, it is intercepted as it’s destined for a particular VM by vPATH, which redirects it to VSG for inspection. VSG then performs its network security service then forwards the traffic, if allowed, to its destination just like a firewall appliance operates.

The closest analogy to describe vPATH’s function is network-based application recognition. That is NBAR analyzes traffic and classifies it, and then performs a function such as prioritization. Thus, vPATH intercepts traffic and sends it to VSG while VSG performs its security service and decides if traffic will be forwarded to the destination VM.

Fast Path

vPATH also benefits from a concept called fast path. Fast path is similar to a cut-through method in that once traffic has been forwarded to VSG for firewall services, for example, the remaining traffic flow, it’s routed directly to its VM destination. Note that fast path can be utilized for most network services. Fast path obviates the need to route all traffic through VSG once the first packet of the flow has been processed by the firewall. Therefore, all traffic does not require packet-by-packet inspection, speeding up flows and reducing processing and latency.

For example, if the first packet of a flow passes through VSG without alteration, then the rest of the flow should pass uninspected as the security rules are the same. However, this wouldn’t be the case for an IPS system, where the entire payload is inspected to assure there is no malware residing in the flow. Fast path will evolve to support various traffic scenarios too.

Network Service Chaining

Cisco’s UNS provides a solution to the challenge of providing network services to traffic flows within a virtualized infrastructure that stick to VMs as they move and change physical location in the data center. The next challenge is to provide virtualized network service chaining. Chaining network services is the ability to create a single policy for traffic flows as it ingresses to a VM for multiple network services. For example, a policy may apply firewall, load-balancing, WAN-optimization, etc., to a flow and route that traffic through subsequent services, as opposed to having to create unique policies, intercept each one and route traffic accordingly. Chaining is a huge operational time saver, and it hastens the flow of traffic within the data center. vPATH is one underlying mechanism that can steer traffic to services in the right chain/order.

The UNS Value Proposition

From a data center network design perspective, UNS is developing a set of network service building blocks that brings physical network service appliances and virtual service nodes into virtualized environments along with the tools to apply policies to govern their use. As more and more data centers become virtualized so too will network services. In addition, as physical and virtual data centers will co-exist for many years to come, the ability to offload physical network appliances with virtualized ones as well as pass traffic between them offers a transition path and a means to extend the life of existing appliance investments.

As mentioned above, physical data centers are equipped with stacks of appliances offering load balancing, WAN acceleration, firewalls, IPS, etc. Now with service chaining and vPATH, all of these physical and virtualized appliances can be put to work servicing VMs and their applications. Most importantly though is that UNS offers a way to control network services so that VMs, virtual applications and mobile workloads can be scaled up and down plus moved within a dynamic network that allows provisioning services easily. For all intents and purposes, the industry has not had a multi-service chaining mechanism in the physical world. IT operations have done this manually via provisioning VLANs, policy routing, Web Cache Communications Protocol or WCCP, etc. But the old approach is static, and when servers, applications, appliances, etc., move or change, manual intervention is required. The beauty is that chaining network services in a virtualized infrastructure enables elastic scale-up and scale-down much more seamlessly.

Why Unify Network Services

One of the key strategic elements behind UNS is to change the mindset in which IT leaders deploy network services. Traditionally network service appliances were deployed at the edge of the data center or in front of a specific application server. But servers and application are often moved creating the manual re-configuration problem discussed above. Having common accessible network services in private and public data center clouds could offer huge provisioning benefits. For example, there could be, potentially, a vWAAS instantiation in Amazon EC2, Rackspace, GoGrid, etc, which IT leaders who have deployed WAAS in their branch offices could leverage, meaning their WAN would be accelerated thanks to a common WAAS image in the branch and cloud providing that network service independent upon these two application deployment models. This new network services deployment model attempts to blend the worlds of Cisco’s borderless and data center initiatives to the fullest extent.

What’s the intrinsic value of making a network service virtualization? In the case of vWAAS, Cisco is able to give IT leaders flexibility of placement and IT delivery. vWAAS is easier to scale up, licensed in a “pay as you grow” model, offers fewer devices to manage with less power and cooling cost plus is overall more flexible in its placement. In addition, vWAAS and WAAS can both offer WAN acceleration services to virtualized applications thanks to vPATH increasing the usefulness and value to both. vWAAS may be deployed by cloud providers too, which could offer IT leaders a WAN acceleration option independent upon application hosting.

Distributed Deployment with Centralized Management

Value is gained by being able to deploy network services in a distributed fashion, thanks to UNS. UNS changes network service deployment from a centralized model to distributed. But while virtualized network services are distributed, its management is centralized, offering operational efficiency and deployment flexibility. Distributed network service deployment with centralized management is the only approach that works as virtualized network services tend to be distributed widely. In fact, large data centers and clouds will see their instantiations of a particular service grown from a few hundred to thousands, if not more. Therefore, centralized management of virtualized network services provide the control knobs to provision, develop policy, steer traffic, etc., for thousands of virtualized network services distributed throughout a virtualized infrastructure. For example, in Cisco’s UNS, vWAAS and VSG run in their own VM, either on a single physical server or multiple physical servers, offering a highly distributed network service option.

Other companies, such as A10 and at least five others, are virtualizing their application delivery offering too. And cloud service providers are seeking virtualized network services, which will offer IT business leaders the ability to deploy applications from either private or public clouds with a common set of network services over time. For example, many public cloud providers would like to place load-balancing services on top-of-rack and deploy it in a small-medium-large type format. Further, many would also like to place load-balancing services on a compute platform to give customers the ability to deploy load-balancing pseudo-traditionally. That is to deploy network services where a compute platform would be largely dedicated to that service, or, alternatively, distributed so that it does not necessarily reside top-of-rack, or centralized, but resides “logically” next to a VM or sets of VMs so that as VMs move the network service benefit followings.

UNS: A Product Set or Next Evolution of Networking and Computer Services

Now Cisco isn’t the only IT firm developing a unified network service framework, but it is the only company that has all the components to deliver a comprehensive and thoughtful solution. For example, HP, IBM and Oracle do not develop load balancing, application delivery, WAN acceleration or softswitch network services, placing them at a disadvantage. Oracle, HP and IBM usually partner with others for these services such as F5, Riverbed, VMWare, etc., eliminating the opportunity for this level of virtualization and unification development. In HP’s case, its networking gear is increasingly made in China which lacks the forward-looking foresight to get in front of this opportunity. IBM usually does a really good job here, but it’s limited on these major network service components.

Many of the niche players, such as F5, Riverbed, Infoblox, A10, et al, will and are virtualizing their network service appliances and will do it very well, emerging as feature functional leaders. But these firms’ virtualization strategies will lack the broad view of multiple network services and most importantly, how the network nodes (L2-3 infrastructure) or hypervisor can steer traffic to them. To gain a broader UNS view and solution, these firms could organize a consortium to develop a comprehensive UNS strategy and implementation that matches Cisco’s UNS. But consortium is driven by committee, which usually moves slowly. Cisco’s UNS framework will be emulated by others while key technology layers can be standardized, such as Cisco’s proposed VN-Link for traffic steering to physical devices from a virtual/softswitch. Hopefully, an ecosystem can be created that allows all vendors to participate, because UNS is not just another vision and product line, but it’s the next evolution of networking and computing services.

Tweet

By Force10 Networks

Deriving a return on virtualization investments means deploying effective automation techniques that can simplify the virtualized environment while allowing a policy-based deployment model. While many network vendors have chosen a proprietary path to automate network changes, Force10′s approach is to utilize open and industry standard technologies based on the extensible and modular Force10 Operating System (FTOS), across a range of the heterogeneous Force10 switch and routers portfolio.

Find out how to automate network changes in virtualized infrastructure by downloading this white paper

Get the White Paper

Tweet

By Extreme Networks

Server virtualization brings with it a set of network operational challenges: from configuration challenges around Virtual Machine (VM) switching to managing virtual machine mobility, to providing VM location and inventory in the network. There are few tools available to the network administrator that provides visibility, control and insight into the VM environment until now. Extreme Networks® XNVTM provides network-level visibility and control of the server VM environment in a hypervisor-agnostic manner and without requiring any changes to the server virtualization operating environment.

Find out how by downloading this white paper:

Get the White Paper

Tweet

A typical “non-virtualized” data center has three network layers, Top-of-Rack, End-of-Row and Core switches. But virtualized infrastructure adds two additional layers—the virtual switch and blade switch—raising the number of tiers from 3 to 5. This significantly increases latency plus the number of network elements within the data center resulting in increased data center management complexity. I talk with Shehzad Merchant, Senior Director for Strategy at Extreme Networks, about Extreme’s flattening approach to data center network fabric through its DirectAttach.

Listen to the Podcast

Tweet

Andy Bechtolsheim, Founder, Chief Development Officer and Chairman of Arista Networks, joins me to discuss how the public and private data center cloud network market is emerging and the new type of networking it’s requiring.

Listen to the Podcast

Tweet

Ethernet networking is now the single most important data center technology to assure the new IT economic model of centralized application delivery. Yes that’s right—Ethernet as the data center fabric is the stability point in data center design that will dictate if a data center or cloud facility can scale to support huge application and storage traffic loads. And if you think that Ethernet switch performance is not important then you would be as right as the engineers who designed the Tacoma Narrows Bridge. In this Lippis Report Research Note, we explain why network performance of data center Ethernet switching products matter more now than ever.

Data centers are becoming IT black holes where no application can escape the gravity of its economic force. A few facts are in order:

Mobile Applications and Devices Soar: Mobile application use is expanding exponentially, thanks to the popularity of the iPhone and increasingly Android smartphones. Most important about this is the traffic load these applications are placing on data center Ethernet fabrics. The vast majority of mobile applications are hosted in data centers and/or public cloud facilities. The application model of mobile devices is not to load them up with thick applications like Microsoft Word, PowerPoint, Excel, etc, but to load them with thin clients that access their application and data in data centers, private and/or public cloud facilities. As of this writing, there are some 205,000 plus smartphone applications.

A New Tier of Computing Emerges: A new and rapidly growing tier of computing has emerged in 2010. This tier is the Android tablet and iPad. According to the Wall Street Journal, sales of tablet devices (Android plus iPad) are expected to hit 19.5 million units in 2010 and 54.8 million in 2011. In contrast, Garter predicts that PC shipments will be 352 million units in 2010. In just six short months, tablets now represent some 6% of PC shipments and are expected to displace nearly 10% of PC shipments by 2014!

What is important about this new tier of computing is its application model, which is nearly the same as smartphones. That is these 10s of millions and growing numbers of tables are relying on data centers plus private/public cloud facilities for their applications placing further traffic load on Ethernet data center fabrics.

Virtualized Desktops: 2011 will be the year of the virtualized desktop. Frustrated with Microsoft’s enterprise application licensing, plus desktop support model, IT business leaders will turn toward virtualizing desktops at increasing numbers in 2011. The application model of virtualized desktops is to deliver a wide range of corporate applications hosted in data centers and/or private/public clouds over the enterprise network. While there are no estimates to the traffic load this will place on campus and data center Ethernet networking, one can only assume it will be huge.

Storage Traffic over Ethernet Fabric: Converged I/O or unified networking where storage and network traffic flow over a single Ethernet network will increasingly be adopted in 2011. A single converged network adaptor or CNA plugged into a server provides the conduit for storage and application traffic flows to traverse over an Ethernet fabric. The number of suppliers offering CNAs has grown significantly, including Intel, HP, Emulex, IBM, ServierEngines, QLogic, Cisco, Brocade, etc. In addition, the IEEE opened up the door for mass deployment as it has ratified the key Ethernet standards for lossless Ethernet. What will drive converged I/O is the reduced cost of cabling, NIC and switching hardware.

The above trends are just starting to take hold. Over the next five years, a sea change in IT delivery will occur. It’s clear that the number of mobile smartphones and tablets will only increase as will their reliance data center hosted applications. Virtualized desktops too will force an increase in centralized application delivery while storage traffic increasingly flows over Ethernet fabrics. Corporate application portfolios will change dramatically as will their application traffic profiles with loads being ever more unpredictable. There will be surprises or unforeseen changes that may very well accelerate these trends.

From a data center design point of view, IT architects discovered over three years ago that they can scale compute resources to nearly unlimited dimension thanks to multi-core processors, virtualization and cloud spec design. And with centralization comes huge corporate advantage that being centralized complexity to manage IT more effectively. But more importantly is the fact that IT represents on average only 2% of corporate revenue but has a profound impact on the other 98% of corporate operational spend and competitiveness. With application centralization, IT business leaders can more easily control IT and target it toward reducing corporate operational spend through streamlined business processes or launch new services to respond to market dynamics.

At the center of this massive application centralization transition is networking as it ties compute, storage and internet access together. Ethernet networking, in particular, is now the single most important data center technology to assure the new IT economic model of centralized application delivery. Now most corporations and cloud providers are scaling up their data center bandwidth with 10GbE. In fact, over the last quarter, many networking companies have reported greater than 60% shipment growth in their layer 2 and layer 3 fixed and modular Ethernet switches. So the above trends are driving network demand.

But IT architects and business decision makers need to understand the underlying performance and power consumption metrics of the switches they deploy. The only way to be assured that the Ethernet fabric that is being deployed now in the data center will scale to support increasing application load and storage traffic is to review public, independent, credible and repeatable network throughput and latency performance numbers across multiple vendors.

During the mid 1990s, Scott Bradner of Harvard University and Nick Lippis of the Lippis Report offered independent comparative Ethernet switch performance test evaluations to guide IT business leaders with their purchase decisions. But network purchase decisions have much greater weight to them now as over 80% of IT budgets are spent in the data center. Further, HP wouldn’t have purchased 3Com or IBM wouldn’t have purchased BLADE if they didn’t realize how critically important networking has become to successful data center and cloud computing design.

It’s for the above reasons the Lippis Report has teamed with Ixia to deliver an open data center fabric evaluation of 10GbE switches. Several network equipment manufacturers will participate in this industry-first evaluation, including Alcatel-Lucent, Apresia, Arista, Blade, Juniper Networks and Voltaire. The testing, which is taking place at Ixia’s iSimCity location in Santa Clara, will use Ixia’s Xcellon-Flex load modules to evaluate the performance of the participating vendors’ top-of-the-line 10 GE data center devices.

We’ll publish a report on our findings in mid January so stay tuned.

Tweet

For as long as I have been following Avaya—and it’s been a decade since it was spun out of Lucent back in October of 2000—it has undergone three fundamental transitions. First, Don Peterson, Avaya’s first CEO, managed to fix Avaya’s balance sheet after Lucent saddled it with heavy debt. He also pointed the way toward IP telephony in his six years at the helm. Then came Louis D’Ambrosio, with high energy and confidence, to point Avaya in the direction of unified communications, and a software and services business model, while bringing the company private in 2007 through TPG Capital and Silver Lake Partners. In 2008, Charlie Giancarlo became chairman, while Kevin Kennedy took the helm, ushering in a new wave of innovation and nimbleness while re-engineering sales and channels plus absorbing the Nortel enterprise business. Yes, what a long, strange trip it’s been, but Avaya is now the most innovative in its history and well positioned for the post-recession business cycle. In this Lippis Report Research Note, we examine Avaya’s prospects and challenges.

If Peterson’s contribution to Avaya was “Righting the Ship,” and D’Ambrosio’s was “Energy and Purpose,” Kennedy is ushering in “Nimbleness and Innovation.” With each phase of executive leadership came a resetting of corporate culture. Peterson and the executive management team nearly all had AT&T/Lucent culture where the enterprise business was a rounding error. D’Ambrosio brought a customer focus, energy and big blue reliability. Both of these cultures were grounded in East Coast high tech. Kennedy reset the culture button to a West Coast pace of “go, go, go” with phased product roadmaps, advanced technologies and broad channels to market. The new Avaya has taken shape with a slew of product announcements and new technologies. Here are a few of its most novel directions:

The Flare Experience: In September of 2010, Avaya introduced the Flare Experience, which is a new human-machine metaphor to easily conduct videoconferences and collaboration. Flare seeks to provide a seamless video experience from desktop to softphone to video conferencing systems to android tablets, etc. The most notable aspect of Avaya Flare™ is the introduction of the Avaya Desktop Video Device, an android tablet that creates video sessions with the ease of dragging and dropping contacts from an address book to the center of the screen via touch screen technology. Key to Flare’s innovation is the linking of presence, directory and call establishment/tear down between Avaya one-x Communicator 6.0, the Avaya Video Conferencing solutions based upon joint development work with LifeSize, the Avaya Desktop Video Device, Avaya Video professional and managed services as well as Avaya’s web.live. But it’s the Avaya Aura collaboration server in the back end providing the magic code to create an enterprise wide video experience.

What’s impressive about Flare is that Avaya has created a user interface that integrates voice, video, web conferencing, IM, presence, email, contacts, calendar, messaging, browsing, business applications and social networking that’s controlled by touch. Desktop or user interface design is usually not offered by communications companies, other than phones, so this is a significant innovation point for Avaya.

The Skype Relationship: With Skype and Avaya being owned in part by Silver Lake Partners, a friendly business channel was easily created. For years, most industry observers and IT business leaders sort a way to integrate Skype calls into enterprise communications and collaboration. Avaya was the first to do so by granting access to U.S. customers to Skype Connect™, from their existing UC endpoint via a SIP connection between Avaya Aura and Skype. The Avaya-Skype link becomes more feature rich in the second half of 2011 when federation is established so that Avaya and Skype business users can engage and interact via presence, IM plus voice and video. Beyond the cool factor, there are hard economic reasons why a Skype connection makes sense for Avaya customers. There are three value points…those being low international calling rates, access to Skype’s global community and inter-company collaboration via modern communications.

The SME Roadmap: To show how nimble Avaya has become, when it acquired the Nortel enterprise business in late 2009, it had six products focused at the same Small to Medium Enterprise (SME) market. Those products were Avaya IP Office, Integral 5, PARTNER ACS, Norstar, BCM and SCS. The road to a single product was introduced in January 2010, and the Avaya IP Office was chosen as the SME platform. In 10 short months, Avaya has integrated the full feature sets of PARTNER ACS and has added support for BCM IP handsets into the Avaya IP Office 6.1 image. The next major software revision for IP Office is 7.0, due out in early 2011, and if all goes well, it will include complete BCM and Norstar features plus handset support. The integration value is huge as there are fewer products with overlapping features to support the large SME market simplifying IT executives’ lives, plus channel partners and Avaya’s businesses.

The Avaya Virtual Enterprise Network Architecture or VENA: With the Nortel acquisition, Avaya picked up the enterprise data networking group and associated products that include Ethernet switching, unified branch, WLAN, network access and network management portfolios. To organize these products and demonstrate an investment cycle, Avaya recently launched Virtual Enterprise Network Architecture (VENA). VENA focused the Avaya product set on a major inflection point occurring in the industry; that is virtualization in the data center as well as on the desktop via VDI and storage. There are clear problems with existing network architecture and design that has focused on physical versus virtual ports since the mid 1980s. New thinking in network design is needed if IT business leaders are to reap the benefits of virtualization as it spreads throughout an enterprise.

VENA defines a virtual service network layer that maps IT services to unique virtual networks that run over a virtual services fabric, which is built upon enhanced IEEE Shortest Path Bridging. According to Avaya, this provides resiliency, simplicity and a consistent interconnect that transparently supports co-existing services. In short, applications and IT resources are assigned to virtual networks that are independent of physical ports, allowing more freedom and much less operator intervention during changes to applications, Virtual Machines, etc.

Avaya’s Prospects

One thing is clear, and that’s Avaya is not tree hugging any technologies or products from the past. It has aligned its UC, contact center and data businesses with major market demands. The Flare Experience is a bold new approach to UC and video collaboration matched only by Cisco and in part Microsoft. It has executed the Nortel integration with speed rivaled only by much larger high-tech firms Cisco and IBM. It could not have picked a better market inflection point than virtualization to add value and investment for its data-networking portfolio. Avaya seems to be firing on all pistons from an operations, engineering, channel expansion and product innovation points of view. If its bets are right, it should be rewarded with market share gain.

Avaya’s Challenges

Avaya does have challenges too. By keeping the data networking group, it has, in essence, become a little Cisco, being about one eighth its size. But Avaya does enjoy very loyal data networking customers, particularly in the financial services industry, which date back to the days of Wellfleet and Synoptics. Avaya and Cisco could not be more different, however. If Avaya is a voice company with some data networking, Cisco is a data networking concern with voice technology. While Cisco and Microsoft have significant pull through sales of communications for their data networking and software, respectively, Avaya does not. Avaya has to compete for data networking and communications business, by and large separately, unless and until it provides a compelling value proposition to supply an architected solution consisting of networks, communications, collaboration and contact center.

Opportunity: Service Delivery Process

One of Avaya’s biggest opportunities lies within its ability to add value to a company’s “service delivery process,” thanks to its rich customer data afforded by Avaya’s Contact Center (CC) business. For example, just this past July, the Avaya CC business introduced the Avaya Aura CC Suite, which is designed to enable end-to-end service experience management. The Aura CC Suite’s Assisted and Automated Experience categories include multi-channel work assignment, self-service and proactive contact applications that drive communications and transactions with customers via voice, email, web chat, SMS or social media. Aura CC Suite also delivers a Performance Management category that includes Avaya’s analytics and reporting platforms, Avaya Call Management System and Avaya IQ, which provide companies detailed customer information that helps to improve profitability and customer retention. In addition workforce optimization and workforce management capabilities were added under the Avaya Aura WFO category.

Service is a key competitive differentiator during this economic cycle where the service economy is the bright spot, and 82% of millennials stop doing business with a company after one bad CC experience. Avaya’s CC customers are equipped with vast customer touch points with every interaction, be it voice, chat, IM, email, etc., being a data point of needs. And Avaya does have a loyal CC customer base. Paradoxically Avaya and Harley Davidson are similar in the fact that both enjoy customers who never switch products. A Harley rider would be hard pressed to switch to another motorcycle as would an Avaya CC customer.

It’s all of these tools to monitor and control customer touch points that deliver so much value to customers and Avaya that it can now be leveraged to another level. Avaya can add value to its enterprise customers by synthesizing, aggregating and monitoring the huge number of customer touch points it offers its customers to afford them deeper market knowledge and allow them to be more adaptive, responsive organizations that deliver differentiated experiences and favorable business outcomes.

Key Avaya platforms are Aura—especially as it invests to make Aura a media-agnostic application platform with special attention to video—and Agile Communications Environment (ACE). ACE facilitates the development of communications-enabled business applications to speed workflow. ACE 2.2 includes an Event Response Manager—a new packaged application that reduces downtime and increases efficiency by automatically notifying the right people with the right skills to respond to and manage unexpected events, such as inventory shortages, security breaches, disasters, stock crashes, etc. A new ACE developer toolkit seeks to make it easier to embed timely and personalized communications into business applications. With Avaya ACE, enterprises can communications-enable their business applications up to 80% faster than by using traditional methods, according to Avaya.

Leveraging Avaya’s huge DevConnect community to write applications around Aura that leverage Avaya’s UC and CC resources while riding over VENA is one sure way to elevate Avaya’s importance and consideration in the Enterprise market. With Avaya’s new nimbleness and innovation, it clearly has the ability to weave its UC, Collaboration, CC and Data products and services around “service delivery process” for its customers, as well as differentiate itself in a significant way. If there isn’t a value proposition around all of Avaya’s products and services, then it may find itself competing in four separate markets, with four separate customers, channels and competitors.

Tweet

Move a Virtual Machine (VM) from one physical server to another, and network port profile, VLANs, security settings, etc., have to be reconfigured. Many networking companies haven’t taken the critical step of giving complete visibility and control of the VM lifecycle from an infrastructure perspective. But with XNV, Extreme Networks is bringing this functionality and visibility to network administrators, tracking VMs and applying policy as they move throughout the network. Shehzad Merchant, Senior Director for Strategy at Extreme Networks, joins me for a discussion about Extreme’s approach to network automation and visibility of virtualized infrastructure through its XNV software module.

Listen to the Podcast

Tweet

Nick Lippis of the Lippis Report announces the participating vendors in the 10GbE Data Center Network Fabric test at the iSimCity lab during the week of Dec 6-10, 2010.  Watch it here

Tweet

By BLADE Network Technologies

To exploit the benefits of server virtualization, data centers need to enable the dynamic and automatic movement of Virtual Machines while protecting their security and maintaining accessibility. Data center network plays a large role in delivering these and other important services for virtualized environments. Current networking switches are not aware of Virtual Machines, and this creates security and availability issues for both server and network administrators as they try to fully exploit the value of virtualization and manage this new environment.

Find out how to build a VM-aware network by downloading this white paper.

Get the White Paper

Tweet

By Siemens Enterprise Communications

One of the greatest obstacles of virtualization is the assignment and allocation of appropriate network resources as virtual machines (VMs) are provisioned amongst diverse network locations. Virtualizing a single data center introduces a number of challenges, not the least of which necessitates moves, adds and changes of virtual images, which adds network provisioning complexity and impacts IT administration workload. Virtualized data center networks need to provide automation, visibility and mobility to support multi-vendor storage, virtualization software and server environments.

Find out how by downloading this white paper:

Get the White Paper

Tweet

Move a Virtual Machine (VM) from one physical server to another, and network port profiles, VLANs, security settings, etc., have to be manually reconfigured, adding cost, delay, security issues and rigidity to what should be an adaptive infrastructure. BLADE Network Technologies’ data center 1/10 and 40 Gigabit Ethernet blade and top-of-rack switches incorporate what its calls VMready with Virtual Vision that enables data centers to deploy VM-aware networks that automate network change management as VMs come online and move between physical servers within or across data centers. Vikram Mehta, CEO of BLADE Network Technologies (BLADE), joins me to talk about BLADE’s data center network virtualization strategy, plus the company’s pending acquisition by IBM. It’s a great talk, and we hit lots of industry nerves in this thought-provoking podcast.

Listen to the Podcast

Tweet

Ronnie Kenneth, CEO, and Asaf Somekh, VP Marketing of Voltaire, a leader in data center networking, are my guests as we dive into Ethernet data center fabric design options. Data Center vendors are discussed too, and you have to hear what they had to say about Cisco. Enjoy, Nick

Visit the Link

Tweet

One significant trend that has emerged during the current business/economic cycle is that IT projects that reduce cost are winners. This savings trend is as strong as I have experienced in my twenty-five years within the IT industry. In particular, it’s propelling data center consolidation, server virtualization and mobile computing projects. As enterprises consolidate data centers and miniaturize them with virtualization, cloud-computing providers are busy offering a new lower cost IT delivery economic model. In short, a new tier of computing has emerged were endpoint devices are mobile and applications are delivered via corporate data centers and cloud computing facilities. This new model of computing that also increases convenience and productivity is lacking in one important area; network security for both mobile endpoints and the ability of data center security appliances to keep up with application demand.

And keeping up with application demand is one of the most challenging tasks IT business leaders are encountering. Not only has information demand skyrocketed during this business cycle but content in the form of web pages has become dynamic, where a single page request opens a multitude of connections pulling content from various sources to satisfy user expectations of real time information access. For example, a single web page request can easily spawn more than fifty network connections over physical and virtual infrastructure placing extraordinary demands on network speed, latency, reliability and security. For the uninitiated, just point your browser to any of these sites—disney.com, cnn.com, nytimes.com, et al—and notice rich content in action. As the page is presented, it serves up video, photos, audio, rich text and more, all of which are pulled from various sources within a data center fabric over virtual and physical infrastructure. The calculus IT leaders are seeking to solve includes massive growth in information demand plus Brownian motion traffic flows, thanks to dynamic content plus densely packed data centers, thanks to virtualization. Even with consolidation and virtualization information/application, demand is forcing the overall data center market size to expand from 108 million sq. ft. in 2009 to a projected 117 million sq. ft. by year end 2010, according to Frost & Sullivan. Part of the solution to IT leaders’ calculus problem is found in a data center network fabric that supports millions of connections/session of east-west and north-south traffic flows securely.

To put the mobility trend into perspective, Apple sold over 3.3 million iPads in its first 3 months; the highest uptake of any endpoint device. Google activates 100,000 Android-based phones per day. Cisco recently announced its CIUS android-based table for business use with tight links to its unified communications (UC) and videoconference systems. Every major UC provider will be offering similar devices while traditional computer vendors serve up android-based tablets over the next few quarters. The iPad and Android tablet is a new tier of computing, which are driving users to access applications over mobile and wireless networks in addition to their wired and VPN networks.

And therein lays the rub. In today’s modern IT world, applications are being extended over multiple networks, e.g., wired, wireless, mobile and remote, where users shift their application access back and forth between these different network access methods and expect the same or consistent experience. Security is paramount to user experience and IT asset protection. While IT security executives have fortified their defenses of IT assets within corporate boundaries or perimeters, exponentially growing numbers of mobile endpoints being connected into corporate networks and data centers present significant security challenges that are unfortunately outside the control of IT.

The nature of mobile smart phone endpoints is to combine personal and business IT services, thereby creating a unique user experience. Part of that experience includes information access from a plethora of online destinations, such as public WIFI hotspots, SaaS applications, e.g., Salesforce.com, workday.com, netsuite.com, etc, corporate VPN, and a wide range of personal sites for social networking, banking, music, videos, news, communications, etc. Therefore, for every employee equipped with a mobile endpoint, security vulnerabilities and threats are opened unless IT mitigates with network security. Clearly mobile devices are becoming ubiquitous, and there are security solutions available, such as VPN support, data wipe after loss, cloud-based security services, etc. But mobile devices need a security solution that works in real time, meaning it’s always-on protection and provides comprehensive coverage.

For example, mobile endpoints, and thus corporate assets, need to be protected from users accessing the corporate network from insecure home WIFI networks and hackers. Internal applications need to be secured against attacks such as SQL injection/data leakage, request forgery/impersonation, cross site scripting/phishing, etc. SaaS access needs to be secure against unauthorized access, exposure from password reuse, layer 7 attacks and more. Also the same level of reporting for mobile users as wired users needs to be supported to assure activity/audit trail, regulatory compliance plus governance and reporting. In short, IT needs the same level of control over mobile endpoints as it does over devices within the corporate perimeter without ruining the mobile experience.

Mobile Endpoint Policy and Enforcement

The most important aspect of real-time mobile security is policy enforcement as it places control of corporate asset and SaaS access back into the hands of IT. Not only does policy and enforcement mitigate threats from being transmitted from mobile endpoints onto corporate networks, it makes them safer devices, too, by providing a means to adhere to corporate policy as corporate devices, even though they are used for business and pleasure. This is important as many mobile devices are purchased by employees, part of the huge consumerization trend that has been building over the last five years. With IT able to administer policy with a means of enforcement, mobile devices can deliver personal and business IT services. Employees may purchase mobile devices but if they require access to corporate IT, then the endpoint has to comply with corporate policy and IT needs a means to enforce such policy. In short, policy and enforcement enables IT to extend the corporate perimeter around mobile devices to creating a virtual perimeter around IT assets.

Consider the following example of policy and enforcement creating a virtual perimeter… A user may be accessing an SaaS application while at his/her desktop. This flow traverses the corporate firewall with associated policy and enforcement. When this user is outside the corporate perimeter, he/she could access the SaaS application directly without corporate policy or enforcement opening vulnerabilities. However, with mobile policy and enforcement, this same user could access the SaaS application with the same policy, enforcement and protections as available when within the corporate perimeter mitigating any vulnerability. Solutions to this usually require the mobile device to first pass through the corporate firewall or a security cloud service where IT controls policy before the user connects to the SaaS application.

New Security Performance Demands

With mobile endpoints under corporate IT policy and enforcement, this huge security vulnerability can now be managed and mitigated. At the same time that mobile devices are becoming ubiquitous, data center security appliances are failing to keep up with the huge demand for information and application access. As more compute power is concentrated into smaller spaces, traffic volume increases exponentially, and security appliances need to adjust accordingly.
Consider how web sites serve up a rich media web page. Every time a user requests a webpage, its server typically needs to request 50 to 100 different objects just to display the one webpage requested. Now consider a data center with thousands of servers and five-thousand connections per second of requests each spawning 50 to 100 server requests. The backend east-to-west traffic flows between servers are one to two orders of magnitude larger than the north-to-south user request flows with the combination of both flows being immense.

New Firewall/IPS Performance Metrics Needed

From a security point of view, not only is firewall throughput an important performance metric, but “connections per second” is becoming more important. A high number of “connections per second” supported assures IT that backend server flows are being screened without delaying user experience. In addition to the number of connections per second, another performance measurement is “maximum connections” supported per second to assure that the number of server-to-server flows to deliver a webpage can be securely delivered. The combination of throughout, connections per second and maximum number of connections can be defined as “true scale performance.” Typically a firewall can deliver hundreds of thousands of connections per second, but this is too slow for most demanding data centers by at least a factor of 2 to 3. Typical maximum number of simultaneous connections supported per firewall is around a few million, which is too low by at least a factor of 4 to 6. Also consider a more realistic throughput measurement other than a range of UDP packet sizes, which is common in the industry. Real world throughput performance numbers that represent a mixture of traffic profiles is a better measurement to assure throughout quoted is throughput experienced.
In addition to raw security performance, data center rack space too needs to be carefully managed as IT executives quickly start running out of rack space as they consolidate. Security appliances need to reduce their footprint as many appliances occupy 16 to 24 RU or a half rack of space and more consuming footprint, energy and cooling resources. Expect security appliances to start delivering on the above performance metrics at up to an 8th of their size or 2 RU high if not smaller.

Threat Protection

To assure this security infrastructure protects IT assets at the rate in which cybercriminals and hackers wish to penetrate it, the industry is serving up cloud-based threat protection. A few suppliers have launched cloud-based security services, which collect anomalistic data throughout the internet and corporate networks via sensors, analyze/correlate the anomalies with reputation scores and when a new exploit’s signature is detected, the cloud transmits mitigation code/signature updates to corporate IPSs. The speed in which this process takes place is a competitive differentiation. Those that send updates every five or so minutes have the best chance of mitigating exploits from cybercriminals which tend to change IP address every hour to avoid detection. IT business leaders will know when cloud-based threat protection becomes highly reliable. It’s at that point that suppliers will start offering “guaranteed protection” that incorporate penalties to suppliers if protection is penetrated.
Policy and enforcement of mobile devices creates a virtual perimeter while true scale performance enables security appliances to keep up with application demand and new traffic flow realities. Smaller security appliance footprint allows IT executives to maximize data center space while minimizing energy and cooling. Cloud-based threat protection keeps the security infrastructure updated in near real time with signatures to mitigate threats throughout the corporate and virtual perimeter. In short, IT business leaders gain control and manage mobile security vulnerabilities while delivering applications to users securely at speed with small footprint consumption. Mobile, data center consolidation and virtualization plus cloud computing are powerful trends rooted in economic efficiency and increased information demand. To maximize the value of these investments, a new security model is needed.

Tweet

Before HP purchased 3Com it would seldom engage in discussion about the future of networking. It was product focused. But now in an exclusive interview with Mr. Saar Gillai, Vice President of Advanced Technology Group & CTO of HP Networking, HP opens up and talks freely about industry trends it’s watching and most importantly influencing. These trends are guiding HP investment in networking with hopes of market share gain, thought leadership and increased industry influence. It’s a fascinating interview into the mind of HP Networking.

Listen to the Podcast

Tweet

Contact center technology is being virtualized at the agent desktop and server levels offering favorable economics, agent flexibility and a centralized data center deployment design. For Siemens Enterprise Communications, a leader in the Contact Center market, its OpenScape Contact Center V8 software runs on top of a VMware hypervisor. In addition to efficient server deployment V8 adds powerful agent features such as a Web-based agent, supervisor and management desktops making it easier to deploy home-based or remote agents virtually at multiple locations. George Despinic, Global Contact Center Marketing Manager for Siemens Enterprise Communications is my guest as we discuss the company’s OpenScape Contact Center V8 that puts contact center functionality in a virtualized data center offering new design and economic options for contact center deployments.

Listen to the Podcast

Tweet

Major IT Delivery Transitions IT Business Leaders Are Managing
Application owners and developers have been deploying and writing applications as if networks had no boundaries or were borderless. By “application owners” I mean IT departments chartered with IT application delivery and management. By “application developers” I mean in-house corporate software developers, independent software vendors (or ISVs) and software companies. There has always been a disconnect between applications and network architects where developers write applications to run over a network as long as there is connectivity. In addition, service-oriented architecture (SOA) based applications call for greater application componentization, which increases messaging between application components, resulting in the network having a direct impact on application performance. In essence, application owners, developers and application standard bodies assume that networks are borderless as the industry is organized around the OSI model where knowledge and skills at one layer, e.g., the network is not necessarily taken into account at another layer, i.e., the application. Therefore, the normal state of affairs is that network designers have been tasked to optimize applications to improve user experience especially when the application was not written to run over a particular kind of network. This status quo does not scale and needs to be re-thought.

Business Drives Applications that Drive Computing that Drive Networking

Every cycle of computing has brought with it this discontinuity between applications and networks with the possible exception of mainframe computing and SNA. Minicomputer applications designed for local ASCII terminal connections were extended over the Wide Area Network (WAN) and via virtual terminals. Client-server computing applications designed to run over Local Area Networks (LANs) were extended over the WAN. At first the internet was text based until the mid 1990s when the web was developed, bringing graphics, audio and video to a network that needed a massive upgrade to support new media rich applications.

IT today is no different. Application developers are writing mobile applications at a frenzied pace thanks to Apple’s iPhone and iPad, Google’s Android, RIM’s Blackberry and now Cisco’s CIUS plus Avaya’s Flare, etc. Legacy enterprise applications are being extended to mobile platforms too with the assumption of a suitable network for delivery. At the same time, applications are being increasingly centralized into consolidated data centers creating greater distance between users and their applications plus data. Some estimate that over 80% of enterprises have undergone a data center consolidation process, which is significant, but we are just at the beginning of the centralization trend.

Thanks to the economics and performance offered by server virtualization, much more consolidation will occur with associated challenges. For example, IT leaders require application tracking as applications are moved from Virtual Machine (VM) to VM as they tune/optimize their virtual infrastructure or respond to peak loads as well as manage VM failovers. In addition to virtualization, massive data centers we call cloud-computing facilities are being built to host applications at scale plus offer infrastructure, platform and other IT services. According to the Yankee Group, 56% of IT business leaders seek to take advantage of cloud-computing technology and build their own private cloud center while 24% seek a fully-managed cloud-computing facility. In the same study, 32% of IT business leaders will seek a hybrid cloud approach that is, connect their private cloud to a service provider’s public cloud. While these market numbers are impressive, they could be much higher as IT leaders express that their top three concerns as they consider cloud services is application performance issues, according to IDC.

In addition to increased mobile and cloud-computing trends, video communications, both on-demand and real-time, have become the largest percentage of internet traffic type. In fact, Cisco Systems recently predicted that by 2014 video traffic will be greater than 94% of all global internet traffic!

This disconnect between applications and network architects will more than likely continue as application owners/developers/standards continue to view networks without borders and boundaries. However, for most network architects, there is no single network, but a wired network, wireless, campus, wide area, data center, branch office network, telecommuting network, mobile network, etc. In fact, most enterprises have a diverse infrastructure in which they are tasked to delivery applications over and for those applications to perform at high standards. The good news is that network designers and architects are starting to build borderless networks that anticipate unforeseen application changes, are equipped with a portfolio of application performance features and simplify deployment and management of IT services…more on this below.

Application Performance Challenges

From the above discussion, it’s clear that enterprise-computing applications are being demanded and stretched over increasingly borderless networks. Consider that the number of small or remote offices and mobile employees are increasing significantly. It’s impossible to argue the mobile computing surge with over 3.3 million iPads shipped in the first three months of its launch, and new entrants such as Cisco and Avaya offering CIUS and Flare tablets, respectively, for business users. In addition, data centers are being consolidated with cloud computing, offering further consolidation and centralization of applications. Applications are changing too as developers add rich media features, and video becomes a dominate application type. Employees, customers, partners and suppliers will be accessing applications over ever-larger distances, via a plethora of endpoints and different networks.

To assure applications perform their task and deliver an excellent user experience, network architects and designers will be increasingly challenged with network capacity being taxed as a wider application portfolio competes for network resources. Today’s model of application performance optimization is to implement appliances within remote sites and data centers, which increases certain application performance, but at the high capital and operational expense of increased network complexity. In addition to network capacity and complexity issues, latency or application transaction delay and how to efficiently utilize data center resources are challenges faced by network architects as they seek to maintain high application performance over a borderless network. Relating specific application transaction problems to network behavior to ascertain if a correlation exists is yet another challenge.

Application Performance Creates Corporate Value

At the center of application performance is corporate performance. The ability of IT leaders to respond to executive management directives is directly linked to corporate performance. Executive management may be challenged with a competitive threat or a new market opportunity, etc., requiring fast corporate response. IT leaders who can execute directives quickly have built an agile business capable of changing when markets or customers shift under them, placing their corporation in a better competitive position to serve its customers and prospects. For example, consider a retail store under competitive pricing pressure where executive management decides to respond with an alternative offer. IT may be able to display the new offer via digital signage quickly allowing the business to respond.

Key to business agility is the IT attribute of rapid innovation absorption–that is, the capability to deploy new applications and technologies at the speed of business opportunity. Most IT infrastructures consist of innovation and features which are already in place, but IT organizations require knowledge, skills and tools to put them to work when needed.

A borderless network that is capable of application performance delivers these attributes of innovation absorption and business agility. In addition, IT resource utilization can be optimized, and most important to users is that they gain an excellent IT experience independent of geographic location, endpoint device or application, which in the end improves productivity.

As an example of optimal resource utilization, consider Cisco’s ISR G2 branch office router that integrates unified communications, wide area application optimization, network security, LAN/WAN networking plus supports its AXP (or Application eXtension Platform), which run applications at the branch office router. In one branch office, an IT manager can deliver networking, security, voice and video communications and host applications while gaining visibility to applications. This type of resource utilization not only saves on capital cost and energy spend, but offers IT operational efficiency, rapid application deployment and innovation absorption.

To gain the full value of corporate applications, their performance must deliver excellent user experience. An excellent experience should not only occur while working in the office or at home, but anywhere in between, even while talking on a mobile endpoint. Independent of geographic location, a user accessing his/her business services and/or personal services should be the same seamless experience. Application performance is key to excellent experience and should be consistently good whether sitting at a desktop watching a video or engaged in a Web conference, and then immediately transitioning to an iPhone for example. The user should have an excellent experience at the highest level afforded by his/her endpoint. To deliver this seamless user experience, application performance technology needs to be incorporated in corporate IT infrastructure, endpoint devices or a combination of both.

That is, networking silos need to become an integrated network without borders. For applications to offer the best possible user experience, then the use of application acceleration technology as appliances or an overlay needs to be integrated into the network fabric and into network operating systems. This technology, which has improved application delivery for specific applications, needs to become systemic and fully distributed throughout the network fabric. The integration or pervasiveness of application acceleration technology within networks and endpoints is its natural evolutionary next step. Over the next few months we’ll see vendors such as Cisco, HP Networking, Juniper, Riverbed, Citrix, Blue Coat, et al, start to deliver on this vision.

Tweet

By Arista Networks

VM Tracer offers the tightest integration between virtual machines and network infrastructure. Visibility in to virtualized infrastructure is key for management and operations of mobile and virtual environments.

Find out how by downloading this white paper.

Get the White Paper

Tweet

By Arista Networks

The increase in utilization of virtual machines and increasingly virtualized resources in the data center has caused a reduction in network visibility into the virtual infrastructure. VM Tracer provides the visibility necessary for the network team to support virtual environments and the automation necessary for server administrators to be effective and efficient.

Find out how by downloading this paper.

Get the White Paper

Tweet

I interview Marius Haas, Senior Vice President and World Wide General Manager of the HP Networking Business. I asked him

• How HP will compete in the highly competitive networking market,
• What is its salient value proposition,
• Why is HP so focused on networking now, and
• Why should IT business leaders consider HP Networking.
His responses had a ring of market truth to them. As such, any IT business leader worth his/her salt needs to listen to this podcast.

Enjoy, Nick

Listen to the Podcast

Tweet

Networking is entering a new phase or era. During the 1990s, new networking markets opened up, creating multi-billion dollar opportunities for the vendor community and corporate cost savings for IT business leaders. First, it was shared LANs and routing, then switched LANs, then Frame Relay to speed up WANs, then SNA over IP, then remote access via dial-up and VPN, then MPLS, then IP telephony, then Wireless LANs etc… and now, it’s video and cloud networking. You get the picture. But what we didn’t realize as we build these networks is that they are silos with disparate management systems and unique access methods resulting in operational cost overlap and, most importantly, user frustration as they transition application use from desktop, to mobile end point, to remote endpoint. In short, we built boundaries around applications in the form of networks and it is the dismantling of these borders that vendors are now starting to deliver and differentiate upon. It’s not just Cisco that communicates borderless networks, but HP Networking, Juniper, Brocade, Extreme, Avaya, Force10 and others too. Why is the industry entering a new age of borderless networking and what’s in it for IT business leaders, is explained in this Lippis Report Research Note.

As each new wave of computing entered corporate IT departments, a new set of networking requirements arose. To connect remote 3270 terminals via SNA to mainframes, IT implemented an analog multipoint wide area network or WAN. To connect remote ANSI terminals to minicomputers, IT departments implemented pools of dial-up modems and private line WANs. To connect personal computers (PCs) via Client-Server computing, IT departments implemented Local Area Networks or LANs via LAN switches, which we now call wired connections. To connect multiprotocol LANs over the corporate WAN, IT departments implemented routed networks. To gain access to LAN based applications while remote, IT departments implemented Virtual Private Networks or VPNs. And, as computing and applications go mobile, IT has been implementing Wireless Local Area Networks or WLANs. In short, each network was deployed to service a certain computing style and application set. These networks are silos, and with advances in technology, IT business leaders can now design one borderless network to provide a broad array of common access methods to support a plethora of endpoints and applications.

Siloed networking frustrates users, as each access network performs differently depending upon its access method. Siloed networking also frustrates IT, as each siloed network has its own management system creating inefficient IT operations. In addition, siloed networking does not meet today’s IT “any access” requirements.

There are boundaries or silos that need to be broken down in many places of the network. In today’s modern IT world, applications are being extended over multiple networks e.g., wired, wireless, cellular, remote, virtual, etc where users need to shift their application access back and forth between these different network access methods and expect the same or consistent experience. In short, networks need to be borderless so that applications can be accessed independent upon network entry point and IT operations efficient. This “any access” trend is accelerating as IT business leaders seek to connect not only traditional desktops and laptops, but smartphones, notebooks, tablets, iPads, cameras and building control systems into a common general purpose network that support multiple logical network topologies.

Crossing purpose-built silos is difficult for applications, as bandwidth and quality of service issues limit application portability thus their usefulness. These different access methods offer limited consistency resulting in user frustration when they shift application access from desktop to mobile smartphone to VPN and back again.

And this shifting of application access between different networks and endpoints is only going to increase. Apple sold over 3.3 million iPads in its first 3 months, the highest uptake of any endpoint device. Google activates 100,000 Android based phones a day. Cisco recently announced its CIUS android-based table for business use with tight links to its unified communications (UC) and videoconference systems. Every major UC provider will be offering similar devices while traditional computer vendors serve up android-based tablets over the next few quarters. The iPad and Android tablet is a new tier of computing which will drive users to access their applications over mobile and wireless networks in addition to their desktop and VPN networks.

If IT business leaders are unable to get ahead of this curve and think of network access from an architected and unified design point of view, than unfortunately, their users and IT cost will be more frustrated and expensive, respectively, than others. Siloed networks are friction points as they create boundaries between network access types degrading user experience, which results in decreased productivity and increased IT operational cost. The result is a high total cost of ownership and less then optimal user experience, and thus decreased corporate productivity. The status quo of siloed networking is about to change.

Cisco’s Borderless Network Architecture

From a design point of view, borderless networking requires three core attributes: 1) reliability, 2) security and 3) seamlessness. Cisco was the first to articulate a vision for borderless networks, which has resonated with IT business leaders as it represents a solution to their pain. For example, Cisco’s borderless network architecture is built upon five services: 1) mobility or users in motion, 2) Energy efficiency called EnergyWise, 3) integrated network security via its TrustSec architecture, 4) application performance and 5) video management, control and distribution via its MediaNet. These borderless network services are built within switching, routing, security, wireless and wide area application services or WAAS infrastructure products. It’s the integration of these services into existing network infrastructure and their control via policy and management that enable a borderless experience to occur.

Juniper’s New Network

But Cisco is not the only supplier to grasp the problem siloed networks create. Juniper Networks is working to a similar end, albeit it hasn’t articulated it well. It provides VPN, LAN Switching, mobile security through its acquisition of SMobile and is working toward a flat cloud Ethernet fabric through its project Stratus and New Network initiatives. For example, Juniper plans to integrate SMobile security into its JUNOS Pulse endpoint software for network connectivity and acceleration breaking down the boundary between LAN based and mobile network access.

HP Networking’s Converged Infrastructure

When HP Networking launched its comprehensive network portfolio in April of this year it emphasized the elimination of network silos. The HP Networking portfolio strives to eliminate redundant equipment by integrating wired and wireless environments with security from edge to core. From an IT operations perspective, this translates into a “single pane of glass” for management, configuration, deployment and monitoring these networks as if one. HP Networking hopes to implement a common policy management to reduce human error of network operations while creating a consistent user experience across access mediums.

Brocade One

Brocade has jumped on the borderless bandwagon also in June of this year with the introduction of its “Brocade One”. Brocade One emphasizes the convergence of wired, wireless and cellular networking to offer a seamless user experience. In addition, Brocade One describes its view of a simplified virtualized data center network fabric that scales to cloud spec. In essence, Brocade One is about eliminating the boundaries around wired, wireless and data center networking.

Arista Network’s VM Tracer

Arista Networks doesn’t use the terminology of borderless networking either, but its recent VM Tracer strives to eliminate the boundaries between physical and virtual networking environments. VM Tracer does this by being integrated into Arista’s EOS linking Arista switches to VMware’s vCenter. This linkage creates an adaptive infrastructure in which the network responds to changes in the VM network while also providing complete visibility into the virtual machine network.

Extreme’s DirectAttach

Extreme Networks has focused on removing two network boundaries; the wired and wireless boundary and the physical to virtual network boundary. For the latter, Extreme has introduced its Direct Attach approach to data center networking that eliminates the virtual switch layer, simplifying the network and improving performance.

Force10’s Open Automation

Force10’s focus in eliminating boundaries is in the data center between physical and virtual networks. Force 10′s Open Automation initiative seeks to align dynamic data center changes with network configuration and policies, a huge barrier to virtualized data center management and scale.

While each of the above suppliers are at different points in their borderless network initiatives, the direction is clear. The boundaries between siloed networking are coming down be it in the data center, campus, branch office or home. For IT business leaders this means simplified operations and management as a key attribute is the “single pane of glass” approach to network management for siloed networks. The big surprise and delight will be found in enhanced user experience, as borderless networking strives to deliver a common access method for all networking types while enabling applications to be extended across a plethora of different endpoints, depending upon endpoint capabilities and network resources.

In essence, borderless networking’s value proposition is that it enables a corporation to be more adaptive or agile while increasing user experience and reducing operational cost. With the majority of IT business leaders trading off reductions in operational spend for an increase in capital expenditure, borderless networking is the right solution at the right time.

Tweet

It hasn’t been since the mid 1990s that the networking industry was focused on multi-protocol integration or convergence. But the industry is gearing up for a major innovation and competitive cycle fueled by the multi-billion dollar addressable market for data center network fabrics. Over the last eighteen months, every major Ethernet infrastructure provider has been talking about two and three tier network fabrics for high-end data centers.

Companies such as Cisco, Arista Networks, HP/3Com, Force10, Voltaire, Extreme, Brocade, Juniper et al have announced network fabrics for data centers with five thousand and more servers with and without storage enablement. Juniper talks of a one-tier fabric through their Project Stratus work with IBM to be available some time in the future. Brocade recently introduced its Brocade One, which is a converged data center fabric. Extreme Networks launched its DirectAttachTM that eliminates virtual plus blade switch layers. HP has FlexFabric, a virtualized fabric for the data center. Cisco launched its FabricPath Switching System or FSS for the Nexus 7000 that enables massive scale of a two-tier fabric.

In this Lippis Report Research Note, we review the architectural attributes of two tier network fabrics.

The IT industry is at an inflection point as service delivery is becoming more and more centralized thanks to data center consolidation, virtualization, cloud and mobile computing. It is estimated that a third of all IT spend is concentrated in the data center, and this trend is only building thanks to favorable economics, motivating IT business leaders to centralize IT delivery.

The impact of this trend is more and more dense data centers made up of servers in the thousands to tens of thousands and higher. It is at the scale of 5,000 plus servers that a new network fabric is required for high-end data centers. High-end data center design is challenged with increasing complexity, the need for greater workload mobility and reduced energy consumption. Traffic patterns have also shifted significantly, from primarily client-server or as commonly referred to as north-to-south flows, to a combination of client-server and server-server or east-to-west plus north-to-south streams. These shifts have wreaked havoc on application response time and end user experience, since the network is not designed for these Brownian motion type flows.

The main requirements for high-end data center network fabric are low latency, large flat layer 2 domains to enable workload mobility, low power consumption, simplicity of design and significant bandwidth. Storage enablement, meaning consolidated I/O or virtualized I/O, is a growing priority and a new fabric that can support FiberChannel over Ethernet, iSCSI over Ethernet, iWARP over Ethernet or Infiniband over Ethernet, is a major plus. One salient observation is that it’s pretty clear that Ethernet is the network fabric of choice, as it is the only network protocol that enjoys continual innovation such as TRILL, Data Center Bridging, IEEE’s 802.1AQ, link aggregation, multi-pathing, and as recently ratified by the IEEE 40 Gbs and 100 Gbs speeds.

With the above requirements in mind, let us review data center network design options.

Three Tier Data Center Fabric

A three-tier network architecture is the dominant structure in data centers today and will likely continue as the optimal design for many networks. For most network architects and administrators, this type of design provides the best balance of asset utilization, layer 3 routing for segmentation, scaling and services, plus efficient physical design for cabling and fiber runs. By three tiers we mean, access switches/Top-of-Rack (ToR) switches, or modular/End-of-Row (EoR) switches that connect to servers and IP based storage. These access switches are connected via Ethernet to aggregation switches. The aggregation switches are connected into a set of core switches or routers that forward traffic flows from servers to an intranet and internet, and between the aggregation switches. It’s common in this structure to over-subscribe bandwidth in the access tier, and to a lesser degree, in the aggregation tier, which can increase latency and reduce performance. Inherent in this structure is the placement of layer 2 versus layer 3 forwarding that is Virtual Local Area Networking or VLANs and IP routing. Also common, is that VLANs are constructed within access and aggregation switches, while layer 3 capabilities in the aggregation or core switches route between them.

But within the high-end data center market, where the number of servers is in the thousands to tens of thousands plus and where north-south plus east-west traffic is significant, is where a new structure is needed. It is within these data centers where applications need a single layer 2 domain.

Two-tiers of network fabric

A two-tier fabric is designed with two kinds of switches: one that connects servers, and the second that connect switches creating a non-blocking, low latency fabric. In short, there are server facing and fabric facing switches. We use the terms ‘leaf’ switch to denote server facing or connecting switches and ‘spine’ to denote fabric facing or switches that connect leaf switches into the fabric. Together, leaf and spine switches create the fabric.

Many IT leaders in Global 2000 firms will have deployed both two and three tier network structure, as different deployment models are used for different applications. For these leaders, a network equipment supplier that possesses product architecture flexibility, meaning an end-to-end product solution that accommodates tier two and three fabrics would be advantageous. This flexibility is found in product that supports layer 2 and layer 3 forwarding, as well as, a variety of line cards to offer design options.

A common network Operating System (OS) of products configured for two and three tier structure is important as IT operations gain efficiency to manage fabrics, as configuration and management are consistent. In addition, a common network OS offers rapid absorption of innovation to IT operations, as new OS features are available at the same time to all fabrics. The benefit of using a common product set to build tier two or three fabrics offers value around operational efficiency, training, sparing and ease of evolution between fabric deployments. In short, the network fabric needs to be simple and general purpose versus purpose built, which a common set of products creating tier two or three fabrics offer.

A Unified/Converged Fabric

The concept of a unified fabric is to virtualize data center resources and connect them through a high bandwidth network that is very scalable, high performance and enables the convergence of multiple protocols onto a single physical network. These IT resources are compute, storage and applications, which are connected via a network fabric. In short, the network is the unified fabric and the network is Ethernet.

The industry tends to focus on storage transport over Ethernet as the main concept behind a unified/converged fabric with technologies such as Fiber Channel over Ethernet or FCoE, iSCSI over Ethernet, iWARP over Ethernet and even Infiniband over Ethernet. But this is a narrow view of a unified/converged fabric which is being expanded, thanks to continual innovation of Ethernet by the vendor community and standards organizations such as the IEEE and IETF.

Ethernet innovations such as FCoE, Data Center Bridging or DCB, IETF’s Transparent Interconnection of Lots of Links or TRILL, CEE or Converged Enhanced Ethernet, link aggregation, IEEE’s 802.1AQ have enhanced Ethernet networking to support a wide range of new data center fabric design options. In addition to these protocol enhancements, the IEEE has ratified its work on defining 40Gb and 100Gb Ethernet, significantly increasing Ethernet’s ability to scale bandwidth. To demonstrate how Ethernet is evolving to be the unified fabric for high-end data centers, we explore Cisco’s new FabricPath Switching System innovation in this white paper.

The decision to implement a two or three tier network structure comes down to scale. For high-end data centers, a two-tier structure meets the requirements of low latency, movable workloads, scale, simplicity, etc. Many global 2000 concerns will have deployed both a two and three tier network fabric for their high end and less dense data centers.

When shopping for network equipment to construct two and three tier network fabrics, look for suppliers that support both rich Layer 3 routing services and scalable Layer 2 Ethernet capabilities to ensure choice and flexibility of three tier and scalable two tier fabric implementations. Such suppliers offer products that can be configured in multiple use cases and topologies where modules are inter-changeable, skills transferable and operations common between both fabric approaches.

But make no mistake about it, it’s a two-tier network fabric that IT business leaders and data center architects have gravitated toward for high performance computing, cloud scale data centers and just plain high end data centers of 5,000 and above servers.

Tweet

By Extreme Networks

A typical “non-virtualized” data center has three network layers, Top-of-Rack, End-of-Row and Core switches. But virtualized infrastructure adds two additional layers; the virtual switch and blade switch, raising the number of tiers from 3 to 5. This significantly increases latency plus the number of network elements within the data center resulting in increased data center management complexity.

Extreme Networks® Direct AttachTM eliminates the virtual switch layer, simplifying the network and improving performance. Extreme Networks high density BlackDiamond® 8800 series switches with 8900-series modules further enable data center simplification by utilizing high density blades and cabling to eliminate the blade switch, thereby reducing the number of tiers in the data center from 5 to 3.

Find out how by downloading this white paper:

Get the White Paper

Tweet

Networking has become “rigid”. Yes I know it’s almost absurd to attribute inflexibility or rigidity to networking. Look what TCP/IP has done for us. There are nearly 2 billion people connected to the internet and according to the Internet World Stats internet user growth rate increased by 380% between 2000-2009. With 2 billion people and growing online, accessing a plethora of applications via a wide range of end-points there is no doubt that the internet and TCP/IP has been a much bigger success than anyone would have imagined back in the early ’90s. But there’s always a give and take between computing and networking where one drives and changes the other. Right now we are in a compute innovation cycle that’s driving a fundamental change in networking which screams out the need for more flexibility.

Sure networking has increased from a bandwidth point of view and the IETF has added new protocols and network services, but it hasn’t kept up with compute innovation. As data centers pack more compute power and operating systems (OS) per physical server, thanks to virtualization, the need to move containers of OS plus applications and data around have sky rocked. In addition, traffic patterns have shifted tremendously as client-server or north-south flows are layered on top of server-server or east-west flows. And yes, there are new networking approaches being offered by vendors and standard organizations such as Cisco’s FlexPath, Juniper’s Stratus, Brocades VCS, Extreme’s Direct Attach, Force 10’s Open Automation, Arista’s Multi-Chassis Link Aggregation, BLADE’s Unified FabricArchitecture, the IETF’s TRILL and LISP and IEEE’s 802.1AQ, but these may be short term solutions to a much bigger networking problem.

Computing has always driven network design as mainframes drove SNA and analog multi-point wide area networks (WANs) during the ’70s. Mini-computers drove peer-to-peer networking protocols like DecNet, OSI and TCP/IP in the ’80s. Client-Server computing drove LANs and TCP into the mainstream in the early ’90s. The Web drove the internet in the 2000s and now server virtualization and cloud computing is once again changing fundamental networking requirements to make them more flexible.

The rigid label is a powerful one as it creates frustration by not addressing or enabling new business processes. Every time a network protocol or architecture was labeled as too rigid it was replaced and in the process a new market emerged on the scale of tens of billions of dollars. SNA was labeled as too rigid to support peer-to-peer networking. The T1 multiplexer market of the late ’80s and early ’90s was too rigid to support data traffic and thus routing replaced it. The PSTN and TDM were too rigid as they doled out bandwidth in 56Kbs chunks and were unable to support internet and VoIP traffic. The national entertainment network is rigid too as it doesn’t support two-way communications and it also will be replaced slowly but surely.

So where is networking not flexible enough? It’s in virtualized data centers. Some analyst groups estimate that 30% of workloads are virtualized and increasing. Since virtualization or a VM is the new atomic layer of data centers, networking is falling short in public as well as private clouds. Ideally, all resources (compute, storage, and networking) would be pooled, with services dynamically drawing from the pools to meet demand. Virtualization techniques have succeeded in enabling processes to be moved between machines, but constraints in the data center network continue to create barriers that prevent agility, for example, VLANs, ACLs, broadcast domains, Load Balancers, Firewall/IPS Security settings and service-specific network engineering.

The well understood problem is that when a VM is moved from one physical machine to another the network, load balancers, firewalls/IPS, broadcast domains, etc., have to be reconfigured. There is no automation in place, meaning that the network is not flexible or agile enough to make the changes required. Now this problem has scale to it as it’s a growing requirement of both IT executives managing corporate IT assets and service/cloud providers.

There are market solutions available today and more are coming that address “network automation” which enable the network to reconfigure itself as a VM and/or workload is moved within a data center. Cisco’s Nexus 1000V, HP Network Automation software and its Virtual Connect approach, Force 10’s Open Automation, Blade Network Technologies VMReady Network Virtualization, Arista Network’s Virtualized Extensible Operating System or vEOS and others are addressing the problem of network agility or lack thereof in virtualized environments.

But the problem gets bigger and more complex when distance and cloud provider entities become engaged. None of the solutions above address moving a VM from one physical server to another over large distance, be it around town, across state lines, across the country or the globe. Some are using IF-MAP as a registry, sort of like facebook for computers that publish their resources and use this information to automate network configuration to support large distance VM moves.

The problem gets larger yet when workloads move from a private cloud to a public cloud. (Definition note: There is no single definition of a workload, so for my purpose here I assume a container including a VM and associated applications and data that can be moved as simply as drag and drop or some other string of instructions). In short, all the software that is needed to compile and run an application for a set of users is a workload. The network inflexibility problem grows even larger when moving workloads between public clouds.

Now is this a real problem? You bet it is. Consider the value also of portable or mobile workloads to Enterprise and service providers. Workload mobility means capacity on demand, business continuance, and disaster recovery, etc. In addition, as IT leaders explore public and private cloud alternatives, they will want to move workloads from their data center to a provider’s and move the workload back when and if required. For reasons of security and trust, IT business leaders will demand mobility. For example, if your cloud provider goes bankrupt, then you will want to move your workload out quickly. If your cloud provider’s performance drops again then you could move your workload out. If your cloud provider is the target of a terrorist attack or is turned into a large botnet then you can move your workload out.

In addition to security and piece of mind, mobile workloads will fundamentally change IT delivery, capital structure and most importantly business models and processes. Once IT can move workload anywhere in their data center, across their data centers or to a provider they have tiered with, the question becomes when and how fast does IT move workload? If IT can perform all the provisioning in software and enable workload moves to occur transparently and safely with address, identity, security preservation, enabled trust, control and interoperability across providers, then the question is when does IT need to move workload? This level of mobility is an industry-wide initiative as it offers significant and material business value. Business value is created as IT could move workload in a follow- the-sun model, following the lowest cost per kilowatt-hour model; workload could move to avoid a disaster, or for capacity on demand, or for lowest cost of workload execution, etc.

So how can data center networks become more flexible? A key element of the solution is agility or the ability to dynamically grow and shrink resources to meet demand and to draw those resources from the most optimal location. Today, the network stands as a barrier to agility and increases the fragmentation of resources, which leads to low server utilization and prevents portable or mobile workloads.

Tweet

I interview Vikram Mehta, President and CEO of BLADE Network Technologies or BLADE. BLADE has been offering top-of-rack and blade switches since 2006, boast nearly 9 million ports installed and are now the number 2 provider of 10GbE switches. With all this success, I ask Vikram what BLADE is being asked to offer from IT Business leaders as they build out their data center Ethernet fabric. You don’t want to miss his responses. Listen now.

Listen to the Podcast

Tweet

In Lippis Report 151: A Two or Three Tier High-End Data Center Ethernet Fabric Architecture? we detailed the new two tier data center Ethernet fabric that is becoming conventional wisdom amongst business leaders of high end data centers and cloud computing service providers. The networking industry is headed for a major innovation and competitive cycle fueled by a multi-billion dollar addressable market for data center network fabrics. Over the last eighteen months, every major Ethernet infrastructure provider has announced or taken a position on two tier network fabrics for high-end data centers. Companies such as Cisco, Arista Networks, Force10, Voltaire, HP/3Com, Juniper, Extreme, Brocade, BLADE Network Technology, et al have announced network fabrics for data centers with two thousand and more servers that either support storage enablement or not. In this Lippis Report Research Note, we review why it is Ethernet that will be the network fabric of high performance computing or HPC and cloud computing deployments.

For high-end data centers, HPC plus private and public cloud computing networks connecting thousands of servers, a new set of requirements have emerged. Low latency and high performance are the two driving requirements. Yes, there are more, especially when the fabric needs to enable converged storage, but let’s focus on latency and performance for now. Traditional three tier (server access, distribution and core) fabrics designed primarily for north-south traffic flows, that is client-server computing utilized spanning tree protocol (STP) and slower speed Ethernet (100Mbs to 1Gbs). Thanks to web 2.0, mash-ups and social networking sites east-to-west or server-server traffic flows have spiked requiring networks to support both north-south and east-west flows.

As most network engineers know, STP was designed to avoid loops that confused Ethernet as it was designed as a bus topology. STP shuts down redundant links between common switches to maintain the bus. Therefore, connecting access switches to distribution switches utilizing STP would require that network engineers over-subscribe the links between switches as only half of the bandwidth could be used. Oversubscription would also create blocking of packets between points too. To avoid this design, nearly every major switch manufacturer offered link aggregation that is the ability to shut off STP and aggregate links between switches. While this was and is a benefit, the down side has been that vendors only offered the ability to aggregate two links, which still drove oversubscription and blocking.

Recently, industry players such as Cisco and Arista Networks have offered the ability to scale up aggregation of links from 16 to 32, while at the same time delivering multipathing that allows packets to be forwarded across multiple links to arrive at its intended destination. Switch-processing capacity to support these massive inter-switch links have been increased too. These design changes, along with Ethernet’s innovation march, has ushered in the two-tier network design fabric option.

A two-tier fabric is designed with two kinds of switches; one that connects servers and the second that connect switches creating a non-blocking, low latency fabric. We use the terms ‘leaf’ switch to denote server connecting switches and ‘spine’ to denote switches that connect leaf switches. Together a leaf and spin architecture create the network fabric.

In late June 2010, Cisco announced its’ FabricPatch Switching System or FSS and its’ F-Series modules that support 32 ports of 10GbE of auto-sensing 1/10GbE and is essentially for server access and aggregation. FabricPath provides a new level of bandwidth scale to connect Nexus switches and delivers a new fabric design option with unique attributes for IT architects and designers. FabricPath is a NX-OS innovation, meaning that its’ capabilities are embedded within the NX-OS network OS for the data center. FabricPath essentially is multipath Ethernet; a scheme that provides high-throughput, reduced and more deterministic latency, and greater resiliency compared to traditional Ethernet.

FabricPath combines today’s layer 2 or Ethernet networking attributes and enhances it with layer 3 capabilities. In short, FabricPath brings some of the capabilities available in routing into a traditional switching context. For example, FabricPath offers the benefits of layer 2 switching such as low cost, easy configuration and workload flexibility. What this means is that when IT needs to move VMs and/or applications around the data center to different physical locations, it can do so in a simple and straightforward manner without requiring VLAN, IP address and other network reconfiguration. In essence, FabricPath delivers plug and play capability, which has been an early design attribute of Ethernet. Further, large broadcast domains and storms inherent in layer 2 networks that occurred during the mid 1990s have been mitigated with technologies such as VLAN pruning, Reverse Path Forwarding, Time-to-Live, etc.

The layer 3 capabilities added to FabricPath deliver scalable bandwidth allowing IT architects to build much larger layer 2 networks with very high cross-sectional bandwidth eliminating the need for oversubscription. In addition, FabricPath affords high availability as it eliminates STP, which only allows one path and blocks all others, and replaces it with multiple paths between endpoints within the data center. This offers increased redundancy as traffic has multiple paths in which to reach its final destination.

FabricPath employs routing techniques such as building a route table of different nodes in a network. It possesses a routing protocol, which calculates paths that packets can traverse through the network. What is being added to FabricPath is the ability for the control plane or the routing protocols to know the topology of the network and choose different routes for traffic to flow. Not only can FabricPath choose different routes, it can use multiple routes simultaneously so traffic can span across multiple routes at once. These layer 3 features enable FabricPath to use all links between switches to pass traffic as STP is no longer used and would shut down redundant links to eliminate loops. Therefore, this would yield incremental levels of resiliency and bandwidth capacity, which is paramount as compute and virtualization density continue to raise driving scale requirements up.

Designing A 160 Tbps Data Center Fabric

As an example to how multi link aggregation, the elimination of STP, high switching capacity and 10GbE connections create a highly scalable two-tier layer 2 Ethernet fabric, we use Cisco’s FSS and its’ F-Series module in the Nexus 7000. The following details the design of a 160 Tbps switching fabric with FabricPath and the F-Series module for high performance data centers using Cisco’s Nexus 7000 switches. This architecture can support over 8,000 servers connected at 10GbE or 4,000 servers dual homed at 10GbE with attributes of being non-blocking, low latency (5 microseconds), high bandwidth, reliability, plus simplicity of workload movement.

To build a 160 Tbps two-tier fabric, thirty-two Nexus 7018 switches populated with F-Series 10GbE modules would connect servers. These thirty switches are leaf switches. Each leaf chassis provides 256 10GbE ports to connect servers and another 256 10GbE ports to connect into spine switches. Therefore, each leaf is directly connected to each spine with sixteen FabricPath ports at 10GbE equaling a total of 256 10GbE ports for each leaf switch. There are sixteen spine switches each accepting 512 10GbE FabricPath ports. A single leaf chassis connects 256 10GbE ports into a spine equaling approximately 2.5Tbs. Multiplying each thirty-two leaf’s contribution into the fabric yields 80Tbs. As Ethernet is full-duplex, the total fabric switching capacity is 160
Tbps. Therefore, 160Tbps of switching fabric is available across all thirty-two leaf chassis. As 256 10GbE equals 2.5 Tbs, which also equals 16 FabricPath links to each one of sixteen spine switches, yields 2.5 Tbs, the fabric is non-blocking.

As for layer 2 and layer 3 forwarding, the job of the spine is to forward packets from leaf switches at layer 2, creating a single tier fabric. A key attribute of this architecture is that each 16-way FabricPath links are Equal Cost Multipathing or ECMP. What 16-way FabricPath ECMP provides are two benefits: 1) It delivers more paths for traffic to flow, which increases available bandwidth in the fabric and 2) as they’re distributed across all switches, diversity of routes is enabled to distribute packet forwarding. In essence what 16-way FabricPath ECMP provides is a very low latency, high bandwidth approach to supporting both north-to-south and east-to-west traffic flows simultaneously.

While the above is a Cisco deployment example Arista’s new 7500 series of Ethernet switches support 6 Billion packets per second at wire speed. The 7500s can be configured into a massive two-tier network fabric thanks to it support of 32 port MLAG (Multi-Chassis Link Aggregation) affording the connection of 18,000 to 30,000 servers.

Ethernet continues to evolve. The IEEE recently ratified the 40 and 100 GbE standard with vendors such as Force 10, Cisco, Arista, Extreme, BLADE, Brocade, Voltaire, HP et al announcing support and scheduling product delivery. While the above two-tier network example provides the perspective from the large switch provider, below is BLADE Network Technologies perspective, a company focused on server connectivity.

BLADE Network Technologies believes that as Ethernet delivers new levels of speed and intelligence, it will be the dominant two-tier network fabric for high-end next-generation data centers.
For many applications, low latency is a key requirement, and latency is an area where two-tier networks excel. Studies of stock trading exchanges have shown that tens of milliseconds of delay in data delivery can represent a ten percent drop in revenues, and delays of even five microseconds per trade can cost hundreds of thousands of dollars. Industry-specific requirements for uncompressed data and end-to-end deterministic latency within tens of microseconds make attaining such performance even more difficult. These factors have combined to make raw switching speed a top priority, and today’s best-of-breed 10 Gigabit Ethernet switches achieve can operate with under 700 nanoseconds of port-to-port latency while consuming a miniscule amount of power equivalent to that of standard light bulbs.

As next-generation networks get flatter – driven by latency and bandwidth requirements – emerging Layer 2 technologies such as the IETF’s Transparent Interconnection of Lots of Links or TRILL, enable this trend. The idea behind TRILL is to replace spanning tree as a mechanism to find loop free trees within Layer 2 broadcast domains. Using a routing protocol to build forwarding trees within a Layer 2 broadcast domain enables the flexibility and efficiency to route Layer 2 traffic, just like one would Layer 3 traffic, without the overhead associated with Layer 3 packet processing. TRILL will offer important features, such as support for both broadcast and multicast, load splitting along multiples paths, support for multiple points of attachment, and no tangible delay in service after attachment.

In the data center, bottlenecks are moving from the CPU and memory access to the I/O of the servers. Today’s multi-core servers are now able to sustain a great amount of traffic, requiring fast, flat networks, especially now that virtualization is widely deployed. Analysts have predicted that the 10G market will double year-to-year in 2010 and 2011. More servers using 10G increases the requirement for 40G and 100G in upstream networks. With 10G widely available and 40G coming online, Ethernet networks can enable data and storage traffic to use a single wire, using FCoE or iSCSI for example, and provide the raw speed that makes Ethernet with its economies of scale, to supplant InfiniBand for HPC requirements.

The reason Ethernet will be the network fabric for high-end data center networks is that the vendor community continues to innovate and build upon this protocol. Ethernet innovations are many and are beyond bandwidth increases from 10Mbs, 100Mbs, 1Gbs, 10Gbs, 40Gbs and 100Gbs, which are obvious. Link aggregation, multi-pathing and so much more propel Ethernet’s relevance and suitability to new challenging networking requirements.

Tweet

Massively scalable data centers have unique requirements such as low latency, high performance, non-stop operation, simplicity of design, workload mobility and storage transport support. To address these requirements Cisco launched FabricPath, which is a scalable multi-link and multipath technology allowing 2 to 48 Nexus 7000s to be configured in a large, non-blocking Ethernet switch fabric. This fabric, called a FabricPath Switching System or FSS, eliminates oversubscription and creates a two-tier fabric. In addition to FabricPath, Cisco launched its’ F-Series 10GbE module for the Nexus 7000, which offers 32 ports of auto-sensing 1/10GbE targeted at server access, aggregation and FCoE implementations solutions. Craig Griffin, Senior Director of Product Management for Cisco’s Nexus 7000 discusses new Ethernet innovations for the age of massively scalable data centers.

A white paper on the topic is available here:

Listen to the Podcast

Tweet

It hasn’t been since the mid 1990s that the networking industry was focused on multi-protocol integration or convergence. The industry is gearing up for a major innovation and competitive cycle fueled by the multi-billion dollar addressable market for data center network fabrics. Over the last eighteen months, every major Ethernet infrastructure provider has been talking about two and three tier network fabrics for high-end data centers. Companies such as Cisco, Arista Networks, HP/3Com, Force10, Voltaire, Extreme, Brocade, Juniper et al have announced network fabrics for data centers with five thousand and more servers with and without storage enablement. Juniper talks of a one-tier fabric through their Project Stratus work with IBM to be available some time in the future. Brocade recently introduced its’ Brocade One, which is a converged data center fabric. Cisco just launched its’ FabricPath Switching System or FSS for the Nexus 7000 that enables massive scale of a two-tier fabric. In this Lippis Report Research Note, we review the architectural attributes of two and three tier network fabrics and review FSS and its accompanying F-Series 10GbE module.

The IT industry is at an inflection point as service delivery is becoming more and more centralized thanks to data center consolidation, virtualization, cloud and mobile computing. It is estimated that a third of all IT spend is concentrated in the data center and this trend is only building thanks to favorable economics, motivating IT business leaders to centralize IT delivery.

The impact of this trend is more and more dense data centers made up of servers in the thousands to tens of thousands and higher. It is at the scale of 5,000 plus servers that a new network fabric is required for high-end data centers. High-end data center design is challenged with increasing complexity, the need for greater workload mobility and reduced energy consumption. Traffic patterns have also shifted significantly, from primarily client-server or as commonly referred to as north-to-south flows, to a combination of client-server and server-server or east-to-west plus north-to-south streams. These shifts have wreaked havoc on application response time and end user experience, since the network is not designed for these Brownian motion type flows.

The main requirements for high-end data center network fabric are low latency, large flat layer 2 domains to enable workload mobility, low power consumption, simplicity of design and significant bandwidth. Storage enablement, meaning consolidated I/O or virtualized I/O, is a growing priority and a new fabric that can support FiberChannel over Ethernet, iSCSI over Ethernet, iWARP over Ethernet or Infiniband over Ethernet is a major plus. One salient observation is that it’s pretty clear that Ethernet is the network fabric of choice as it is the only network protocol that enjoys continual innovation such as TRILL, Data Center Bridging, link aggregation, multi-pathing, and soon, 40 Gbs and 100 Gbs speeds. With the above requirements in mind, let us review data center network design options.

Two and Three Tier Fabrics

A three-tier network architecture is the dominant structure in data centers today and will likely continue as the optimal design for many networks. For most network architects and administrators, this type of design provides the best balance of asset utilization, layer 3 routing for segmentation, scaling and services, plus efficient physical design for cabling and fiber runs. By three tiers, we mean access switches/Top-of-Rack (ToR) switches, or modular/End-of-Row (EoR) switches that connect to servers and IP based storage. These access switches are connected via Ethernet to aggregation switches. The aggregation switches are connected into a set of core switches or routers that forward traffic flows from servers to an intranet and internet, and between the aggregation switches. It’s common in this structure to over-subscribe bandwidth in the access tier, and to a lesser degree, in the aggregation tier, which can increase latency and reduce performance. Inherent in this structure is the placement of layer 2 versus layer 3 forwarding that is Virtual Local Area Networking or VLANs and IP routing. Also common, is that VLANs are constructed within access and aggregation switches, while layer 3 capabilities in the aggregation or core switches route between them.

But within the high-end data center market, where the number of servers is in the thousands to tens of thousands plus and east-west bandwidth is significant, is where a new structure is needed. It is within these data centers where applications need a single layer 2 domain.

Two-tiers of network fabric

A two-tier fabric is designed with two kinds of switches: one that connects servers, and the second that connect switches creating a non-blocking, low latency fabric. In short, there are server facing and fabric facing switches. We use the terms ‘leaf’ switch to denote server facing or connecting switches and ‘spine’ to denote fabric facing or switches that connect leaf switches into the fabric. Together, a leaf and spine architecture create the fabric.

Many IT leaders in Global 2000 firms will have deployed both two and three tier network structure, as different deployment models are used for different applications. For these leaders, a network equipment supplier is needed that possesses product architecture flexibility, meaning an end-to-end product solution that accommodates tier two and three fabrics. This flexibility is found in product that supports layer 2 and layer 3 forwarding, as well as, a variety of line cards to offer design options.

A common network Operating System (OS) of products configured for two and three tier structure is important as IT operations gain efficiency to manage fabrics, as configuration and management are consistent. In addition, a common network OS offers rapid absorption of innovation to IT operations, as new OS features are available at the same time to all fabrics. The benefit of using a common product set to build tier two or three fabrics offers value around operational efficiency, training, sparing and ease of evolution between fabric deployments. In short, the network fabric needs to be simple and general purpose versus purpose built, which a common set of products creating tier two or three fabrics offer. This type of flexibility will enable IT leaders to address the challenges of scale outlined above.

In addition to product flexibility, some networking suppliers take a systems approach to their fabric design, meaning that a solution is built and pre-tested before it arrives on site. This ensures that IT does not have to perform system integration. With the increased concentration of computing and IT dollars into data centers, it’s only obvious that data centers are long-term corporate commitments. Therefore, it is only appropriate that the networking supplier of choice also has a proven long-term commitment to their product architecture.

Perhaps the best example of this is Cisco’s Catalyst 6000 switching architecture and its’ two-year-old Nexus product line. The Catalyst investment protection is well documented as it has been in operation for over a decade, which Cisco customers enjoy continued innovation and value added to this platform. Competitors view its’ longevity as a weakness. The Nexus product line has a similar investment protection philosophy with a fifteen-year plus lifespan expectation. Common to both Catalyst and Nexus is the fact that these products are built on silicon, developed at Cisco, affording investment protection from one generation of the hardware to the next.

A Unified Fabric

The concept of a unified fabric is to virtualize data center resources and connect them through a high bandwidth network that is very scalable, high performance and enables the convergence of multiple protocols onto a single physical network. These IT resources are compute, storage and applications, which are connected via a network fabric. In short, the network is the unified fabric and the network is Ethernet.

The industry tends to focus on storage transport over Ethernet as the main concept behind a unified fabric with technologies such as Fiber Channel over Ethernet or FCoE, iSCSI over Ethernet, iWARP over Ethernet and even Infiniband over Ethernet. But this is a narrow view of a unified fabric, which is being expanded thanks to continual innovation of Ethernet by the vendor community and standards organizations such as the IEEE and IETF. Ethernet innovations such as FCoE, Data Center Bridging or DCB, link aggregation, Cisco’s VN-Link, FEX-Link and virtual PortChannel or vPC have enhanced Ethernet networking to support a wide range of new data center fabric design options. In addition to these protocol enhancements, the IEEE is scheduled to complete its’ work on defining 40Gb and 100Gb Ethernet during the summer of 2010, significantly increasing Ethernet’s ability to scale bandwidth. To demonstrate how Ethernet is evolving to be the unified fabric for high-end data centers, we explore Cisco’s new FabricPath Switching System innovation in this white paper.

The decision to implement a two or three tier network structure comes down to scale. For high-end data centers, a two-tier structure meets the requirements of low latency, movable workloads, scale, simplicity, etc. Many global 2000 concerns will have deployed both a two and three tier network fabric for their high end and less dense data centers.

When shopping for network equipment to construct two and three tier network fabrics, look for suppliers that support both rich Layer 3 routing services and scalable Layer 2 Ethernet capabilities to ensure choice and flexibility of three tier and scalable two tier fabric implementations. Such suppliers offer products that can be configured in multiple use cases and topologies where modules are inter-changeable, skills transferable and operations common between both fabric approaches.

Tweet

The Unified Communications (UC) market is changing significantly in terms of how solutions are deployed, services are packaged and systems procured. Siemens addressed all of these areas in one fell swoop with the recently released OpenScape UC Server 2010. Siemens announced new licensing options for endpoints and integrated UC endpoint services in packaged solutions including social media plug-ins such as Twitter. Kathy Heilmann, Director, Large Enterprise Voice and UC Solutions Marketing at Siemens Enterprise Communications joined me to talk about the new rules of UC procurement and its value to IT business leaders.

Listen to the Podcast

Tweet

Networking has become “rigid”. Yes, I know it’s almost absurd to attribute inflexibility or rigidity to networking, but we are in a compute innovation cycle that’s driving a fundamental change in networking, which screams out the need for more flexibility and configuration automation. The well understood problem is that when a virtualized machine is moved from one physical machine to another, the network, load balancers, firewalls/IPS, broadcast domains, etc., have to be reconfigured. There is no automation in place, meaning that the network is not flexible or agile enough to make the changes required. Networking companies such as Force10 Networks are driving an open approach to automation to enable network changes. I spoke with Steve Garrison, Vice President Marketing for Force10 Networks about their “Open Automation” approach to networking. Enjoy, Nick

Listen to the Podcast

Tweet

In Lippis Report 148 we reviewed the major drivers and trends that are propelling the high-end data center Ethernet switch market to well over a $1B annual run rate. In this Lippis Report Research Note, we review the major suppliers of these switches. We review Cisco, Arista Networks Force10 Networks, BLADE Network Technologies, HP/3Com/H3C, Voltaire, Avaya, Brocade, and Juniper and identify their unique positions and offerings to participants in the burgeoning market. Our focus is the high-end, high density 10GbE switches that are enabling virtualized cloud computing data centers thanks to Terabits per second of back plane switching capacity, billions of packets per second of layer 2/3 forwarding, hundreds of 10GbE port connectivity per chassis, a new two-tier architecture, microsecond level latency, low power consumption, non-stop operation and software hooks that eliminate network barriers to large scale server virtualization. The engineering in these switches should be celebrated, as they represent the state-of-the-art in computer and network design. In short, they represent the fundamental building block of a new generation of IT delivery based upon cloud computing and virtualization. This Research Note is a must read for any IT executive designing a data center.

After finishing this Research Note, it became evident that this market needs a set of industry neural 10GbE switch test to independently verify vendor claims. We hope to make such a contribution this Fall.

Cisco Systems Nexus Family of Switches

Cisco’s approach to data center Ethernet switching is rooted in its Data Center 3.0 strategy which seeks to scale server virtualization while introducing a platform to enable a unified fabric or converged network and storage running on one physical Ethernet network. Cisco’s data center Ethernet switch portfolio is primarily the Nexus family of switches including the 7000, 5000, 2000 and 1000v. NX-OS is a purpose built data center operating system that runs across the entire Nexus family. NX-OS integrates a number of higher system availability functionalities such as virtual port- channel (vPC), and the capability to upgrade software without disrupting traffic. The Nexus 1000v is a softswitch that resides in a VM hypervisor. The Nexus 1000v’s main job is to eliminate network configuration barriers that exist when moving a VM from one physical machine to another. To accomplish this, the 1000v creates a port profile including VLAN, ACL, policy, security, etc. with persistence, which moves with a VM as a virtualization administer moves a VM from one physical machine to another.

The Nexus 2000 family of Fabric Extenders (FEX) introduces the concept of a remote line card of the parent Nexus 5000 switches and sits on the top-of-rack connecting servers to the switch fabric. The extender concept allows the 2000 and 5000 to be managed as one switch. This configuration reduces cabling requirements and offers an economical approach to server connection, thus providing the benefits of both end-of-row and top-of-rack deployments. The Nexus 5000 Series is 10 Gb Ethernet and Unified Fabric capable switches, connecting Nexus 2000s and servers directly at 100/1/10GbE/FCoE, while providing layer 2 forwarding. Providing layer 3 forwarding, dense 1/10GbE connectivity is the Nexus 7000 Series. The Nexus 7000 Series is available in a 10 and 18 slot chassis and is Cisco’s flagship data center Ethernet switch series. As a point of reference, the Nexus 7000 is now on an annualized run rate of $1B for Cisco, which is more than 10 times greater than any other switch supplier in the data center switch market. The high end 7000 connects 512 10GbE ports with 128 line-rate 10 Gigabit Ethernet ports. The Nexus 7000 Series switches can be segmented into virtual devices, delivering true segmentation of network traffic, context-level fault isolation, and management through the creation of independent hardware and software partitions. Overlay Virtualization Transport (OTV) provides customers a simplified DCI solution by extending layer 2 VLANs over existing IP networks. We have profiled the Nexus 7000 when first released and is available here. The Nexus switches can create a two-tier architecture with the 2000/5000, providing server connectivity and layer 2 forwarding between servers. The Nexus 7000 connects the 2000/5000 to each other and the internet/intranet with high density, high reliability layer 2/3 forwarding.

Arista Networks 7500 Family of Modular Switches

Arista Networks is a new comer to the data center Ethernet market, but its management team is seasoned and customer base growing. It provides six fixed 10GbE switches; five 1/10GbE 7100 and the 1GbE 7048 along with the new Best of Interop awarding winning 7500 modular switch. The 7100/7048 switches connect servers in a Top-of-Rack configuration while the 7500 aggregates these switches and connects them to the internet and intranet. This is a two-tier, “leaf-spine” architecture. The 7500 boasts ultra high performance layer 2/3 1/10 Gb Ethernet switching for high performance computing and cloud computing data centers. The 7500 supports 384 10GbE ports, 5.7Bpps at layer 2 or 3, high packet buffers 18GB deep, ultra low port-port latency of 4.5 microseconds and 10Terabit loss less switch fabric connecting modules.

The 7500 is 10GbE port dense, compact, cloud spec fast, green and prepared for 40 and 100GbE, with a price tag 50% below competitive offerings, according to Arista. While the 7500’s hardware architecture is impressive, its operating system EOS, Extensible Operating System, offers another set of uniqueness. For example, all Arista switches run the same binary image of EOS, easing administration while hastening switch feature upgrades. EOS is a modular OS that allows partners to run their software in the Arista switch, consolidating the number of management and network appliances required, thus increasing performance while reducing energy consumption and physical space. Arista’s EOS modularity was designed as a unique state sharing architecture that separates switch state from protocol processing and application logic. EOS is built on top of a standard Linux kernel. All EOS processes run in their own protected memory space and exchange state through an in-memory database. This multi-process state sharing architecture provides the foundation for in-service-software updates and self-healing resiliency. You can listen to a podcast interview with Douglas Gourlay, VP Marketing and Anshul Sadana, VP Customer & Systems Engineering from Arista on the introduction of the 7500 Series of Ethernet switches here

HP/3Com/H3C’s A12500 Core Data Center Switches

HP has spent 25 years building and selling networking products to its worldwide client base and is currently #2 in the market, with a 21% port count share and the fastest growing networking company in the industry. The combined HP/3COM acquisition brings core switching products, the #1 market share position in China, TippingPoint Intrusion Prevention System and ProCurve edge switches, representing a new choice for clients who are frustrated by today’s current offerings. HP will combine these two entities and operate under the banner of “HP Networking.”

The HP Converged Infrastructure Architecture and FlexFabric blueprint approach the modern data center with a vision that places networking at the center of an integrated data center solution and accelerates deployment of enterprise services and applications. It is designed to drive simplicity through streamlined network designs and centralized management, enhance agility with high performance security, and accelerated provisioning, and reduce cost with energy efficiency and low total cost of ownership. Central to HP FlexFabric is policy-driven network provisioning tightly integrated with server and storage management in an end-to-end data center converged infrastructure.

HP data center solutions are purpose built, using the latest advanced systems and ASIC technologies. “A” family data center networking platforms leverage a common operating system, Comware™ and are managed with a single-pane manager, Intelligent Management Center (IMC). HP switches make use of an HP-developed technology – Intelligent Resilient Framework (IRF) – to create a resilient virtual switching fabric. IRF delivers geographic independence, distributed high-availability, resiliency and millisecond re-convergence across layer 2 and layer 3 protocols. These innovations allow customers to build a simplified, high performing, highly resilient and flat (two-tier) data center network design. They overcome the limitations of low performance/scale, high cost/latency inherent in legacy solutions, which rely on multi-tier network designs, disjointed platform operating systems and complex resiliency protocols.

A key enabler of this transformational design flexibly is the HP next-generation data center switching architecture. This starts with the flagship HP A12500 core data center switch – which is based on a 100G design that uses a multi-level, multi-plane, non-blocking switching architecture to provide high performance and scalability. The A12500 supports 6.66 Tbps of high-performance switching capacity (future support for 13.32 Tbps) and scales to 2.2 billion packets per second of forwarding performance. The A12518 supports 512 10 Gigabit Ethernet or 864 Gigabit Ethernet ports in a single chassis. Its future-proof design accommodates 40/100 Gigabit Ethernet and emerging unified network requirements such as end-to-end FCoE/Data Center Ethernet.

Force10 Networks ExaScale E Series

Force10 Networks was one, if not the first company to offer 1 and 10Gb switching solutions for high-performance computing and data center markets in Fortune 100 companies, Internet portals, global carriers, leading research laboratories and government organizations. It offers a wide range of Ethernet switching and routing products that deliver high port density and resiliency to help customers deploy a high-availability, agile and standards-based GbE and 10 GbE network fabric, while reducing power and cooling costs. Its Ethernet switching products are designed to leverage virtualized data center environments and automate Ethernet networking. For example, its VirtualScale enables management of virtual chassis. Its VirtualControl enables virtualizing logical switching and routing boundaries. For automation, Force10 has developed an architecture, which automates network resource allocation as applications and services spin up and down. This architecture is built upon its HyperLink and SwitchLink technology, two new software features implemented within its Force10 Operating System (FTOS). HyperLink provides real-time communication between Force10 switches and hypervisors or virtual switches to enable automatic provisioning of one or many virtual LANs (VLANs) across multiple switches simultaneously. The SwitchLink feature provides real-time communication with middleware orchestration tools to enable automatic provisioning and management of virtual devices anywhere in the network.

Force10’s modular Ethernet switch data center product portfolio includes the ExaScale E-Series, optimized for core deployments in large-scale, high-performance 10GbE data centers, and the C-Series, optimized for mid-range data centers. Both the E-Series and C-Series come in multiple form factors, run FTOS and are dense high performance switching platforms equipped with redundancy, availability, fault-tolerant operations and many line card options. In addition, Force10 offers the fixed configuration S-Series product line for GbE and 10 GbE ToR configurations. Force10 promotes a vision of simplified data center topologies, using integrated switching and routing in the core, using chassis based E-Series or C-Series products, and fixed configuration ToR access products allowing both 1 tier and 2 tier designs. One tier can be achieved with high density E-Series platform for server aggregation, switching at the server edge, and routing off the same platform to the Internet / WAN. The two-tier architecture can be achieved leveraging ToR switching for server aggregation along with Force10’s chassis based systems in the core. In addition to a large direct sales force, IBM OEM’s Force10’s ExaScale platform as part of IBM’s iDataPlex clustering solution. You can listen to a podcast interview with Steve Garrison, VP Marketing of Force10 on their 40 GbE offering here.

BLADE Network Technologies RackSwitch Family of Ethernet Switches

BLADE Network Technologies (BNT) has been working in the data center switch market since 2006 with much success providing 1/10Gb Ethernet switches for blade servers and top-of-rack configurations. BLADE was launched from Nortel and made up of the successful Alteon Networks group. Their success stems from their ability to identify the top-of-rack and blade switch market in ’06, along with an OEM go to market strategy that included all of the top tier blade server providers such as HP, IBM and NEC. The result is that BLADE has shipped over 8m ports, achieved 25% growth from 2008 to 2009 (in a down economy), owns 50+ % of the blade switch market, is number 3 in the Fixed 10GbE market according to Dell’Oro Group, and has demonstrated scale with at least one customer installing over 16,000 of its switches.

BLADE offers the RackSwitch family of Ethernet switches, which are ToR, 1U high switches. They include the 24-port 360ns latency RackSwitch G8100 10GbE, 48-port RackSwitch G8000 1/10 GbE aggregation and the 24-port 700ns latency RackSwitch G8124 10GbE. Over a year ago, BLADE released its virtualization software called VMready that automates network settings for VM movement ensuring that network settings migrate when a VM is moved from one physical server to another. VMready scales to a 1000 virtual port switch, is based on standards and works with most popular hypervisors.

In addition to VMready, RackSwitch’s unique attributes are found in the fact that they were designed for the data center versus being a wiring closet switch re-formatted for the data center. For example, the RackSwitch BLADEOS supports CEE for unified fabrics, uplink failure detection, virtualization, dual homing for servers, low (80-170Watts) power consumption, back-to-front or front-to-back airflow and very low latency in the 700-360 nanosecond range.

Voltaire’s Vantage 8500

Voltaire has a long history in high performance computing and data center networking as it is one of the key leaders in the InfiniBand market. Voltaire enjoys distribution relationships with HP and IBM, as well as Bull, Fujitsu, NEC, SGI and Oracle. The result is a 100% + year over year revenue growth for Q1 as reported on May 5th. Last October, Voltaire entered the 10 GbE market with the introduction of its Vantage 8500 Ethernet layer 2-core switch. The Vantage 8500 boasts less than 1 microsecond of latency, a low 10 watts per port power consumption and 288 wire speed 10GbE ports in a 15U high chassis. The Vantage 8500’s unique industry contribution is that it’s based on converged enhanced Ethernet (CEE) technology providing InfiniBand-like capabilities to the Ethernet data center. In fact, Voltaire has ported many of InfiniBand’s key characteristics to the Vantage 8500 such as a lossless switching fabric, multi-pathing, virtualization, fabric-wide congestion management and QoS.

From a network design point of view, Voltaire supports a two tier network architecture that enables a simplified, ‘flat’ data center network and puts an end to the era of the over-provisioned network. Voltaire’s design centered on the Vantage 8500 is to support a two-tier data center network that scales from hundreds to a few thousand core ports, which requires high capacity, non-blocking 10 Gigabit Ethernet core switches. By clustering up to twelve Vantage 8500 switches together, IT business leaders can expand their data center to many thousands of servers while preserving the efficiency and price-per-port, without degrading performance or latency which occurs in traditional hierarchical network designs. To support ToR implementations, Voltaire and BLADE Network Technologies announced recently a partnership where BLADE ToR RackSwitches are aggregated by Voltaire’s Vantage 8500, rounding out the two-tier data center Ethernet network architecture.

The Vantage 8500 also features software-based capabilities to address virtualized and converged data center environments. Voltaire’s Unified Fabric Manager™ (UFM) software, application acceleration software and management OS (VT-OS) provide management and performance enhancement tools. These tools were developed and optimized in InfiniBand environments and are now available for Ethernet-based data centers. Voltaire’s recently introduced Unified Fabric Manager™ (UFM™) 3.0 software orchestrates physical and virtual switches delivering guaranteed levels of service per application. It’s the first and only Ethernet fabric management software that dynamically orchestrates end-to-end virtual machine connectivity for multi-vendor, scale-out data center networks.

Avaya’s VSP 9000

During the April 2009 Las Vegas Interop trade show, Nortel committed to the data center Ethernet market with the announcement of its Virtual Services Platform or VSP 9000 switch, which supports up to 27 Terabits per second (Tbps) of backplane switching and 240 10GbE ports per chassis at first release. Avaya announced their commitment to the VSP 9000 and said that it will be generally available in the second half of 2010 while already in controlled availability. The VSP 9000 is built upon the Ethernet Routing Switch 8600/8800 software providing a proven software foundation, mid-plane architecture, a fully programmable network processor unit for flexible data forwarding and carrier-grade Linux.

The VSP 9000 is designed to deliver high-density 10GbE, 40GbE and 100GbE. Its design center is rooted in highly dense connectivity environments that are all mission critical, by definition. Early testing validation of the VSP 9000 promises to provide ultra-high reliability and availability delivering below 50ms failover support, which is critical to eliminate application disruption thanks to its patented hardware failure detection differentiation. The VSP 9000 switch fabrics are lossless Ethernet capable and therefore well positioned to support the next generation Data Center requirements for convergence of storage onto the Ethernet infrastructure.

The VSP 9000’s unique network architecture is found in its ability to cluster four switches together, in that the total architecture exceeds 100 Tbs, with the number of 10GbE ports per rack being up to 720. Avaya continues to invest in Switch Clustering technology (Active/Active resiliency model) such as SMLT (split multi-link trunking) and RSMLT (routed-SMLT), which provides link, switch and router redundancy mechanisms. Three modules are being introduced in the first VSP 9000 release, a 24 port SFP+ for 1 GbE and 10 GbE connectivity, a 48-port of SFP module in addition to a 48-port 10/100/1000 TX module. Future plans include 40GbE and 100GbE interfaces, and even higher-capacity Switch Fabric modules.

Juniper Networks’s EX8200 & EX4500

In January of 2008, Juniper Networks launched its much-anticipated entry into the enterprise Ethernet switch market. Juniper’s focus is on the enterprise data center, campus and branch, as well as the service provider market. Juniper provides a suite of Ethernet switch products, including the EX4200 with Virtual Chassis technology for GbE Top-of-Rack (ToR) and End-of-Row (EoR) data center access, the EX2500 24-port and new EX4500 48-port 10GbE ToR switches, and the EX8200 high-density, high-performance line of modular Ethernet switches.

According to Juniper, it simplifies customer enterprise LAN architectures and advances the economics of networking via its most recently launched initiative called the “new network” for data centers. Juniper’s “new network” promises critical innovations in automation, virtualization and fabric technologies. These innovations are to reduce time to operation by up to 50 percent and eliminate up to 35 percent of data center networking capital expenditures. One aspect of the “new network” is a simplified two-tier network architecture, which may be reduced to one when “Project Stratus” is completed with IBM. The reduction of a three-tier architecture to two is accomplished by utilizing Juniper’s Virtual Chassis fabric technology in the access layer, in conjunction with its high-density, high-performance platforms such as EX8200 and EX4500 in the LAN core, thus eliminating the aggregation or distribution layer. According to Juniper, collapsing the distribution layer reduces complexity in the data center as well as campus networks by reducing the number of managed devices by up to 89%, providing up to 39% savings in space, 44% savings in power and reducing the number of switch interactions by up to 99% compared to three-layer networks. According to Juniper, this approach improves application performance by also reducing latency up to 77% compared to three-layer networks. Note that these claims and numbers are Juniper’s and not mine.

At the core of Juniper’s data center Ethernet product family is the EX8200 line of modular switches. The EX8208 and EX8216 are eight and sixteen-slot modular switches. The EX8216 sports a maximum of 640 10GbE ports and 1.92Bpps and 6.2Tbps backplane speed. The EX8200 is said to support 40GbE and 100GbE interfaces in the future. The EX8200s connect either EX4200 GbE or EX2500 and EX4500 10GbE ToR switches together while providing access to internet/intranet. All Juniper switches run Junos, the network operating system that provides reliability and availability features, developed for the high-performance enterprise and service provider market.

Brocade’s NetIron MLX Series of Switches

In July of 2008, Brocade had purchased Foundry Networks, catapulting them into the Ethernet switch market as one of the top five Ethernet switch/router vendors by revenue. Brocade, with its long history of data center storage, saw that converged I/O was going to happen and prepared the company to participate in this market. At the high end of Brocade’s data center Ethernet switch products is the NetIron MLX-4, MLX-8, MLX-16 and MLX-32 routers, which support 4, 8, 16 and 32 I/O module slots, respectively. We’ll focus on the high end NetIron MLX-32 here, which has been in production since August 2006.

The NetIron MLX-32 boasts a total of fully redundant non-blocking 7.68 Tbps switch fabric capacity. Brocade says that the MLX-32 can forward some 2.284 Bpps of Layer 2/3 packets and support 1,536 and 256 non-blocking 1 GbE and 10 GbE ports, respectively. Note that the new high density 10 GbE was announced the same day as this Research Note was made public. All four NetIron MLX systems are designed for non-stop operation, supporting 1:1 management module redundancy, N+1 switch module redundancy, M+N power module redundancy and N+1 fan redundancy. The NetIron MLX architecture is an adaptive self-routing Clos switch fabric with a virtual output queue (VOQ) design. This non-blocking architecture is optimized for maximum throughput and low latency for all packet sizes.

Tweet

While the networking industry is full steam ahead with the transition from 1 to 10 Gb Ethernet in corporate uplinks plus virtualized and cloud spec data centers, the IEEE has been working on 40 and 100Gb Ethernet standards to be ratified soon. It’s anticipated that the 40GbE standard will be completed first and built with lower cost, long-range optical components than 100GbE. The question on the minds of most IT business leaders is when and where to deploy these ultra high-speed Ethernet technologies and at what cost? We answer these questions with Steve Garrison, Vice President, Marketing of Force10 Networks. We’ll dive into Force10’s 40 GbE leadership position and the new network design options it unleashes upon IT business leaders.

Listen to the Podcast

Tweet

By Arista Networks

Extensibility is a system design principle where the implementation of the operating system takes into consideration future growth. It is a systemic measure of the ability to extend the operating system and the level of effort required to implement the extension. Extensions can involve the addition of new functionality or the modification of existing functionality.

Find out how Arista’s Network Operating System EoS improves system uptime and delivers rapid service restoration in the event of failure by downloading this whitepaper.

Get the White Paper

Tweet

By Cisco Systems

Virtualization is rapidly becoming an essential tool for more fully harnessing and managing the power of today’s data center servers. In only a few years, standard x86 server technology has increased in performance and density so that today, multisocket, quad-core systems with 32 or more gigabytes of memory, are the norm. The combination of multicore computing and virtualization software such as VMware Virtual Infrastructure has enabled IT departments to bring server sprawl under control by running multiple independent workloads on a smaller number of servers. Today, fewer servers are required to do the same work, and their utilization levels have increased — both factors that contribute to greater energy efficiency and lower power and cooling costs.

As IT departments have discovered the benefits of server consolidation, they have also found that virtualization solves an even broader set of problems. Business continuity plans based on virtualization can make disaster-recovery solutions simple, reliable, and more cost effective. Virtual desktop environments can use centralized servers and thin clients to support large numbers of users with standard PC configurations that help to lower both capital and operating costs. Virtualization allows development, test, and production environments to coexist on the same servers, and it helps decouple application deployment from server purchasing decisions. New applications can be deployed in virtual environments and scaled on demand to accommodate the evolving needs of
the business.

Get the White Paper

Tweet

By Arista Networks

Can Ethernet compete with Infiniband in the low-latency trading and high performance computing markets? Administrators and IT professionals face a choice when deciding whether to invest in Infiniband or Ethernet for their low-latency networks. This paper addresses many of the characteristics of Infiniband that have made their way into Ethernet.

To learn more, download the Arista Networks whitepaper.

Get the White Paper

Tweet

Our industry is in a compute innovation cycle thanks to virtualization and cloud computing and it’s changing fundamental networking requirements and design. These changes are beyond increased packet processing performance and ultra low latency. They extend to network design and product features that enable increased server virtualization scale, workload mobility and cloud computing. I discuss a new model for networking born out of data center virtualization and cloud computing with Doug Gourlay, Vice President Marketing at Arista Networks. This is Arista Networks’ first podcast; sure to be a classic.

Enjoy, Nick

Listen to the Podcast

Tweet

By Voltaire

This paper describes the challenges inherent in existing Ethernet solutions and how Voltaire’s new scale-out Ethernet architecture effectively addresses those challenges. Voltaire’s InfiniBand products and fabric management solutions have been addressing Data center architecture changes for years. Today, Voltaire delivers the fabric for the world’s largest supercomputers, the world’s fastest financial trading platforms, and the world’s most scalable database machines. Soon, Voltaire will extend its product offering to include Converged Enhanced Ethernet CEE switches and software, allowing end users with less demanding performance requirements to benefit from a far more scalable Ethernet fabric that also lowers overall fabric costs, lowers power consumption, has greater efficiencies, and simplifies management.

Find out how by downloading this white paper.

Get the White Paper

Tweet

Networks today require predictable performance and are much more aware of application flows than traditional networks with static addressing of devices. Enterprise networks in the past were designed for specific applications while new cloud designs in the data center can address a multitude of applications. This is clearly a radical departure from today’s oversubscribed networks in which delays and high transit latency are inherent.

To learn more, download the Arista Networks whitepaper.

Get the White Paper

Tweet

Arista’s new 7500 Series of Ethernet switches is touted as the fastest Ethernet switch on the planet. Arista Networks has been delivering ultra high performance fixed 1/10 Gb Ethernet switches for high performance computing and cloud computing data centers. But this week Arista shocked the industry by introducing a massively powerful Ethernet switch platform that is 10 GbE port dense, compact, cloud spec fast, amazingly green plus smart and prepared for 40 and 100GbE with a price tag 50% below competitive offerings. I talk with Douglas Gourlay, Vice President Marketing and Anshul Sadana Vice President, Customer & Systems Engineering both from Arista Networks about a new age of network design in the cloud-computing era.

Listen to the Podcast

Tweet

Power over Ethernet (PoE) powers IT endpoints like IP phones and WLAN Access points. But newer devices like IP video phones, IP surveillance cameras, thin client display, next Generation IP phones, 802.11n WLAN access points and Pan Tilt Zoom (PTZ) Surveillance cameras require more than 15W of power so the IEEE recently standardized a new 30W PoE standard called PoE Plus. Soon most, if not all, corporations will have a mix of end points that require old PoE and new PoE plus power. IT leaders can meet these requirements by leveraging the enhanced PoE and PoE Plus features offered by Cisco’s Catalyst 4500 E-Series product line. Moreover Cisco has announced two new Catalyst 4500 Series Line Cards with readiness for PoE Plus and inline power up to 30W per port. I talk with Sachin Gupta, Director and Product Manager at Cisco Systems about these new PoE design features and what new options are available to IT architects as they build out corporate collaborative solutions.

Listen to the Podcast

Tweet

By Arista Networks

What is Arista’s definition of scalability of a data center network? What are the critical requirements and what issues must be solved to address data center scalability? Scalability of the data center network is the ability for network technologies to accept increased traffic or new devices without impacting the contribution margin. In this white paper Arista Networks discuses network design best practices for scaling up Data Center Networks.

Find out how by downloading this paper.

Get the White Paper

Tweet

A panel of IT business leaders discusses their experience with Network Virtualization as we dive into motivations, design options, economics and business outcomes. On the panel is Marilyn Hay, Manager of the Network Management Centre at the University of British Columbia, Frank Hoonhout, Senior Lead Network Engineer at the State of Oregon’s State Data Center and Hasan Siraj, Director of Product Marketing at Cisco Systems. This is a podcast you surely want to listen to.

Find out the real world value and business outcome of investing in Network Virtualization by listening to this podcast.

Listen to the Podcast

Tweet

By 3Com/H3C/TippingPoint

The same general benefits that enterprises derive from server and application virtualization—more efficient resource utilization, driving down deployment and management costs, and reducing network complexity— can also be realized through virtual networks. Additionally, network security components such as Intrusion Prevention Systems (IPSs) and firewalls can provide more sophisticated policy enforcement in the network fabric. Most enterprises have by now at least started to take advantage of data center virtualization to drive down the cost of application deployment and more efficiently use server resources. While the first step of virtualization usually happens in the application server, enterprises should also be thinking about ways to reduce hardware costs and management complexity by taking advantage of the same virtualization concepts in the design of their data center and campus networks.

Find out how by downloading this white paper.

Get the White Paper

Tweet

Hear how virtualization has spread beyond the data center into the core network to increase utilization, security, and functionality.

Listen to the podcast here

Visit the Link

Tweet

By Arista

The standard for 10 Gigabit Ethernet (IEEE802.3ae) was ratified in 2002. While 10GbE deployments have grown every year since then, the technology has primarily been used to interconnect switches and routers. Almost all of the server connections in data centers have remained at 1 Gbps, limiting the amount of network throughput available to each server. With recent enhancements in CPU performance, system I/O, and storage I/O the gigabit network has increasingly become the application and workload performance bottleneck.
The primary reason for staying with Gigabit Ethernet has been cost-performance. Until recently it has been more cost-effective to have multiple GbE connections rather than a single 10 GbE port. In addition, most installed servers typically cannot utilize the full bandwidth of a 10 GbE connection. However both of these factors are changing, which are leading to widespread adoption of 10 GbE for server connectivity over the next few years.

Download this white paper for an overview of the factors that are driving the growth for 10 GbE in the data center.

Get the White Paper

Tweet

By Cisco Systems

Cisco IT is transforming its data centers with solutions that help to realize the company’s Data Center 3.0 vision, which employs a unified network fabric to connect servers and storage devices in a way that is resilient, scalable, and easy to manage. The transformation occurs in three stages: 1) Consolidating I/O and increasing throughput by implementing unified I/O running on 10 Gigabit Ethernet (current stage); 2) Increasing the power available to compute resources by reducing the power consumed by the network infrastructure and; 3) Making applications location-independent, which will simplify changes and possibly eliminate the need for change requests

Learn how Cisco is deploying consolidated I/O in their data center by downloading this paper.

Get the White Paper