Just about every new technology or new way of doing things goes through a period of over-hype, and application service provision is no exception. During the past few years, the application service provider (ASP) model has gone from the greatest thing since sliced bread— applicable to almost any situation—to much more modest expectations. The ASP model CAN work, but the model's applications seem to be more industry-specific (as they were when the model first became popular), and the number of total subscribers remains small when compared to total computer users.

We've discussed some of the reasons for this low adoption level: Even apart from inertia, people continue to use inhouse applications because of availability and security concerns. ASPs are still useful for some situations. Companies that don't have IT staff—or companies that benefit from scale because they can't afford to purchase, support, or maintain the high-end applications they need—can benefit from outsourced application and data hosting. Although there aren't any "nevers" when it comes to technology (20 years ago, who would have thought that PCs would be as widespread as they are now?), at this time the ASP model of software as a service isn't very widespread.

Given the the ASP model's relatively low saturation level, you might reasonably ask why Microsoft thinks it can make software as a service work. "ASPs with solid business models, efficient operations, and new services are continuing to deliver value to their customers," says Janet Smith, lead product manager of Microsoft's Windows .NET Server Marketing. "Software as a service continues to create diverse and expansive opportunities for ASPs providing users with access to resources, data, and services any time, any where, on any device. Today, a variety of partners are successfully delivering software as a service to customers via the ASP model. For example, CertifiedMail.com is using .NET as a platform to deliver upon Software-as-a-Service, offering value to their customers."

I agree that successful ASPs offer value that customers recognize and can't get anywhere else, but good ideas and good technology to back them aren't always enough. For ASP adoption, image is important, too, and Microsoft has image problems—especially related to security. When Gartner recommends that people not use Microsoft IIS because it isn't secure, that recommendation has to hurt any kind of hosting—-even if Gartner is wrong. Inertia is also an issue. No matter how easy it is to integrate a new service with an existing one, the integration still costs something in terms of time and effort. A company that wants to push a new messaging solution or online access will have to make the case that the new solution is more valuable than the time required to implement it.

In the end, Microsoft will make software as a service work if it wants to do so. First, the company has deep pockets. Unlike most other ASPs, which need venture capital because their customer subscriptions don't cover costs, Microsoft has a profitable business it can use to subsidize its ASP services, if necessary. Second, Microsoft has the marketplace; its OSs and application software are widely used and switching is not as easy as fans of StarOffice (a free Java-based productivity suite from Sun Microsystems) might have you believe. If the company that makes widely used applications decides that it's going to offer those applications as software as a service, that company will be able to pull it off. The bottom line, though, is that the past couple of years have demonstrated that an ASP's success isn't just the success of a good idea or good technology, but widespread adoption by a paying customer base. If Microsoft succeeds with this model, it won't be just because .NET works, but because the company has the money and the muscle to pull it off.

And on that note, I'm sorry to say that after 2 and a half years of Application Service Provider UPDATE and its predecessor, Thin-Client UPDATE, this is the last issue of the newsletter. I'll continue to write about terminal services for Windows 2000 Magazine; the November B issue will include a Web exclusive about Citrix MetaFrame XP for Windows and Feature Release 1 (FR1), and more articles are in the pipeline. (Please keep those terminal services questions coming; the best articles I write are based on how-do-I-make-this-work questions.) And beginning next year, I'll explore another great enthusiasm of mine—VBScript scripting—when I write the Scripting Solutions column each month. I hope to see many of you there! (Hey, why not? You can write scripts that apply to terminal services, too.)