IDC adds VA Linux Systems to server survey elite
Market research giant 'clarifies' its mistake, but the numbers still don't quite add up...
A red-faced IDC has admitted that VA Linux Systems is indeed one of the top five Linux server vendors, and that it's recent report to the contrary was... well... wrong. IDC released its report on the state of the Linux server market during Q4 1999 last week. It put Compaq in the lead, followed by IBM, Dell, Hewlett-Packard and Fujitsu Siemens. VA was consigned, along with a number of hardware companies more associated with the open source OS than the top five, into its catch-all 'Others' category. VA immediately cried foul, pointing out that its numbers showed it clearly ahead of Fujitsu Siemens. It also said it had reached an agreement with IDC under which the market research company would "clarify its earlier statement". It's taken a week or so for IDC to make that clarification, but finally overcoming its embarrassment it has, and today issued a released stating that it will "will re-categorize and track VA Linux Systems as a 'brand-name' vendor in its Commercial Systems and Servers program". That's good news for VA, but it does raise a degree of concern over IDC's research procedures here. The statement implies it's only tracking what it considers to be 'brand names'. The fact that it didn't consider VA a brand name in this sector itself suggests that it wrote off VA without first checking the vendor's sales figures to see whether its assumption was correct. And if IDC is only tracking "branded" companies, what else is it missing? In any case, the Q4 1999 Linux server market now looks something like this: Compaq, 25 per cent; IBM, ten per cent; HP, seven per cent; Dell, seven per cent; VA Linux Systems, five per cent; and Fujitsu Siemens, three per cent. However, one thing still puzzles us. IDC puts VA's unit shipments during the period October to December 1999 at 3645 servers. The only trouble is, that's exactly the same figure VA cited as its unit sales for the period November 1999 to January 2000, its second fiscal quarter, which suggests VA experienced exactly zero sales growth during between the two periods. Either that or, to save time (not to mention face), IDC simply cut and paste the figure out of VA's email without realising it covers a different period from its own numbers. And surely IDC wouldn't do that, would it? ® Related Stories VA Linux Systems disputes IDC server market findings Linux server market dominated by IT giants
Sponsored: Benefits from the lessons learned in HPC