Which virtualization platforms will matter in 24-36 months time?

I was asked this question recently at the first Opscamp in Austin, TX. I sure don’t know who will win or lose but there are some big trends emerging.

There will be a few big winners. It is a crowded space with all the big players vying for market share, or at least install base. Given the costs incurred by ISVs, service providers and end-customers supporting every platform, it’s only a matter of time before we see a handful of clear leaders and the rest drop out. Why only a handful?

Virtualization is an OS feature. This is clearly already happening – Windows Server 2008 R2, RHEV are the products you buy today, not Hyper-V or KVM. In the race to differentiate themselves and grab mind/market share I think we’ll see more ‘Enterprise’ features to be baked-right-in. Live guest migration; block-level replication; virtualization aware I/O. Check; check;check! I can hear the old-school IBMers laughing from here.

It won’t matter whose platform you pick. We already have a de-facto standard for manipulating these virtualized containers, courtesy of a bookseller in Washington State. While it doesn’t have everything we need yet, there is a nice ecosystem of tools and vendors bridging the gaps in the current implementations.

It’s all about the applications. While the analyst community keep (or do you mean “are keen” on measuring the number of virtualized servers, a more interesting (and harder to obtain) metric might be the number of virtualized applications. We’re rapidly reaching the point where mid-sized companies can virtualize all the key applications they’re still running in-house. Not all of them will choose to do so, but ops guys managing the business-critical, custom applications at the heart of many of small and medium-sized businesses sure could use the efficiency bump. Start moving software bits instead of atoms!

What do you think? Have I got it all wrong? Join the discussion below.