The friction in the extremely competitive applications server market turned up considerably this week, as one of the companies perceived as tied for third in market share issued a buyout bid on Tuesday for the company it's tied with. That fact might get buried on the back pages were the bidding company not Oracle Systems, the maker of the 10gAS Java Application Server, which late yesterday saw its $17 per share cash buyout bid for competitor BEA Systems rejected as too low.

"This proposal is the culmination of repeated conversations with BEA's management over the last several years," stated Oracle President Charles Phillips this morning. "We look forward to completing a friendly transaction as soon as possible."

Indeed the pairing of BEA and Oracle has been the subject of much speculation since last March, when financial analysts heard the first whispers. While those same analysts earlier this week were speculating over whether BEA was actually up for sale, the terms of its rejection to Oracle made it apparent to the rest of the market that it would entertain a higher offer.

Some are expecting such an offer to come from either HP, which is seen as needing to either shore up its flagging software resources or get rid of them altogether; or SAP, already Europe's largest software manufacturer.

But it took BEA's letter of rejection to remind analysts that it and Oracle are direct competitors in the application server market. The letter itself reportedly said it could not go into specifics about the terms of its rejection due to the fact that the two are competitors, and such communication would come under scrutiny.

BEA may also be appeasing billionaire investor Carl Icahn, who already has a 13.2% stake in the company and who could acquire more, especially if all this merger talk certifies the likelihood of an upswing. Icahn himself may have pulled the trigger of the starting gun for the bidding war, after having suggested in his own SEC filing that simply entertaining a merger would be in the best interest of BEA shareholders.

This battle is about who gets to deliver business functionality throughout the networks of large businesses. At long, long last, corporations are moving their critical business processes outside of COBOL and onto a more manageable platform. It's difficult to talk about the contents of that platform without sounding esoteric, but Web-driven applications have actually managed to break companies free from the constraints artificially imposed by the first generation of object-oriented services. Now business activities can be based on tangible, malleable models rather than instances, interfaces, and versions.

The big boon for Java applications servers has been AJAX, which shifts much of the burden for process control and user interfacing off the shoulders of the server and more toward the client. Web applications are smarter than they were just last year, but they're surprisingly also faster. Because of this, there's been a lot of movement in the design of applications servers, and no one company has the clear-cut lead in innovation.

So by anyone's standards, the Java applications server market is a healthy and competitive one. Here are the latest market share numbers from the Sixth Annual Java Use and Awareness Study, conducted last December by BZ Research:

App. server brand	Market share
Apache Tomcat	64.3%
IBM WebSphere	36.9%
JBoss	32.0%
BEA WebLogic	23.7%
Oracle App Server (10gAS)	22.4%
Sun Enterprise Java System	19.0%
Apache Geronimo	11.8%
SAP NetWeaver	6.0%

While IBM claims the market share lead with WebSphere in its own advertising, a survey of Java applications server customers typically includes open source software as well, which is why Apache Tomcat has the lead in the BZ Research survey. Meanwhile, ever since JBoss was acquired by Linux distributor Red Hat for $350 million in April 2006, it also frequently lays claim to the #1 spot. If you've noticed the above percentages add up to way more than 100%, that's because most respondents deploy more than one brand simultaneously.

Among the commercial brands, however, note how close they are to one another. No single commercial brand has absolute command of this market.

A $17 per share cash buyout would put the cost of a BEA acquisition at $6.7 billion. But if Oracle were just interested in buying intellectual property for its software portfolio, it would have acquired JBoss in 2005 when it clearly had the chance, and before JBoss was grafted onto a publisher with real muscle. A combination of BEA Systems with Oracle would most likely have the effect that everyone expects: one fewer competitor in the Web app space.

Of course, that's not the image that Oracle wants to project, with its president stating today, "We intend to protect the investment customers have made in BEA's products by supporting those customers and products for years to come."

Oracle's two other major acquisitions in its history include its September 2005 acquisition of Siebel Systems, a company ironically founded by a former Oracle salesperson. That cash buyout for $5.85 billion gave Oracle the immediate lead in customer relationship management software, and what analysts currently believe to be a market share over 50%, especially in Europe.

The other was the January 2005 cash buyout of business applications software provider PeopleSoft - a nasty battle which left Oracle paying $10.3 billion, but left PeopleSoft's upper ranks in tatters. There too, PeopleSoft had enjoyed healthy market share but just behind the #1 player, SAP, with Oracle lagging in the distance.

No one can fault Oracle for being kind to its former competitor's customers, having granted PeopleSoft's customer base a lifetime support guarantee. But in that market at least, Oracle may not yet have seized the overall lead, with some in Europe still claiming SAP has leadership in key segments.

For those who might have studied the consumer software market so long that they might have forgotten what one looks like, the preceding was an example of a competitive software landscape.