Information Age: News, analysis & insight for IT & business leaders

Sooner or later, every new market faces consolidation. When that time comes, the transition can be painful for both the companies and their customers alike.

Persistence Software, an early pioneer of application server technology, is a case in point. Revenues at the company are fading fast. In fiscal 2001, sales fell 23% to just $19.4 million, a figure that compares poorly with market leader BEA Systems. In the year to the end of January 2002, BEA achieved revenues of $975.9 million.

In its current fiscal year, BEA will become a billion-dollar company, while Persistence – which has been in business almost five years longer – will struggle to survive. Why did Persistence stumble?

The company readily admits that a number of strategic mistakes were made. For example, it released a general-purpose application server package that, although powerful, lacked the ease-of-use needed to win widespread market acceptance.

Furthermore, it did not introduce full support for the critical application server standard Java Two Enterprise Edition (J2EE) until August 2001, more than a year after BEA and iPlanet first trumpeted their J2EE-compliant servers. Until then, Persistence's flagship had been the PowerTier Application Server for C++.

"One of the tactical mistakes we made was to introduce a general purpose application server when our real skills were in managing distributed data. So BEA, for example, had more user-friendly front-end interfaces, which allowed people to do more mainstream applications more easily," says Ed Murrer, senior vice president of marketing.

Persistence was originally founded in 1991 as a supplier of tools for mapping from object-oriented databases to relational databases and it is to this niche that the company is now retreating.

This move has been met with approval by analysts, such as Gartner Group's Deborah Hess. The company has "great, valuable" technology, she says, which is well targeted at a market that she says is hotting up. "This ability to do object-to-relational mapping and data caching in memory is a market that is not really visible to the end user, but it is a really important market," she adds. Providing fast and synchronised access to data is vital for organisations handling large volumes of online transactions across distributed networks. But this requires a far more complex technology than caching software on a web server.

One method is to cache frequently used transactional data on an application server so that users do not access a database directly, thereby reducing load on the server and improving performance.

PowerTier's mapping tools provide the capability to bi-directionally map Java objects or components and relational database management system tables. "This is not easy to do by hand – especially for methods and interfaces – and Persistence has applied its patented object-to-relational mapping technology to this task," says Hess.

Persistence's two additional products are EdgeXtend and Dynamai for customers not using its PowerTier AS. While Dynamai provides distributed caching software for users of BEA's and IBM's application servers, EdgeXtend is targeted at users of other application servers.

But despite her enthusiasm for the technology, Hess doubts whether Persistence has the finances or marketing clout to compete effectively against BEA and IBM. Her conclusion? "I think Persistence is ripe for acquisition."

"Organisations that have already implemented or are planning to roll out large web-based applications are going to need secure caching and object relational mapping tools in the future," says Hess. But while Persistence's technology is still likely to be around in a few years time, the company may fail to live up to its name.