Citrix buys XenSource: what next?
And what will Microsoft and VMware do now?
Last time we wrote about virtualisation - only just over a month ago - we said: "the virtualisation market is in a bit of turmoil at the moment, shaken up by the prospect of billions of dollars of revenue." How prophetic those words have proven to be, albeit not in the manner we expected. Market gorilla VMware now has a rival. But what does it all mean?
If at the start of last week you were taking bets on the virtualisation market, you would have seen growing potential for consolidation - not just of servers but also of vendors. And so it has come to pass, with the acquisition of open source developer XenSource by Citrix. Few guessed ahead of time that Citrix was a possible buyer for one of the many and growing number of virtualisation software houses but, from the desktop provider's point of view, it makes a lot of sense.
Citrix branches out
Citrix has, effectively, burgeoned on the back of a single product: Presentation Server. While it's undoubtedly grown its product portfolio over recent years, the bulk of its revenue still comes from server-based software that pushes Windows desktops running on heavy-duty, centralised hardware out to thin clients. It's a model that works in many circumstances but one that, from an end user's point of view, trades flexibility and power for centralised control.
The most obvious way of giving users back the power and control they need but without removing the central control that's essential in many environments is by provisioning desktops via virtual machines running on datacentre-located servers - desktop virtualisation. This is the technology that Citrix acquired last week when it bought XenSource for US$500 million - a tidy sum for a tiny company with only a few hundred customers.
While obvious though, desktop virtualisation has yet to become a major trend. The market is way smaller than that of the server virtualisation market, probably - and analysts are in agreement that server virtualisation has huge potential for growth. Analysts regularly produce forecasts for server virtualisation with numbers in the billions of dollars, while desktop virtualisation is set to grow to a mere US$900 million by 2010. Between 6-10 percent of all servers have so far been virtualised, according to estimates, which means that desktops have even bigger potential in the longer term, given their far greater numbers - all other things being equal of course.
Analysts Philip Winslow and Dennis Simson from financier Credit Suisse reckon that Citrix has made the right move: "We view the price paid for XenSource as a reasonable/attractive value, given that the acquisition triples Citrix’s addressable market over the next five years and uniquely positions the company in the desktop virtualisation [sic] market, which is estimated to grow to the size of the current presentation virtualisation [sic] market in the next 4-5 years. From a market share perspective, $50 million in XenSource revenue represents approximately 2.5 percent of the virtual machine market in 2008 - which we view as achievable."
Newly public VMware didn't react well. It currently owns some 85 percent of the virtualisation market and looks set to continue raising the barriers to entry into its market with the new software - ESX Server version 3.1 is due for release in a couple of months and the company plans to show a new version of its free but more limited VMware Server 2.0 product at its annual developer and customer conference VMworld next month, for example.
But that hasn't stopped the company sticking the knife into the Citrix deal. In a letter sent to channel partners, the market leader wrote trashing XenSource's market share, its technology and its strategy.
"XenSource failed to gain measurable market share. Its upcoming XenEnterprise v4 release still has major shortcomings compared to VMware Infrastructure 3. Citrix provides impressive technology, but its competencies and resources are not in the field of system-level virtualisation, which is what XenSource needs in order to deliver an enterprise virtualisation solution," wrote VMware.
Coming shortly after VMware's IPO, which raised US$957 million, it appears that VMware lacks neither resources nor confidence.
Yeah, but so what?
The question is what it all means for the rest of us. While it's fun to look at the numbers and wonder whether and how much of a competitor for VMware Citrix will become, what complicates the matter hugely is the fact that Citrix is a long-term collaborator with Microsoft. In fact Citrix' presentation layer technology, developed for the purpose of transmitting desktops to thin client hardware, is now resident in every standard copy of Windows. Newly acquired XenSource also had a technology-sharing deal with Microsoft.
Both those deals look set to be at the very least strained when Microsoft finally makes its much-anticipated entrance with Viridian, the codename for its hypervisor technology. However, you can expect this be relatively limited in functionality, with the next rev more likely to be intended as a VMware killer.
Citrix CEO Mark Templeton on the other hand appears content that its deal with Microsoft will hold, provided it continues to support Microsoft products - which will include Viridian, the first public beta of which is due out before the end of 2007. Templeton is also reported to have said that XenSource will retain its Cambridge HQ; Xen was developed there.
In practical terms, expect Citrix to develop products that deliver desktops via virtualisation as well as its current ICA protocol in the medium term, and for that product to be fairly efficient thanks to the company's closeness to Microsoft technology. However, the fact that Citrix will have a product to rival Viridian is bound to put strains on that relationship.
In the shorter term, you can expect Citrix to repackage Xen as a server virtualisation system, one that will eventually support Presentation Manager directly rather than via an intervening OS layer.
But what of the gorilla in the room? VMware positions itself as an OS-agnostic software company - indeed, co-founder Mendel Rosenblum recently went further and predicted the death of the OS in favour of the hypervisor. But in increasingly swallowing up market share, VMware is upsetting an increasing number of other industry players. Might we witness a defensive move in the form, perhaps, of an industry grouping that excludes VMware?
If nothing else, the events of the last couple of weeks demonstrate that the virtualisation market - although still in its nascent stages - is growing up. The gloves are off as the value of the market emerges: there's something really worth fighting for. This has also stimulated new companies such as Virtual Iron, which re-purposes the Xen hypervisor as the basis for its product - to enter the market.
In the long run, this frantic level of activity will undoubtedly result in more product choices and lower prices. And that can't be bad.