Coming-of-age for app servers - the burgeoning application-server software market - Internet/Web/Online Service Information

The chaos in today's application server market will eventually lead to consolidation, observers claim

Analysts agree that the application server market is coming of age. Defining the technology is still a point of contention, however.

"I'm not sure there is a good, single definition," says Scott Lundstrom, vice-president of research for Boston-based AMR Research Inc.

New reports focus on the emerging role the Internet will play in enterprise applications, listing a range of vendors and products. While the field could prove lucrative for some, researchers still haven't pinned down what it all entails.

Steve McClure isn't sure there's even a clear understanding of the concept. "It's a pretty fluid thing to try to define," said the director of Internet tools research for International Data Corp. (IDC) in Framingham, Mass. Oracle Corp.'s Application Server product used to be called the Web Application Server, he said, claiming the move proved how uncertain the market really is at the moment.

"It's an early market, and in an early market there's a certain amount of diversity," Lundstrom argues.

One analyst's diversity is another's confusion. Today's bevy of competing developers and standards amounts to nothing less than chaos, according to Ted Schadler, a software analyst with Forrester Research Inc. in Cambridge, Mass.

Application servers connect thin clients with an array of distributed computing applications, including databases and enterprise resource planning software, according to a report co-written by Schadler released in August. He describes the technology as consisting of middle-tier, Internet-based computing platforms, with more than two dozen products currently available.

Lundstrom breaks the market into three "camps" of developers. Database vendors, tool developers and independent or stand-alone companies are all vying for their share of enterprise-level customers, he says.

A glance at an August report from Dataquest Inc. reveals another litany of product categories, including database management systems, transaction-processing monitors, object request brokers, object transaction monitors and Web application servers. Despite their different backgrounds, however, vendors are all competing in the same space, Dataquest claims.

As for explaining how that space is structured and what business strategies will win out, the Gartner Group Inc. subsidiary says its study "gracefully brushes past the other two questions until later this year when the market has matured slightly."

At stake is a growing, if ill-defined, market for supplying business applications across enterprise systems without having to install software on desktop systems or adding more servers to networks. Dataquest estimates 1997 application server product licence revenues to total about $350 million (U.S.), excluding income from MVS (multiple virtual storage), used on IBM mainframe computers.

For its part, Forrester predicts the market will grow to more than $2 billion by 2001, with only a handful of vendors remaining. Microsoft Corp. will be among the leaders, according to Schadler.

"Microsoft's vision for domination here is to extend NT," he says.

While the software giant may have built up a set of business applications, as well as a strong database contender with SQL Server 7.0, it still needs to integrate the components to produce a winner. "They've got to make the parts work better than they do now," Schadler says, adding the firm also needs to develop specific products for different sectors.

Lundstrom is also bullish on Microsoft. New technology to be included in Windows NT 5.0 will help make integrating two-tier or client-server systems into application server environments that much easier. It will also make it easier to expand networks as needed.

Other key players also stand to gain ground. "This becomes a real footprint war," Lundstrom claims, arguing that companies with the strongest presence and most comprehensive coverage will win out in the long run.

Schadler predicts IBM Corp., Netscape Communications Corp., Oracle and Sun Microsystems Inc. will all eventually enjoy the lion's share of the market, while "wild cards" such as Inprise Corp., Sybase Inc. and BEA Systems Inc. could also remain strong. Survivors will have to combine effective middleware offerings with a variety of business applications to build a reliable, expandable product, he says.

McClure maintained that Oracle and Microsoft are still at loggerheads when it comes to enterprise networks. Oracle is promoting a "server-centric world," which is at odds with Microsoft's approach to distributed applications on numerous servers, he said. Even with third-tier support in NT 5.0, Microsoft is "still trying to sell lots of NT servers," he claimed.

Schadler originally predicted gains made from increased application server sales would be slowed by Microsoft-led price cuts, what he considers to be a typical strategy from the Redmond, Wash. company. Two factors, however, have made him revise his thoughts on the matter.

"We believe the price of NT on bigger boxes will go up," he says, explaining that enterprise buyers are used to paying more for high-end merchandise. Microsoft itself also needs to grow, he adds.