Cisco Puts the Focus on Blades

Even as the enterprise industry forges ahead with broad architectural revisions playing out in the long term, immediate needs are not going unmet.

One of those needs is enhanced networking among increasingly consolidated legions of blade servers that still make up the brunt of data center operations.

Cisco Systems made a key move on this front recently when it announced the Nexus 4000 blade switch that fits in a standard blade enclosure with the intent to aggregate multiple blades onto a 10 GbE fabric. Naturally, the device supports the NX-OS, providing instant compatibility with the existing Nexus line, particularly the 7000 core switch and the 1000V virtual switch, designed to form the backbone of physical and virtual environments. It also gives you a choice of Fibre Channel, FCoE, iSCSI or NAS connectivity for converged data and storage networking.

The natural partner for this type of switch is IBM, considering that HP just came out with a set of integrated ProCurve switches for the BladeSystem line. The fact that Cisco recently entered the server market itself probably still rankles a few at Big Blue, but business is business after all, and it wouldn't do well for IBM to let HP run away with an all-in-one enclosure at a time when enterprises are looking for easily deployed solutions as they ramp up both virtual and cloud offerings.

The only fly in this ointment is the eventual disposition of Brocade, now that the company has placed itself on the auction block. IBM, Dell, HP, Oracle and even NetApp have been named as possible buyers. IBM, in particular, would be problematic for Cisco as that would move the number-two blade vendor into the column of those who will likely have an integrated switch of their own soon.

For Cisco, however, the big prize here isn't switch sales, but the Unified Computing System, according to systems architect Jason Nash. The UCS platform, he says, is designed to provide an extremely dense environment for enterprises running applications with high resource requirements. Combine UCS with a Nexus-based 10 GbE fabric and you have the foundation for an integrated environment capable of handing the workloads of the multicore future. It's also important to note that Cisco's architecture provides Nehalem-based servers with twice the memory of comparable systems, allowing you to pack more VMs per blade than normal.

The Nexus line goes to the heart of the fact that even consolidated commodity server environments live or die by their ability to communicate with each other. Cisco knows as well as anyone that networking will be the key component of both virtual and cloud environments going forward.

And as the dominant networking provider at the moment, it's their market to lose.