Red Hat announced two important moves this week; open sourcing of Red Hat Network Satellite, and its own virtualization hypervisor oVirt. Open sourcing RHN Satellite is fundamentally about showing the industry Red Hat is still the keeper of the open source flame but the real strategic move is the development of oVirt. OVirt is built upon Kernel Virtual Mode, or KVM, which is virtualization built right into the Linux operating system, and has been maturing over the past two years. Until now Red Hat's virtualization strategy has been built around open source Xen, much like other players such as Oracle and Sun.
When Citrix took over Xen, it created a sticky situation for vendors depending on the Xen open source software for its virtualization strategies. Such is the way of open source these days. What is free now, may have some corporate and competitive entanglements down the road. Red Hat's now separated itself from the crowd with oVirt. That's an important strategic move, giving Red Hat control over its own virtualization destiny. But it also means Red Hat must carve a niche in the virtualization market for yet another virtualization technology. Being built upon KVM, something now core to the Linux OS, mitigates that somewhat but the proving is yet to be done.
Red Hat has to court hardware manufacturers to make their oVirt the hypervisor of choice. Red Hat has to develop all of the fancy doo-dad features VMware and others have been putting in their products. OVirt has to gain enough mind share so that it's imperative that virtualization management software, like Microsoft's System Center Virtual Machine Manager, add support for oVirt much like Microsoft did for VMware.
The good news is Red Hat's virtualization strategy is now back in its control. The bad news is there's still a ton of work ahead for Red Hat and no guarantees of success. But the good outweighs the bad in this case. OVirt is a very important strategic move for Red Hat. And it's a smart one.
Like this? Here are some of Mitchell's recent posts.
Live Mesh Dumps UAC Live Mesh Dumps UAC
iPhone's "Just Good Enough" Business Strategy
Apple User Saves Microsoft From Collapse
Microsoft / Nortel Move UC To The Cloud
Ozzie, You Must Transform Microsoft
Microsoft NAP: Above The NAC FrayProduct Reviews:
Microsoft Live Mesh Google App Engine LiveNewsCameras.com Xobni Outlook pluginRock Star jobs in SaaS: SaaS Jobs
Recent Converging Network Blog Posts:
Get Ready For XaaS Everywhere Unbelievably Bad Web Password Security Back From Hiatus, Saved by Web 2.0 Technology It Takes a Village.. ah, actually, being there first and tons of hard workFavorite Book Recommendations: The Big Switch Zero Day Attack Clear Blogging
Check out Mitchell's Converging On Microsoft Podcast. Current Podcast Episode: Security Mike Gets Serious About Security
Also visit Mitchell's personal blog The Converging Network and SSAATY Security Podcast.
Visit Microsoft Subnet for more news, blogs, opinion from around the Web.
Sign up for the bi-weekly Microsoft newsletter. (Click on News/Microsoft News Alert.)
Mitchell Ashley is CEO and Chief Strategist of Converging Network, LLC, providing product and technology strategies to emerging technology companies. A serial entrepreneur, Mitchell has created many successful products and services in the networking, security, convergence, Internet and IT industries. In addition to blogging for NetworkWorld, Mitchell regularly blogs at TheConvergingNetwork and co-hosts the widely popular Still Crazy After All These Years podcast.
|
|
KVM is not a hypervisor and never will be
The very idea that Red Hat can wrest control of anything open sourced, from any other vendor, indicates that this author fails to grasp the power of the open source movement. Red Hat is a contributor to open source, and the beauty of it is that it is free for use by anyone, and that it can be developed by anyone at any time.
With regard to hypervisors, it should be clearly understood that KVM is not a hypervisor, but a hosted virtualization model similar to VMware Server or Microsoft Virtual PC. The technology that KVM purports to have innovated is, to put it bluntly, well established in competitive free products. Most notably, none of those free products is used in production for enterprise workloads. Why? The hosted model of virtualization offers a simplistic way for an OS to host more OSes as virtual machines. But customers want their virtual infrastructure to be independent of their OSes - witness the fact that VMware's revenues eclipse those of Red Hat.
The KVM technology is an interesting, immature, Linux based model for Linux to virtualize other guest OSes. It's convenient for Red Hat to adopt because it's just part of the Linux kernel. But Red Hat is committed to Xen, and indeed at the current Xen Summit in Boston, where 47 vendors, 10 universities and 15 countries were represented in a developer group of more than 175 individuals, Red Hat engineers continued to offer contributions to Xen. Red Hat does have a very perplexing challenge though, namely the fact that KVM and Xen VMs are incompatible, and so they will find themselves offering their customers two incompatible virtualization technologies. I am not sure that this is particularly compelling as a value proposition.
Ultimately Xen and KVM will co-exist nicely as independent open source projects. Those that want a simple way for Linux to virtualize more Linux VMs will use KVM, whereas those who want a secure, minimal, type 1 hypervisor as a component of a holistic virtualized infrastructure in which server resources are pooled into a seamless set of resources for any VM workloads, will use Xen. Customers have already voted with their feet - VMware, XenServer and Hyper-V represent the only viable choices in the market for virtualization.
Simon Crosby, CTO Virtualization, Citrix.