Cisco Nexus Ushers in the Unified Fabric

Arthur Cole

Is it a case of switch envy or will enterprise networking soon undergo some real changes?


Cisco Systems is the latest to offer an enterprise switch platform for companies looking to capitalize on virtualization and consolidation technologies already being deployed. However, the company is talking about more than simply boosting bandwidth into the stratosphere; it's looking to fundamentally alter the way both data and network resources are manipulated.


First, the news: The company has announced a new line of core switches called the Nexus 7000 series that can scale up to more than 15 terabits-per-second of throughput across more than 500 10 gigabit Ethernet ports. And the company is working on designs that would support 100 GbE.


But more significantly, the switch operates as a bridge between Ethernet and Fibre Channel protocols, providing a unified fabric for both IP and storage. Right off the bat, that means no more parallel networks that have to be monitored, maintained and upgraded. But when you couple that with Fibre Channel over Ethernet (FCoE) and systems like the Application Control Engine, you'll have a single network that will allow you to seamlessly shift data and computational resources not just across the data center or the home office, but over the wide area as well.


I had a chance to speak with Deepak Munjal, Cisco's manager of data-center solutions, who said the Nexus switch is the key element in the company's DataCenter 3.0 strategy announced last summer. By unifying storage and computing networks, enterprises finally will realize the efficiencies and management benefits of centralized environments as well as the flexibility of distributed architectures.


"We're now seeing the benefits of centralization, but on a more intelligent level," he said. "It's still a distributed architecture based on open standards, but we will manage it as a mainframe."


I also asked him whether customers, having already made the transition from centralized server rooms to distributed networks, seem willing to rework those networks yet again for an admittedly intriguing, but largely unproven, vision. He said customers who are breaking into high-bandwidth applications such as video networking or are running large data loads to and from consolidated virtual environments, existing core switches like the Catalyst simply aren't powerful enough.


"We are going to recommend that most customers with Catalysts today, that the position network insertion point (for the Nexus) be in the core of the network," he said. "The core Catalysts can be redeployed onto the aggregation and server layers, so we're not talking about a major, immediate disruption, but a slow redeployment from the core."


Cisco isn't the only networking vendor with grand plans for the future. In the past week alone, we've seen new enterprise proposals from Brocade and Emulex, with a likely announcement coming soon from Juniper.


At $75,000 a pop, the Nexus ain't cheap, but radical new technologies rarely are. The company plans to start shipping them in the second quarter. It will be interesting to see whether the switch, and the unified fabric in general, lives up to its promises.

Add Comment      Leave a comment on this blog post
Nov 28, 2008 6:05 AM wan accelerators wan accelerators  says:
I' ve had occasion to try out taksi, it worked well for GDI capture, but for Direct3D capture on the engine I used it failed in CTaksiDX9:: GetFrame during GetRenderTargetData. I' ve found a solution by disabling the avi feature (I didn' t need it) and using screen capture through the texture api with a direct surface to file save- I used D3DXSaveSurfaceToFile. GetRenderTargetData failed with INVALIDCALL- I didn' t investigate further, but your comments and the msdn documentation suggest it could happen due... Reply

Post a comment





(Maximum characters: 1200). You have 1200 characters left.



Subscribe to our Newsletters

Sign up now and get the best business technology insights direct to your inbox.