Last week from its official corporate blog Mike Neil, General Manager for Virtualization Strategy, announced a critical decision for upcoming release of Windows Server Virtualization (codename Viridian):
Microsoft has made some hard decisions and now plans to defer some features to a later release. The features are:
- No live migration
- No hot-add resources (storage, networking, memory, processor)
- Limit support to 16 cores/logical processors
This move has a serious impact on the whole virtualization strategy adopted so far.
On a side now that codename Viridian has been deprived of all its key features, will be much harder to justify its release in mid 2008.
Microsoft has been too slow updating its current solutions, Virtual PC and Virtual Server, since Connectix acquisition.
Both still lacks of some mandatory feature, like 64bit virtual machines or USB support, which market leader VMware and newest competitors like Parallels are offering since a lot of time. But despite that some potential customers decided to wait just because of those features now dropped.
As I wrote one year ago on the Microsoft, the big absentee of virtualization essay, software giant is limiting its effort in virtualization space releasing just a minor update every 1-1,5 years. And now that hardware virtualization is becoming a mainstream technology and mass amount of companies started recognizing its benefits, this pace is no more acceptable.
This last change in release plans further damaged company credibility, most of all if we consider Microsoft is doing same errors in the application virtualization space: despite the smart acquisition of Softricity one year ago, Redmond software giant did almost nothing to push SoftGrid technology to customers, or to seriously integrate the product in its offering.
On another side without Windows Server Virtualization key features, will be much harder to attract new customers or move VMware ones away from market leader.
As I wrote on the The Microsoft Virtualization Chance essay, Microsoft has an opportunity passing through the SMB market, offering new hypervisor for free, and pushing it with Windows Small Business Edition.
This option is still there, since smaller companies probably find limited benefits in hot addition of virtual hardware or live migration features. But now VMware has more chances to block this potential threat with a more consistent action on SMB segment.
On the enterprise market instead there are fewer openings: the imminent release of System Center Virtual Machine Manager (SCVMM) 2007 doesn’t seem enough, needing integration with too many other products (Operation Manager, Data Protection Manager, etc.) to provide a rich feature set.
Is this move impacting on competitors strategies? Probably yes.
As I said VMware has now much more time to consider its SMB strategy, and even a chance to beat Microsoft on its own game, providing virtual hardware hot addition for enterprise segment before any other on the market.
Parallels, expected to launch a couple of server virtualization products this summer, has new opportunities as well: challenging an absentee competitor to take its place is easier than challenging market leader.
In similar way XenSource and Virtual Iron, both using Xen as virtualization engine, will have more time to become more competitive than Microsoft, since upcoming Xen 3.1 (formerly Xen 3.0.5) will introduce VMotion capabilities as well.
Is this move also impacting on market? Absolutely.
First of all VMware will maintain its monopoly on enterprise segment still for a long time with little or no efforts, maintaining current prices and quality of support (both considered unsatisfying for some customers).
Secondarily Microsoft partners who are investing today in codename Viridian and Virtual Server may decide that the new hypervisor will take too long to take a serious share with new feature set, and may jump on a competitor bandwagon: VMware, XenSource, Virtual Iron, SWsoft.
Microsoft progression in virtualization space isn’t uncommon: they started offering a hosted virtualization platform and are now moving to a hypervisor architecture. Unfortunately lack of features and incredible slowness completely mined this transition.
Such big feature drop would be much more acceptable if it would lead to release new hypervisor inside the Windows Server 2008 (formerly codename Longhorn) RTM. Which is not the case: Microsoft only promised first public beta for that time, but at today few are expecting Redmond company will really be able to keep such promise.