KVM Pushed as Open Alternative to VMware
The Open Virtualization Alliance (OVA) was announced on May 17 and much of the analysis since then has focused on the formation as an attack on VMware. I concur that the OVA is an attempt to coordinate support for KVM (Kernel-based Virtual Machine), which is an integral part of Linux. But if this is a war machine, it's got some contradictions that IT managers should notice.
For one thing, nearly the entire OVA membership, including three of the four governing members (HP, IBM and Intel) have extensive dealings with OVAs nemesis.
BMC and Eucalyptus Systems, both regular members, also have extensive support for VMware vSphere, Microsoft Hyper-V and Citrix XenServer. And SUSE Linux plays the field when it comes to x86 virtualization.
That leaves Red Hat. Driving a competitive offering that creates virtualization choice is a good thing. So I'll say that IT managers should look at the OVA through this lens, understanding that Red Hat has a vested interest in wresting IT dollars from market leader VMware.
Here's my 1-2-3 for looking at OVA.
1. OVA is a marketing, coordination, and advocacy group for a product that currently has very little market share. It's not a product or standards group. Although member vendors are required to make or use KVM-based products as part of their membership, they all do this already. So, don't look for standards, specifications or certifications to come from OVA. Mainly, look for propaganda.
2. If you keep in mind that this is a competitive challenge to VMware, you'll be better able to decode the propaganda and find the chunks of information that will be useful to your organization. Related to this, it appears that nearly all OVA members are covering their virtualization bases. In this case, it's worth taking the hint and covering your bases so as not to be caught on autopilot in the case that a VMware alternative would be useful to your organization.
3. The OVA has something to prove and in pretty short order. The OVA web site spends almost as much time talking about what the OVA won't do (no products, no standards, advisory not directorial relationship with the upstream KVM development community) as what it will do. I'll be interested to see if the organization produces anything of benefit for IT managers.
For related news, see my colleague Darryl Taft's story here.