Cisco UCS: The morning after

So here we are, two days after Cisco has made their big announcement, and the mud-slinging is in full blossom. Juniper, HP, and BLADE Technologies (the blade switching vendor that HP and IBM sell via OEM) have taken their shots, calling Cisco “plumbers” and bemoaning their proprietary approaches. Sun even lobbed a grenade at Cisco’s core business, claiming they can use open software and standard computing hardware to compete in the router and switch business (not a new concept, by any means, and one which still lacks any real credibility). And this morning, IBM appears to be making a consolidating move in the server marketplace by sizing up Sun for acquisition – pure coincidence, or defensive move? The competitors are saying "caveat emptor", but you should also beware the sour grapes. No one likes it when a tank pulls up and stops on your street corner. You’re not sure if will be staying, but it’s awfully imposing and threatening. Why not throw some rocks to see if it’s possible to put some dents in it? The real question is – how much damage is this tank capable of doing to existing suppliers of enterprise data center computing? And while we don’t know that yet, everyone sure loves a fist fight. One of the darker sides of human nature, but also one of the virtues of the technology market – on a regular basis, new ideas get injected into the mix and ultimately drive the continued and frenzied pace of innovation, to the eventual benefit of enterprise IT. After all, there is a new idea here, despite the naysayers. It is the tight integration of networking and computing technologies that Cisco alone is in a position to deliver. Others have only been able to offer loosely-coupled alternatives, which are more open but are also much more difficult to manage in a holistic manner. But enough about the scrapping - it’s fun to watch, but ultimately it will die down and you’ll be left with the challenge of how to deploy and operationalize UCS, once a Cisco rep has successfully convinced your CIO of the compelling TCO. And while Cisco has made an honest effort to address this, they are not out of the woods yet. The BMC components do a great job of covering the bulk of the new technology being offered here – the B-Series compute blades and the virtualized server environments they will host. Cisco is even touting the concept of “service profiles” that can be defined for each computing load and which can follow VMs as they get provisioning, balanced, and moved around (though they also admit that these have not yet been put through the paces at their early field trial sites). Cisco is also promising APIs that can be used to integrate UCS into your favorite flavor of infrastructure management tools, if you don’t like the BMC-based default offering. What is lacking right now is the tight data and workflow integration that will be needed to deliver on the promise of a seamlessly contiguous virtual computing and networking architecture. The management tools that are being introduced with the UCS are not coming from Cisco’s Network Management Tools Group – they are coming from the Data Center business unit. And what this means is that there could still be silos of management within this unified solution, both at the “element management” level as well as the broader integrated management perspective, resulting in more special cases for operators to accommodate. Cisco has a couple of months to work this out – hopefully there is more to the story on this front….

Join the Network World communities on Facebook and LinkedIn to comment on topics that are top of mind.

Copyright © 2009 IDG Communications, Inc.

SD-WAN buyers guide: Key questions to ask vendors (and yourself)