Gartner's analysts argue that conflicting server needs are increasing, neccesitating Windows to adopt a multi-kernel approach or perish under bloat and complexity. (Source: Gartner)
The analysts say that Windows is simply to big and simpler and more targeted versions need to be developed. (Source: Gartner)
A controversial analysis by Gartner looks into Microsoft's Vista woes
Its no secret that Windows Vista wasn't the resounding success that Microsoft hoped for. While Microsoft managed to push decent sales, it failed to surpass Windows XP's success and received a largely critical reception from the press. Microsoft is taking the situation seriously, and is pushing aggressively to replace Vista, with Windows 7 in 2010, hoping for better luck.
Much of Vista's failings were due to something entirely out of Microsoft's control including poor driver support from hardware partners. However other parts were admittedly Microsoft's own doing, such as inaccurate hardware compatibility claims and its UAC popups.
Fortunately for Microsoft, its competitors have failed to present a cohesive alternative, with Linux still struggling to gain marketshare and Apple OS X still regulated to Apple-endorsed hardware.
A new analysis from Gartner warns though that Windows is on the verge of collapse under its own weight and that if doesn't do something quickly it is a matter of inevitability that a more able competitor will eventually dethrone it.
Gartner's Michael Silver and Neil MacDonald carried out the analysis. During a survey, which was just one part of the analysis, the pair posed the question of whether Microsoft needed to radically alter its approach to Windows to stay viable to a room full of IT managers and executives. Half the room raised their hands.
The problem, the pair says, stems from Microsoft legacy support, followed by increasing complexity and hardware issues. Silver offered the anecdote that his customers have been calling him asking if he advised waiting for Windows 7 and skipping Vista, as Windows 7 is allegedly going to be more lightweight and modular. Silver acknowledged that his advice to these inquirers was to only adopt Vista on an attrition basis; as XP PCs died, he then would advise replacing them with Vista PCs.
MacDonald pointed out that server demands are evolving in different directions, making multiple kernels one possible approach for Windows. He says that Windows is already too bulky and complex, and that simply adding on more complexity will lead to the downfall of the OS. The pair both advice that Windows create multiple versions focusing on individual customer niches.
This may be a somewhat controversial proposal, as many already complain about the multiple Windows SKUs. However some of these complaints focused on varying levels of functionality, so perhaps an approach delineated purely by target group would be better received. MacDonald suggests these different versions be virtualized, but warns that Microsoft won't take warmly to such a proposal. Says MacDonald, "Microsoft doesn’t like anything in between Windows and the hardware. Ninety-five percent of its revenue comes from OEM’s."
The analysis demands Microsoft must reduce its development times, provided more innovate, deliver more inter-platform consistency, and solve compatibility problems among various Windows varieties. The pair also suggest that Microsoft tailor fit the operating system to specific applications, limit what applications can do to improve security (the One Laptop Per Child project's OS is among the few platforms to currently do this), make migration to new versions easier, and simplify licensing. Says MacDonald, "Something as common sense as ‘I’d like Office to go with me’ doesn’t work under current licensing."
The take home message of the analysis: replace Windows, lock-in needs, and adopt more predictable product schedules. The type of Windows that the analysts desire would be tailored to location and to a users identity, or as they put it provide "a composite adaptive work space".
The pair's analysis will likely be extremely controversial. While Windows may eventually move away from legacy support, it makes billions in revenue off it, making it unlikely to want to leave it. Further, fear of alienating customers will likely dissuade it from radical strategies such as multiple kernels. Nonetheless, the commentary is an interesting take on Windows' problems and what might be done to fix them.
