Put brakes on Windows upgrade escalator, Gartner urges MS
Whistler to kill off NT, Blackcomb Win2k?
A recent Gartner report points up the tensions between the company's objectives and those of its customers, and urges enterprises to pressure Microsoft to put the brakes on its OS obsolescence programme. Microsoft's official policy is only to support two generations of OS at a time, but if the company decides Whistler constitutes a new OS, this would mean the end of NT 4.0 and Win95 in the second half of this year.
As Whistler is intended to be the successor to both Win2k and Win9x, and Microsoft has now put out a new version of Win9x three years in succession, it's surely a racing certainty that Microsoft will decide it's a new operating system, rather than a cleaned-up version of Win2k. Which of course it will be, but those three 'new' consumer OSes in a row were cleaned-up version of Windows 95, right?
Gartner points up the arbitrary nature of this; Microsoft doesn't provide guidelines about what constitutes a new OS, nor does it make its intentions sufficiently clear to its customers in advance. The OS support policy is to support the current operating system plus its predecessor, which currently means Win95 should move out of support (because Win98 SE isn't deemed a new OS, Win98 as a whole is OK until Whistler), and that Windows 2000 and NT 4.0 are both supported.
On the consumer side, the application of the policy would mean support will be reduced to WinME once Whistler ships (if it is a new OS, that is). That however has a knock-on effect for business, because Win98 is still the official Microsoft recommendation for Win9x in business, and WinME is aimed squarely at consumers, not business. So killing off Win98 would drive businesses to Win2k (or possibly even straight to Whistler), rather than giving them WinME as a stopover.
Another problem is the speed of the escalator. A converged codebase could (no doubt in the view of the Redmond bean counters, should) result in the consumer annual refresh cycle being applied in business as well. So the next next version of Windows, Blackcomb, shipping in 2002 (or more likely, another rev of Whistler) would kill off Win2k by 2003.
To get some perspective on this, note that Gartner's report and recommendations are aimed at businesses running Win95 and NT 4.0, and that the outfit reckons 95 is probably viable in business through to 2003, while NT 4.0 can make it as far as 2004. Gartner says: "If Microsoft waits until Whistler is announced and decides that Whistler is a new generation of Windows, enterprises likely will not have enough time to react if they decide they need to migrate to a fully supported platform. Enterprises should pressure Microsoft to give two to three years notice before changing any NT v.4 QFE [Quick Fix Engineering] support policies."
As history tells us that Microsoft seldom knows what its OS strategy will be two years (we're being generous) down the line, we doubt very much that this is achievable. ®
Full Gartner report
Sponsored: Benefits from the lessons learned in HPC