InformationWeek

Within the next couple of years, businesses of all sizes and shapes will get their applications in a manner that's far different from today's norm. So, if you're uneasy with the notion of renting software applications as services, you'd better find a way to get comfortable with it. On the other hand, if you're in love with the idea and eager to make the leap right now, it may be best to temper your enthusiasm. There's going to be a major shakeout, and the last thing you want is to be a casualty of it.

The state of the business-applications market just a year ago was that of an industry ready for a good pruning. At the high end, large companies paid lavish prices to license applications, then paid more to have those applications customized to their business needs. At the low end, shrink-wrapped software fit a small business like an untailored department-store suit: a little tight here, a little loose there.

The middle tier of the market sat drastically underserved. Unable to afford the business elite's tailored versions and dissatisfied with the low-end solutions, the middle market ended up with a hybrid. Unfortunately, that hybrid too often was priced like a custom solution but worked like the shrink-wrapped stuff.

Considering this sorry state, it seems logical that new vendors should have popped up to address the midtier's needs. But this didn't happen to any great degree until recently. Midsize companies contracted for a custom application or two if it was critical to the business and resorted to shrink-wrapped software otherwise. These custom applications had to be virtually flawless to justify the investment in them. From the developers' point of view, there was generally inadequate implementation expertise serving this market to support customizable applications geared specifically to this size company. Perhaps more important, Microsoft rules the market for midtier applications. What developer would go up against a monopoly known to seize new markets away from pioneers?

The mainstreaming of the Internet brings three very important factors to the forefront. First, the explosive demand for bandwidth and the commensurate drop in cost, driven primarily by the data-transmission needs of the Internet, enables centralized application deployment. Client-server enterprise resource planning implementations are notoriously intolerant of slow WAN links. Historically, slow links have vexed implementers that attempt to support remote sites within an en-terprise. If an ERP application won't support remote sites, then there's no room to build a business supporting entire remote enterprises.

Second, the Java-enabled browser is the first ubiquitous presentation platform adequate enough to deliver business applications. While most ASPs are still pushing fat clients for end users, the real ASP model will eventually support only relatively thin clients. HTML alone isn't up to the job of supporting these applications; Java adds just enough smarts to the front end to make the user interface truly usable.

Third, the perceived requirement for businesses of every size to become E-commerce-enabled will require new back- and front-office applications, thus opening the door to different application models.