Talking More Seriously About Integration

As I predicted ahead of Microsoft’s developer conference, the company plans to tightly integrate Web services, Longhorn and its developer tools together. In a Monday Weblog, here, I addressed some of the broader integration issues.

But the staggering amount of integration is even greater than I predicted, and there is good reason for Microsoft customers and competitors to be concerned about this.

Argument for integration: Microsoft is by no means the only company favoring increased integration as a development strategy. Other high-tech vendors rightly realize that a nice, tidy, integrated package of software can offer better performance and overall greater customer satisfaction. For example, employees working with an interface common to multiple applications should require less training than those moving among various application interfaces. This is one of many reasons Web browsers are so widely used to connect to multiple back-end corporate data stores.

In concept, integrated software packages can deliver better total cost of ownership. That assumes lower training cost, lower maintenance cost and reduced downtime because the pieces work well together. Apple, IBM and Sun are among the companies favoring integration; in the case of these vendors, that can include hardware and software.

Argument against integration: Microsoft’s approach to integration differs dramatically from many of its competitors. In the case of Apple, consumers receive a media player, Web browser and photo editor, among other applications with Mac OS X. Many of these components integrate with others. Windows XP includes similar software, but there is a significant difference. The majority of Apple’s integrated applications are separate from the operating system and can be removed at will. Microsoft’s software is integrated into Windows XP and cannot be easily removed, if at all.

Microsoft also is beginning to greatly cross-integrate many disparate types of software, a practice I covered in the recent report, "Microsoft’s Integrated Innovation: Weighing Up Customer Benefits, Risks." The extent of that integration would appear to be rapidly accelerating with Longhorn products.

During Microsoft’s antitrust trial with the Justice Department and 19 states, trustbusters argued that Microsoft’s type of integration has an anticompetitive effect since developers would be more likely to write for the exposed OS application programming interfaces (APIs) than for separate, competing products. The consent decree covering Microsoft’s settlement with the Justice Department compelled the company to disclose more APIs and communications protocols than the company had in the past.

With Longhorn, Microsoft’s offering dramatically new APIs and the new Indigo communications subsystem could be cause for concern. No longer will developers have to write separate Windows and browser applications. That’s a good development. On the other hand, the move means tighter integration between the browser and Windows and the introduction of new APIs. That could affect competing browsers and software developer interest in them.

In the past, Microsoft’s woo-the-developer strategy largely focused on this kind of application integration and exposed APIs. Longhorn goes further. Microsoft will be tightly integrating the .Net Framework into the operating system, the new WinFX programming model and Visual Studio .Net Whidbey developer tools. As I blogged on Tuesday, during Microsoft’s developer conference this week, company executives pulled off an almost magical feat: Showing an application written with 27,000 lines of code using the current developer tools reduced to three lines of code with the forthcoming Whidbey tools. That kind of feat could only come from tight integration between the aforementioned .Net Framework, OS and developer tools. Heck, Microsoft is putting a compiler into Longhorn.

Web services also factor into this integration strategy. Microsoft’s development tools and Longhorn heavily rely on Extensible Markup Language (XML). Longhorn’s new file system, WinFS, incorporates XML, and Microsoft plans to use XML metadata as a means of creating relationships between disparate data types so that searches will produce more meaningful results.

Microsoft talked quite a bit about schemas, which basically are dialects that define XML data. To date, while on one hand backing Web services standards, Microsoft has taken a more proprietary approach to schemas (You can read about that here).

The other integration area I should mention is security. Microsoft plans to add new security measures with the Longhorn group of products. One, a concept called managed code, I blogged about yesterday. Another is the rights management technology formerly known as Palladium. Under its current conception, Palladium would basically offer a second, secure OS operating mode.

All of these new integrated scenarios share one thing in common: They are, at the least, partly proprietary or tied to Windows. They also could discourage developer involvement in other platforms. I want to be clear: That discouragement isn’t necessarily through any onerous means. Based on what Microsoft showed off this week, the company has started down the path to delivering excellent tools that would let developers more easily create applications. Microsoft shouldn’t be faulted for outflanking its competitors with better tools.

That said, the tools also leverage off Microsoft’s existing Windows monopoly. Competitors don’t share that advantage and may want to plan a response now. Microsoft disclosed enough information this week so that competitors would have plenty of opportunity to respond. After all, Longhorn is as much as three years from release.

For businesses planning future technology purchases, the amount of integration almost certainly will affect their buying decisions. For some enterprises, increased integration might make sense. But, those companies with largely heterogeneous infrastructures might have to make some tough choices. Because Microsoft’s plans are far from set in stone, customers have a chance to offer feedback on the move towards increased integration. Microsoft is particularly sensitive to competition from Linux. Large customers experimenting with Windows alternatives are perhaps in the best position to open Microsoft’s corporate ears.

Posted by Joe Wilcox at October 30, 2003 06:16 PM