Cisco Debuts FabricPath

 
 
By Cameron Sturdevant  |  Posted 2010-08-06
 
 
 

Cisco is turning on yet another new feature in the Nexus 7000 platform that basically eliminates the ancient and universally deployed spanning tree protocol from network designs, thus enabling much larger Layer 2 networking domains within the data center. The FabricPath Switching System (FSS) is the foundation of a push by Cisco to combine NX-OS software features with the hard capabilities of the Nexus 7000 F1 series module to build the next generation of high-performance network systems.

Make no mistake, unlike incremental feature changes Cisco previously released for the Nexus 7000 platform including OTV (Overlay Transport Virtualization), FabricPath is a much bigger jump up in terms of bandwidth, network performance and virtualization feature support. At the same time Cisco went out of its way to ensure that FabricPath-enabled Nexus devices and software features wouldn't interfere with existing network implementations. Thus, while FabricPath is a radical departure from network designs built around the spanning tree protocol, it should be on the "evaluate now" track for large data center managers.

When used together, FabricPath-enabled modules will benefit organizations that want to implement workload mobility such as VMware vMotion in a much larger resource pool, while also gaining significant bandwidth and N+1 redundancy and fast network convergence after a link failure.

During prerelease conversations with Cisco engineers that included a demonstration of the first-generation release, the performance numbers were pretty astounding. FabricPath-enabled modules use active-active links between devices--losing the line blocking that spanning tree required--and instead use ECMP (equal-cost multipath) to direct traffic. First-generation modules can support up to 16-way ECMP, which can be combined with 16-port 10G bps PortChannels, for a total of 2.56T bps between switches.

When I looked at Cisco's OTV  feature of the Nexus 7000 in February of this year, I concentrated on Cisco's extension of Layer 2 functionality to enhance data center interconnect. OTV was still dealing with the limitations imposed by spanning tree protocol. FabricPath builds on technologies that were involved in creating OTV and rachets up the performance and simplification case for enterprises moving to the Nexus 7000 and NX-OS.  

At the time, I wrote, "the measure that data center managers will need to use when evaluating OTV is how much labor and network implementation effort...will be eliminated relative to the cost of putting in Cisco's hardware." The same holds true for FabricPath, except that the cost savings will likely be much bigger and easier to measure. This is especially true for organizations that move large amounts of network traffic between data centers. When you add to that the resiliency that is enabled by the active use of all links between devices, the case becomes even more interesting.

Those pesky, long-lived Catalyst 6500 chassis will start to look a lot more aged with the release of the FSS (FabricPath Switching Service). This is going to be especially true for organizations that are embracing server virtualization platforms of any variety. As the number of virtual systems increases--and the use of business continuity features including vMotion become more prevalent--the more keenly Layer 2 network limitations will be felt.

And it's clear that Cisco plans on increasing the capacity and capability of the Nexus 7000 and FabricPath to accommodate the tempo of virtualization adoption. For one thing, Cisco has "skin in the game" in the form of its UCS offering. For another, the obvious benefits of server virtualization are an escaped genie that cannot be rebottled. When combined with the ever-increasing deployment of x86 multicore systems, the two-year outlook for existing, spanning-tree-bounded networks is pretty bleak. 

Rocket Fuel