|
|||||||
|
|
![]() |
|
|
Microsoft Explains Its Vision of the Autonomic Windows Future by Timothy Prickett Morgan Everybody who's anybody in the IT racket is talking about how their engineers are developing self-healing, self-administering, autonomic computer systems. The Herculean human labor that goes into running data centers makes that labor an increasingly large piece of the IT budget pie. While people have always added up in the IT budget, the rapid decline in the cost of client and server computers and their software has only made the people costs seem like they are ballooning. But there's more at work with Microsoft's just-announced Dynamic Systems Initiative, or DSI, which was unveiled in Las Vegas last week at the second annual Microsoft Management Summit. DSI--or Dizzy, as I will probably call it because it rhymes with IBM's eLiza (let's say Lizzy) autonomic computing initiatives for its servers, operating systems, and middleware--is the fulfillment of the credo for the IT industry, and indeed for industrialization in generation: that which can be automated should be automated. There's more to DSI and similar autonomic computing undertakings (an apt choice of words, as you will see in a moment) than simply getting rid of people to manage systems, databases, storage, and other gadgetry because they require pay raises, vacations, sick days, and health insurance. Yes, the people costs of IT solutions is somewhere in the range of 70 percent, according to statistics cited by Microsoft in the Dizzy announcements. But, frankly, distributed computing systems--and particularly the evolving Web services variety, which spreads applications across multiple geographies and platforms--are simply too complex for any set of people to manage by hand. So are nuclear reactors and Space Shuttle reentries into orbit, just to name two that have experienced lots of automation and some catastrophic failures. Automation such as the kind that Microsoft is espousing with Dizzy and IBM is with Lizzy is about automating computer systems and applications to avoid catastrophes like those that are commonplace in IT shops today. It is also about changing the way that applications and administration tools work together, so that applications can change in such a way that the administration tools and techniques of those applications do not get completely upset as the applications change. DSI, as I understand it, is not so much a product as an attitude and a collection of partnerships among software and hardware makers that play on the Windows platform. It's like the Trustworthy Computing initiative of Microsoft's chairman and CEO, Bill Gates, in this regard. There are, as usual, some core DSI technologies that will be embodied in the Microsoft portfolio of products in the future. In the mid-1990s, the engineers and theorists at Microsoft Research, the research arm of the software giant, started looking into how the development and administration processes for Windows-based applications could be married and better managed to reduce costs and complexities. After two years of research, the team behind what has become Dizzy moved over to the Windows Server 2003 group to start making real products in order to bring the Dizzy vision to life. One of these products is something called the System Definition Model (SDM), an XML-based "blueprint" that codifies the operational requirements of system administrators and the databases and applications they manage and the policies and agreements that data centers make to provide services to company end users. The idea is that getting upper level management to tell coders what policies they have put in place will allow developers to actually write code that adheres to the deals IT managers are cutting with upper management. Microsoft also announced a new tool called Microsoft System Center, which will be comprised of Microsoft Operations Manager 2004 and Systems Management Server 2003. These two tools will be integrated with the SDM XML blueprint, and will also presumably hook into other systems management frameworks to allow better harmonization and control of application development and system administration. Exactly when this will all happen is unclear, so don't fire your system admins just yet, and if you're a system admin, don't polish your resume above and beyond what you normally do after a hard week's work. Microsoft says that over the next three to five years, it will deliver unspecified Dizzy technologies that span all of its products, not just its core server operating system. In any event, the Dizzy technologies that Windows Server 2003 includes are as follows:
Microsoft says that the Automated Deployment Services component of Dizzy is now in beta for Windows Server 2003, which is just about ready to be launched. It is available on a limited basis to early adopters, and the final code is expected to be available sometime in the second quarter of 2003. Not one word in the Dizzy announcements said anything about the IT platforms outside of Windows, which still accounts for the majority of the data processing that gets done in the world. Microsoft, as usual, figures that if it can make its own platforms do everything more efficiently, customers will dump their old apps and platforms to move to Windows. It might not be a bad bet, really, if DSI turns out to be the impetus behind real products and not just a bunch of press releases and pronouncements from Microsoft, because it, too, has to talk about autonomic computing.
|
Editor
Contact the Editors |
| Copyright © 1996-2008 Guild Companies, Inc. All Rights Reserved. |