Who will buy BEA?

If all that talk at BEAWorld about thawing frozen assets leaves you cold, here are some hot tips on who's most likely to warm things up by liquefying BEA's own assets in a takeover bid:

• Starting off by noting a trail that's now long gone cold, Oracle at one time had BEA high up its shopping list. But after digesting half its lesser competitors in the enterprise software space, the company has now called a halt to large acquisitions, CEO Larry Ellison told customers last week at the company's annual conference.

• Another one to rule out from the start is Computer Associates, whose track record of buying up failing software companies and living off their maintenance revenues rules BEA out of consideration. Open source competition means that BEA won't have any maintenance revenues in a few years' time. That's why it needs to find a buyer, fast — but it needs to find a buyer who can recognise immediate value from the deal.

• A lukewarm possibility but still pretty unlikely is IBM, which has been concentrating instead on acquiring market share from BEA in its core application server business. IBM does need to plug gaps in its SOA strategy, but since BEA is missing almost as many pieces in its own, getting together would help neither company.

• Some observers have been quietly tipping HP — but usually only so that they can go on to crack jokes at HP's expense, recalling its previous acquisition of an application server vendor, only to close down the division a couple of years later. HP's SOA strategy is in dire need of a shot in the arm, but if it made a move on BEA, most people would say it was just pulling their leg.

• A long shot but not as unlikely as it may at first sound is Microsoft, which was at BEAWorld today announcing an alliance with BEA's open-source rival JBoss. Acquiring BEA would instantly give Microsoft the cross-platform credibility that's always held it back in the enterprise market, while at last giving BEA's WebLogic Workshop tool the opportunity to command the mass market developer appeal that BEA chief Alfred Chuang has always envied.

• A much hotter prospective suitor is SAP, which could do with shaking off the legacy of its proprietary ABAP language. Acquiring BEA would show how serious it is about becoming a platform vendor, although the downside of moving too fast in that direction is that its customers might just start to forget that SAP also sells applications ...

• Most intriguing of all, though, is the idea that Indian IT outsourcing giant Wipro might put in a bid. With the kind of services opportunities that BEA's application server would offer Wipro, it wouldn't care two hoots about continuing declines in licence revenues; it might indeed decide to open-source the lot and just make money from the services. Wipro would also gain a lot from BEA's intellectual property investments in development tools, especially its moves to make Workshop more accessible to mass market developers and other non-programmers. It may sound crazy, but it could just make sense.

In the meantime, BEA is going to continue to pursue its ambitions as an independent company. It might just survive if it can transition its main business across to next-generation application assembly, which as I've previously hinted (but perhaps should spell out more explicitly some time), is the hidden agenda behind the company's Liquid Computing strategy. Hence its purchase today of M7, one of the more successful and longest established of the motley band of specialist application assembly vendors. This is a smart buy that adds considerable substance to BEA's ambitions, and demonstrates the company is serious about making them come alive.