Lippis Report 197: Cisco’s Unified Fabric for Physical, Virtual and Cloud Environments

Over the past several years, the term “fabric” has been used to describe a weave of computing, storage and networking that make up the foundation of data center infrastructure. Many networking firms use the term “fabric” to describe converged networking and storage. But today’s data centers are built with a combination of physical, virtual and cloud networking elements that need to interoperate and span multiple data centers. A modern fabric needs to incorporate all of these forms of networking consistently to deliver flexibility through a wide range of design options. Cisco Systems has one of the most expansive fabric definitions that is backed up by over 36,000 customers who deployed NX-OS. Cisco modifies the term “fabric” with the word “unified” to describe a functionally-rich fabric, yet control is simplified through integrated management. In this Lippis Report Research Note, we review Cisco’s Unified Fabric from business driver, architectural structure and outcome point of views.

Smarter Networking Webcast: Get Ahead of the Game with Virtualization and Edge Virtual Bridging

Visit the Link

Cisco defines its Unified Fabric architecture with three pillars of functionality, including 1) scale of performance and geographic span, 2) virtual machine (VM) networking plus workload mobility intelligence including network services such as L4-7 plus security, and 3) a converged wire once LAN/SAN structure managed with a single pane of glass that consolidates devices. In short, Cisco’s Unified Fabric promises scale, intelligence and convergence across physical, virtual and cloud infrastructure consistency. Architectural flexibility is a key design center for Cisco’s Unified Fabric so as to offer specific or unique designs to support a very large set of business requirements independent upon vertical industry, application and/or data center type. Unified Fabric is not a general purpose fabric, but a fungible fabric that can be weaved and wrapped around unique or diverse data center requirements.

Broadcom and VMware Come Together

Visit the Link

So why is it important for a fabric to be flexible enough for IT architects to be able to design unique data center solutions? It comes down to being able to meet business requirements, demands and goals at speed. Or put another way, it’s time and money; that is the time to innovate and the budget to deploy. But data center infrastructure and economics, unfortunately, often get in the way. Typically 70 to 80% of IT administrator time is consumed by maintenance activities; yes that’s 70 to 80%. This includes managing existing assets that are siloed, adds/moves/changes to optimize resource utilization plus dealing with security vulnerabilities and application restraints. When 80% of IT resources are dedicated to “keeping the lights on,” there is little time left for innovation that benefits the business.

Cisco on VXLAN For Logically Isolating Virtual Subnets At Scale

Listen to the Podcast

The data center network sits at the core of IT and is key to how IT can deliver service and value back to the business. But is your network ready to help in the huge transformation taking place in IT? IT is moving from a cost center to IT as a Service or ITaaS. This is a strategic transformation that helps businesses become more agile and evolve quickly by rebalancing the economics of the data center; that is drastically reduce the 80% of operations so that innovation can occur. If IT moves from its current methodology of addressing business needs with dedicated IT infrastructure to addressing them with more efficient, automated on-demand services, IT offerings can be more consistent, timely and responsive to business needs.

Cisco’s Familiar Approach To A Unified Multi-Protocol Storage Network

Listen to the Podcast

Every IT organization has its own journey toward ITaaS. Each phase usually expands into a series of key initiatives, such as consolidation, virtualization, cloud-based services for end-to-end delivery, mobility, elasticity, etc. One key IT time saver is achieved by enabling secure, efficient mobility of VMs to attain the agility that comes from automation and orchestration in the data center. In addition to consolidation, virtualization and cloud initiatives, IT business leaders are challenged to implement initiatives such as business continuity, disaster recovery, virtual desktops, fast application rollout, adding new data centers quickly and seamlessly, etc. Cisco’s Unified Data Center includes Unified Compute, Unified Fabric and Unified Management, and provides a holistic approach to enabling IT organizations move toward ITaaS. Its Unified Fabric is at the center of this strategy, delivering much of its value.

The Benefits of Building Networks With Shortest Path Bridging

Listen to the Podcast

Cisco’s Unified Fabric provides the foundational connectivity for physical, virtualized and cloud-based data centers. It unifies storage, data networking and network services, which is well beyond the fabric offerings of all other networking concerns. It consists of Nexus and MDS portfolios, along with Layer 4 to 7 solutions, including Virtual Wide Area Application Services or vWAAS for WAN optimization, VSG or virtual security gateway for security of in and out of line traffic flows, Application Control Engine or ACE for load balancing, and Network Analysis and Monitoring or NAM for application visibility. Unified Fabric includes the largest virtual networking product portfolio, including the Nexus 1000v, a virtual switch, virtual network service data path or vPath which is aware of L4-7 policies associated with individual VMs, the above virtualized L4-7 services, etc. Unique to Cisco’s Unified Fabric is NX-OS, a common network operating system, plus its Data Center Network Management or DCNM, an integrated, single pane of glass, management platform between MDS and Nexus that spans virtual, physical and cloud-based networking and data centers.

Guide to Cloud-Ready Branch Office Networking Is Your Wide Area Network Cloud Ready?

Get the White Paper

The value proposition of scale, convergence and intelligence of Unified Fabric is delivered through the combination of MDS, Nexus plus L4-7 services. Unified Fabric offers scale beyond switch performance to system scale that is geographic independent. This is fundamental for business and IT agility as it’s achieved via a flexible and highly-available secure fabric that supports dynamic resource allocation, changing traffic patterns, complex workloads and their mobility within and across data centers

Unified Fabric offers unique scale technologies such as its FabricPath, which is an active-active network protocol that eliminates spanning tree so that all links connecting Nexus devices are fully utilized. With FabricPath, the interconnection of Nexus 7000/5000s can scale a data center to tens of thousands of 10GbE ports. In addition to data center scale, Unified Fabric has a range of options to extend data centers over the wide area, including OTV or Overlay Transport Virtualization, a Layer 2 transport mechanism. It also includes LISP or Locator/ID Separation Protocol that has multiple purposes, including path optimization to connect data centers over the wide area as well as the highly desirable VM/application mobility it delivers by separating IP addresses from physical ports.

Gain the Advantages of Cloud Computing and Get Your WAN Ready for the Cloud

Get the White Paper

Most network fabric definitions start and end with the convergence of storage and Ethernet networking. This is one of Unified Fabric’s value proposition points as it enables a converged fabric, thanks to the MDS and Nexus product lines that support all of today’s storage protocols. Financial efficiencies and investment protection is achieved through consolidation, multi-protocol solutions and single point of management for LAN and SAN enabling an evolutionary adoption without disruption to existing infrastructure and operations. Fiber Channel over Ethernet (FCoE) simplifies the data center network by converging LANs and SANs over a single lossless Ethernet network providing a “wire-once-connect-anything” approach. It reduces network hardware sprawl through consolidation of Ethernet and SAN switches, consolidates LAN and SAN cabling onto a single Ethernet cable, simplifying data center management while reducing overall ownership cost including Capex and Opex expenses.

The Changing Face of Data Centre Networks

Get the White Paper

Cisco’s multi-protocol storage initiative brings all storage protocols into its Unified Fabric. Just like the days of multi-protocol routing, IT managers were able to manage the transition to IP networking after routing supported multiple network protocols such as DECnet, AppleTalk, etc. In short, applications that relied upon vendor specific protocols were supported and thus the transition to IP was regulated by how fast the application could support IP. The same is true in modern data centers; build a Unified Fabric that is capable of supporting multi-protocol storage, and IT business leaders can transition to a single fabric with fewer storage protocols over time and immediately gain simplification, lower life cycle cost and faster application deployment.

Broadcom Smart-Buffer Technology in Data Center Switches for Cost-Effective Performance Scaling of Cloud Applications

Get the White Paper

In addition to scale and convergence, Unified Fabric encompasses L4-7 plus virtualization aware intelligence, thanks to the Nexus 1000v that simplifies IT operations. Operations are simplified through embedded virtualization-aware policy-based security that’s applied consistently across physical, virtual and cloud environments. This consistency of applying L4-7 services across physical, virtual and cloud environments enables IT organizations to deliver network services to servers, storage and applications, at lower cost and at speed.

This intelligence comes into play in many scenarios, but is especially noticeable during VM moves. When VMs are moved with or between data centers, the Unified Fabric accomplishes a vMotion initiated VM move with all its security policy intact and without network reconfiguration, thanks to the embedded intelligence within the fabric such as ASA, LISP, vPath, VXLAN, etc.

Spring 2012 Open Industry Network Performance And Power Test Report

Get the White Paper

Cisco’s Unified Fabric focus is to offer a broad fabric solution with a simplified operational model via integrated management, network services that automate key aspects of fabric management plus consistent policy configuration, etc., across physical, virtual and cloud deployment scenarios. Adoption numbers are helpful to gage success. For example, over 36,000 firms have deployed NX-OS, with many extending their data center fabrics via OVT and scaling up their fabric with FabricPath. And from a market share by revenue point of view, during Q4 of 2011, Cisco led in data center Ethernet switching and FCoE SAN switching with 72.4% and 84.8%, respectively.

And those that deployed Unified Fabric have gained its advantages too. For example, 15 to 75% reduced infrastructure cost through converged LAN/SAN was realized by Almaviva and St. Joseph Hospital. The HayGroup and University of Salford were able to deploy applications faster by as much as 50 to 80%, which reduced deployment times from months to hours. Apollo Hospitals were able to increase the network diameter by a factor of two while keeping headcount the same.

And while the above numbers are impressive, one aspect of Unified Fabric deserves amplification and that’s its architectural flexibility. As Unified Fabric is deployed, the range of new IT services available increase, thanks to its dense engineering of a wide range of protocols and network services; services that include business continuity, disaster recovery, virtual desktops, faster application rollout, adding new data centers quickly and seamlessly, etc.

For example, consider data center disaster recovery or DCDR. As Unified Fabric includes OTV and supports ACE, Nexus 1000v, VSG, VXLAN, LISP, SAN, etc., these technologies come into play to support DCDR. OTV provides the basic Layer 3 connectivity between data centers so that workload can be moved from one to another data center during a disaster recovery. ACE provides dynamic resource elasticity between data centers, while the Nexus 1000v and VSG provide migration for virtual networks across data centers. LISP enables workload to be moved without re-provisioning IP addressing. From a storage perspective, Cisco’s MDS 9000 Family I/O Accelerator (IOA) feature improves the performance of SAN Extension to meet business continuity and disaster recovery requirements. All of the above technologies increase business productivity and agility and also come into play during DCDR. That is, in essence, the beauty of a Unified Fabric is its architectural flexibility to respond to demanding and changing business requirements that become critical during disasters too.

All fabrics are not created equal; Cisco’s Unified Fabric offers the widest range of design possibilities and options while enabling ITaaS.

Leave a Reply

You must be logged in to post a comment.