BIG BLUE SWINGS AT MICROSOFT

From Business 2.0's Future Boy:

IBM's Microsoft End Run
Big Blue hopes to make mini-servers out of your laptop, cell phone, and PDA -- and turn Microsoft Office into just another plug-in.

Business 2.0
By Erick Schonfeld
May 14, 2004

"This is not about a competitive battle," insisted IBM (IBM) software chief Steve Mills to a small ballroom full of reporters Monday. "Our intent here is not to displace Microsoft Office." But despite such protestations, IBM's "new software model," which revolves around its Workplace family of products, is a direct jab at Redmond.

While it may be true that Mills is not trying to displace Office per se, what he is trying to do may be even more damaging. He wants to reduce Microsoft's (MSFT) productivity software (Word, Excel, PowerPoint) to nothing more than a series of plug-ins for Workplace. Today, Workplace consists of IBM's popular Lotus Notes and Domino e-mail software, as well as other collaborative applications, including calendars and shared-document handlers. What IBM announced Monday is that companies will be able to customize Workplace by mixing and matching different applications -- such as e-mail, calendars, word processors, and Web browsers -- all on the same screen. In IBM's vision, each application becomes just another component, just another plug-in.

Try to visualize your desktop applications all running in different segments of one computer window, with the main application in a bigger segment in the middle and the other applications in smaller segments all around. Workplace can call up Microsoft Word or one of IBM's "document editors" (scrubbed-up versions of programs from the OpenOffice suite of free productivity software) or even enterprise software from Siebel Systems (SEBL) or SAP (SAP). Each application is just a plug-in for Workplace, and most can be updated and managed from a central Web server.

IBM wants to make this happen by putting a miniature version of some of its server software on your desktop or handheld computer. That is, IBM has designed a mini-database, a mini-application server, an agent for downloading and setting up new software, and an agent for synchronizing the software with a central server. By bringing all of this so-called middleware to client computers, IBM is hoping to replicate its success on server computers. Its WebSphere application-server software and DB2 database software are called middleware because they sit between the operating system and the applications that run businesses and websites. In fact, it hardly matters what operating system a server runs -- whether it's Windows, Unix, or Linux -- because programmers write applications to the middleware.

Now IBM is trying to do the same thing on the client side, the seat of Microsoft's monopoly power. As client devices proliferate -- Windows laptops, Palm Treos, Nokia (NOK) phones -- IBM is trying to entice software developers and corporate customers with an easy way to extend applications to all devices, no matter what the underlying operating system is. IBM is effectively saying, "Write to our middleware; we'll take care of the rest."

Why would anyone do this? It is not yet clear that anyone will. But IBM is trying to sell the idea as a lower-cost way for corporate IT departments to manage the proliferation of computing devices. For software developers, it would be an easier way to take software code written to run on a server and make it run on any device (because IBM's client middleware is designed to run smaller components of the same server code). And, unlike Web-based software, these applications could run even when the client is disconnected from the Internet.

IBM's real bet here is that computer software is ready to take its next evolutionary step. With help from the Internet, stand-alone client software tied to a PC was able to become centrally managed Web software accessed through a browser. The browser, though, is reaching its limits. What IBM is proposing is a way to combine the richness of client applications with all the benefits of a networked approach. Instead of using the old client-server model, IBM wants to turn client computers into mini-servers complete with middleware. This will allow client devices to constantly download and share applications with more powerful servers, and to be controlled by them (that mainframe mentality never dies). IBM, of course, makes a lot of money on server hardware and software, so it is a logical place from which to try to dislodge Microsoft's hold on client computers.

In the 1990s, Oracle (ORCL) and Sun (SUNW) tried a similar server-based approach, with "thin client" computing, and failed miserably. This time, IBM is taking more of a "thick client" approach. Its success will hinge on both its ability to lower corporate IT costs and its ability to create a new class of applications that improve worker productivity. On top of that, IBM will have to convince customers that going down this path is not simply trading one computing dependency for another.