Forrester: Fake servers like recessions
Even if you don't
Ah, there's nothing like a recession to kick IT industry trends into a higher gear - and get those IT analysts out there preaching change and cost cutting.
Today, it was Forrester Research's turn to talk up virtualization on servers and PCs as well as cloud computing as part of its sales pitch for two new reports it has put out casing the climate in data centers these days.
Those reports - called The State Of Emerging Enterprise Hardware Trends: 2008 To 2009 and The State Of Emerging SMB Hardware Trends: 2008 To 2009 - are based on a survey that Forrester did with 2,600 IT managers and other decision makers in the United States and across Europe to gauge their current and future plans with regards to various technologies and strategies.
This is a much large sample than we typically see behind such studies, but you have to bear in mind that this survey was done in the third quarter of 2008, before the global economy went on the fritz in a big way. Moreover, with the studies costing two grand a pop, you would expect a larger survey sample and more details.
According to the tidbits that Frank Gillett, the principal analyst at Forrester who put together the two reports, some 54 per cent of large companies polled have already implemented server virtualization in one form or another on some of their machines or will do so in the next 12 months. The statistic, as it turns out, is nearly the same for small and medium businesses. 53 per cent of SMBs reported they have done the fake server thing already or will do so within the next year.
Now, the fact that a little more than half of the companies polled report that they have monkeyed around with server virtualization in its many forms is not the same thing as saying half the servers at these companies have been virtualized, and it is surely no indication of what percentage of production servers have been virtualized. Test and development took a shining to virtualization early, and given the nature of the job - messing around with lots of software stacks and testing code before putting it into production - this is what you would expect.
Getting web servers and other infrastructure workloads virtualized is also comparatively easy and has a fairly big and immediate payback. But virtualizing the kind of mission-critical applications that tend to run on bigger boxes (application servers, database servers, and such) is much more problematic.
Sponsored: Hyper-scale data management