Recent Cisco Win Shows Power, Problems of Network-Centric Approach
Teaching a new trick to an old dog is simple compared to convincing businesses to change the ways they think about solving problems. That came to mind as Cisco finally gained the first major customer, Savvis, for its unified computing system. For all the criticism that Cisco has gained, including on this blog, in its quest to challenge the server oligarchy, I'm starting to wonder how much of the corporate resistance has to do with the vendor's view of enterprise computing being too new for most CIOs. Not just in terms of asking them to accept a new view of a supplier, but a new view of what computing is supposed to be.
Stacey Higginbotham pegs much of the reason for settling on Cisco as because The Savvis cloud "relies on Cisco switches as well as VMware for its hypervisor and some management capabilities." But although there is a convenience factor, the bigger issue is that cloud computing strives for something that traditional IT concepts of products and services don't easily support: atomization. Computing and content are being broken into unbundled bits, and to sell someone on the benefits, first you have to get them nodding about the underlying concept.
When recently speaking with the CIO of a large service provider to businesses, he used such words as "fabulous" and "tremendous" about UCS. Realize that his company has been making heavy use of virtualization for close to a decade and that one of his aims is to think of providing resources in terms of CPU cycles or exact block of storage, not the more usual buy-a-server-for-that-app approach of IT, so he was particularly enamored of the network-centric version. When vendors depend most on their strength in networking, they tend to make everything subservient to that end, just as the usual servers vendors think in terms of every solution being another server.
This is a substantially different way to think than most IT departments -- largely because until recently, they haven't had the option. But it helps explain why Cisco may be having more trouble than you might otherwise expect. Not only does it have to sell CIOs on its abilities in delivering servers that will do what they need (That alone is a large hurdle.), but it must get them to think differently to prepare the ground for the competitive advantage argument. So it could easily take Cisco a good year or two to even start getting some uptick, because it has to wait for multiple attempts at education and explanation to take hold. The danger for Cisco is that in the process, competitors might try shifting directions and taking a similar route, which could help put a stop to the UCS efforts before they really get off the ground.
Image via stock.xchng user vjeran2001, site standard license.