Virtualisation company VMware has publicly disputed the way in which key research into the virtualisation market was undertaken by analysts IDC, and disputed figures that appear to show a rapid growth by a competitive product from Microsoft.
According to the IDC research published last week, Microsoft has done well in virtualisation since the launch of its hypervisor, Hyper-V.
"[In] its first quarter of general availability Microsoft Hyper-V delivered a strong showing, and when combined with Virtual Server 2005, Microsoft's market share is 23 percent of new shipments," according to the IDC research note.
Then in a blog posting on Monday, Mike DiPetrillo, a principal systems analyst at VMware, questioned IDC's figures. "So did Hyper-V really ship enough units in two days to get 23 percent market share? I doubt it," DiPetrillo wrote.
The figures have been seized upon by Microsoft and emphasised in the company's corporate blog on Microsoft TechNet. "Customers now have choice in market. VMware is no longer the dominant server virtualisation vendor (less than 50 percent [of the market]). Customers now have choices that include Microsoft, Citrix, Sun, Novell, RedHat and Oracle."
The blog post continued: "Microsoft's server virtualisation software already has half of VMware’s share in the market."
On Thursday, a spokesman for VMware confirmed the company's low opinion of the findings. "IDC says that Hyper-V delivered a strong showing in the first quarter of general availability, but how do they know?" asked Reza Malekzadeh, senior director of products and marketing for VMware in EMEA. "If you look at when Hyper-V was launched in October it was only on the market officially for a couple of weeks before the end of the quarter. So how do they know it had a strong showing and how did it pick up 23 percent market share in that time?"
IDC reacted strongly to what it called "the confusion and misinformation being sown by the folks at VMware". Spokesman Mike Shirer told ZDNet UK that "[in] response to some of VM's comments, you'll note that the 23 percent share is the Microsoft share (Virtual Server 2005 + Hyper-V), not Hyper-V share alone". He also emphasised that "we did take into account OEM shipments".
Malekzadeh said the whole process of producing the figures was somehow "different from usual practice" with IDC surveys. "Usually IDC will check the research with the vendors to see if it tallies with their figure, before they publish. It gives you a chance to compare what they are finding with your own research. For some reason, this time it did not happen," said Malekzadeh. "I don't know why."
This was denied by IDC. Brett Waldman, Research Analyst, System Software at IDC told ZDNet UK: "The unit shipment tracker is the result of IDC analysis of publically available financial data, interviews with software vendors, and competitors' information cross-checked through IDC demand-side studies. We interviewed 2,500 people in 35 countries, and we asked about production usage, which could have included use of the Hyper-V beta which was available months before the official launch."
When asked whether there had been any difference between this study and previous ones, Waldman said: "It's a biannual study. We did exactly what we did last time."
Asked what the analysts thought had gone wrong in the process, Malekzadeh said he could not say, adding: "I would ask IDC the question: who had sponsored the research?" When pressed on the point, Malekzadeh would not elaborate, but said: "Just ask them".
If the implication was that, in drawing up the figures, IDC had been influenced by Microsoft because the company had paid for the research, the analysts were keen to make it clear this was not the case. "IDC's Server Virtualisation Tracker is not a sponsored product," said Waldman. "It is available on a subscription basis to all IDC clients."