From Computer Business Review, a sister publication
When industry veterans Bill Coleman of Sun Microsystems Inc and Ed Scott of Pyramid Technology Corp got together to ruminate on the potential opportunities for a new information technology business venture back in late 1994, they noticed a gaping hole in the emerging middleware sector. No single company offered a complete range of middleware software.
By Steve Bell
Coleman and Scott promptly went out and enlisted the services of Alfred Chaung, one of Coleman’s colleagues at Sun, in a bid to plug that gap and, by early 1995, BEA (Bill, Ed and Alfred) Systems Inc had been born. Getting venture capital funding was not quite so easy. When the company first approached Warburg Pincus with the idea of creating an all-encompassing middleware company, the bankers were lukewarm, advising the founders to come back when they had a detailed business plan. It was obviously a persuasive document. When BEA returned the second time, Warburg Pincus ended up giving BEA $50m, the largest venture capital investment in an information technology company that Warburg has made. The signs are that Warburg Pincus’ investment is paying off. A little over two years after BEA’s birth, it has grown from six to 650 staff, from seven offices last summer to 29 today, and boasts partnerships with almost every major hardware vendor. Revenues are also increasing.
In the first quarter of this year, its first as a public company following its flotation on the US NASDAQ exchange in April, BEA posted revenues up more than 300% at $30.4m, leading to a net income of $1.3m, compared to a loss of $3.5m last time. It sounds like a fairy tale beginning for a start-up company, but Coleman, now BEA’s chief executive and Scott, executive vice president of worldwide field operations, prefer to put the success down to astute planning. What was missing from client/server technology was the ability to perform the kind of functions that had been performed on mainframes. We thought if we could bridge the gap between mainframes and client/server we would be able to build a great company, says Scott. In order to take the first steps in fashioning itself as a full service middleware provider, BEA has been on a whirlwind acquisition spree. It snapped up the Information Management Company and Independence Technologies, the two largest suppliers of Tuxedo transaction process monitoring software. The company has also formed a partnership with Novell, the creator of Tuxedo, to exclusively develop and distribute the product, giving BEA the lion’s share of the on-line transaction processing software market. Bolstering these moves, earlier this year the company also bought the highly rated DECmessageQ from Digital Equipment Corp, considered to be the fastest and most advanced message oriented middleware on the market. It also snapped up another highly regarded piece of middleware from Digital – ObjectBroker, an object request broker. BEA now says that it has all the constituent parts for a full-scale middleware solution and its primary focus now is to extend the functionality of all of its offerings. Digital’s decision to sell these supposedly highly rated pieces of software suggests, however, that the market is not as bright as BEA makes out. According to Gerry Sunderland, DEC’s enterprise software marketing manager, the reason DEC decided to sell its products is that it believes middleware competition is going to get very tough. The key competitors facing BEA are formidable – Microsoft Corp, IBM Corp and Oracle Corp. While middleware as a concept has been around a long time, the market is only just starting to take off and the large vendors are still in the process of positioning their products. Although analysts believe BEA has a good understanding of the middleware market and the direction in which it is evolving, they question both BEA’s ability to cope with the long sales cycles associated with middleware products, and its ability to respond to the threat posed by Oracle’s network computer architecture. BEA acknowledges that the Network Computer versus personal computer debate is going to have an impact on its business and is now attempting to tailor its all of its products to deal with a Network Computer architecture.
The biggest challenge is, however, probably posed by Microsoft. In BEA’s favor is the fact that Tuxedo has been around about 15 years while Microsoft’s comparable offering, Viper, is relatively new and proprietary, only working on NT. BEA’s products, in comparison, can operate across most hardware platforms and is currently being used across 50 platforms, including NT. Certainly BEA is already viewed as a heavyweight. According to a survey conducted by International Data Corp, BEA’s market share in the transaction processing sector has grown from 30% to 39% over the last year. IBM has also opted to use BEA middleware for its new high-end SP2 Unix platform rather than its own Encina product. The reason? BEA’s product is faster and more functional, said one IBM Corp product manager. Karen Scherberger of Gartner believes that the key threat facing BEA is not another company but its own ambitions. It is, she says, in danger of overstretching itself, and is perhaps establishing too broad a remit. So is BEA a hot prospect? In the immediate future, the answer is probably a resounding yes, but once Oracle and Microsoft begin to roll, the company may find that the safest place for it to be, is tucked away in a niche.