Cisco on Tuesday quietly rolled out its Nexus 4000, a blade switch which works with blade servers from other vendors. Thus Cisco has taken a significant step toward extending its Nexus switch family -- and in turn its unified computing footprint -- to third parties, who will in turn embed the new Nexus inside their own blade chassis.
The 4000 appears to be an OEM-oriented variation of the Nexus 5000, which is an FCoE network switch for the data center.
For this reason, the news gives additional heft to Fibre Channel Over Ethernet (FCoE), which is the centerpiece protocol of Cisco's Unified Fabric architecture. This enables flexible switching, where servers can access outside resources such as SANs -- admittedly not rocket science, but the flexibility is the key here when you're talking about an increasingly virtualized world where routing changes on the fly.
My colleague George Crump pointed out in a post the other day that FCoE has become just another protocol, which "you may or may not implement depending on your data center's needs." I'd agree with that. George was using a Brocade announcement as evidence of that. I'd say the Cisco news solidifies the message that FCoE is simply another widget in one's toolkit.
Back to the Nexus 4000: Cisco didn't issue a press release about the news, which is unusual. Rather, they talked about it in a live Webcast. I believe that approach was taken because Cisco was less interested in touting, per se, a product (the Nexus 4000) that's destined to be rebranded by third parties than they were in emphasizing their broader Unified Data Center (aka Data Center 3.0 strategy.
The more Cisco corrals all the pieces of that Data Center 3.0 puzzle and becomes, in effect, a one-stop shop for its implementation, the more they become a go-to vendor for extracting the efficiencies everyone's now looking toward. (I might add that the first step, before efficiencies, is getting a handling on a) the complexities of virtualization, which includes getting insights/dashboards etc. into exactly what resources you have and how they're being utilized and b) figuring out the best and fastest way to communicate amongst those resources, most notably managing the links between all those virtualizations servers and virtualized storage (and memory!).
Which brings us back to one of those key link protocols, namely FCoE. So in closing, here's a grab from a Cisco blog, which explains the utility of FCoE as implemented in the 5000 (which should apply to the 4000 as well):
"[This] lets us build more sophisticated network topologies as we deploy an FCoE based unified fabric. Up to this point, End Nodes have had to be directly attached to the FCF; however, with FIP, you can have intermediate "passthrough" switches between the End Node and the FCoE Forwarder (FCF). A quick example of where this might be helpful is a blade server chassis, where you might not want the cost and/or complexity of a full FCF in the chassis, but do want a switch that can serve as an FCoE passthrough to properly forward the FCoE traffic to an FCF."
Follow me on Twitter. Write to me directly at [email protected].
