Network Virtualization: The Cisco Factor

Analysts get paid to have opinions. Some observers think that those opinions have gotten too homogenized over the last several years, more groupthink is evident, and analysts as a community are in general more risk-aversive than they used to be.

There may be some truth to this. But the analysts that I’ve always admired as colleagues were those who didn’t mind taking a contrarian stance now and then. That’s why when I saw a research report suggesting that three vendors --- namely Cisco, VMware, and Citrix --- would dominate the virtualization market in the next year, it caught my eye.

The obvious question of course: where’s Microsoft? I spoke with Charlie Burns, one of the report’s authors to dig a little deeper. At the root of their contention is a key premise. Saugatuck sees network virtualization as coming on strong next year. The report states that “through 2010, server virtualization will have the single largest impact on budgets for IT hardware and support. The second largest impact will be network virtualization.”

If you accept this premise, then of course it’s a breeze to pass through a logic gate to Cisco being a dominant force. But this is where it gets tricky because everything depends on how exactly you define network virtualization. And this is a segment of the market that in my opinion has the potential for lots of fuzzy logic and less than rigorous interpretations.

So let’s take a shot. There are really two kinds of network virtualization. First, what’s been around for years with VLANs, VPNs, and other virtual types of partitioning. And then there’s this new beast, what might be called network virtualization 2.0 which involves a new array of products to support server, desktop and other kinds of virtualization. Cisco is involved in both and, as a minority investor, is working with VMware on the next generation data center where a lot of these new products and approaches are still being developed.

So here’s the definitional problem I alluded to earlier: network virtualization 2.0 products are still in development. So as the Brits might say, not bloody likely that they’re going to make a huge impact on the market next year especially against the likes of Microsoft in other areas of virtualization. However, if you want to throw the first kind of network virtualization into the mix, sure Cisco is selling a ton of gear (such as Catalyst switches) that use virtualization (which, of course, has nothing to do with hypervisors per se.)

As to Microsoft’s conspicuous absence in the top three, Saugatuck contends that Microsoft will not be a fully engaged market force in 2009 as it continues the process of rolling out Hyper-V. “Microsoft’s new solution will not be production-ready for most users for 6 to 12 months after its release”, the report states. You could certainly argue that a lot of early deployments will be used for test and development. You could further argue that VMM 2008 still hasn’t gone into G/A and won’t be available until later this year.

In fairness to the report’s authors, Charlie Burns and Bruce Guptill, the report was issued earlier this year, well before the release of Hyper-V. And the report really got me thinking which is what good research is supposed to do even if you don’t agree with its conclusions. But without better distinctions between network virtualization 1.0 and 2.0, additional detail of which Cisco products we’re talking about, and some hard numbers on the same (which coming from Cisco might be pretty hard to get), I’m not prepared to buy into the “Cisco in the top three” conclusion.

Posted by Tom Valovic on 07/08/2008 at 12:49 PM