Will MOM Catch the Bus?Will MOM Catch the Bus?

Has the enterprise service bus (ESB) architecture propped up or killed off the MOM market? Research firm IDC asks that question in a recent report. "Everything needs middleware, but message-oriented middleware [MOM] sales had fallen flat last year, so we wanted to see if ESB was the reason, as there has been a lot of noise there," says Dennis Byron, vice president, Business Process Automation and Deployment Software Research at IDC.

Byron notes that the "heritage" middleware market — composed of MOM, transaction servers, and access integration — has been collapsing for a few years, matched with steady growth of Web servers, application servers, and integration servers. "The latter three possess the functionality of the older middleware, and they can do things both asynchronously or in real time." Thus, he's concluded that ESB will not be the rebirth of MOM: "MOM will deflate slowly and ESBs will proliferate," adds Byron, who says IDC conferred with 1,000 users and 30 vendors on the topic in the last year.

"We don't think [this shift] really reflects anything new, as [ESB] is a form of MOM from a computer science point-of-view — once you get beyond the marketing hype," says Byron. Because an ESB acts as a shared messaging layer, services participate in the ESB using Web services messaging standards. By supplementing a business's core asynchronous messaging backbone with intelligent transformation and routing, messages are supposed to be transported more reliably. By supporting standard interfaces (such as SOAP, COM, .Net, and JDBC), ESB is supposed to facilitate developer-friendly integration by opening packaged integration approaches to Java and Web services standards.

Byron projects that application servers of the future will have ESBs built in. "For 30 years or so, middleware was built into packaged applications; then, the trend was to separate it out. We think that ESB will drive a trend back to putting middleware in the packages." He points to the fact SAP has built an ESB under R/3 in NetWeaver. "It's an attempt to ease the integration of legacy applications into SOAs based on SAP technologies." Others will follow, Byron predicts. "It will all shake out in the next five years."

In the meantime, Byron believes that MOM — which initially met demands for integration — will give way to XML and Web services standards, which enable enterprises to drive down integration costs. "MOM is unifunctional, whereas ESB allows for multifunctional capabilities — necessary for Web servers, integration servers, and application servers," says Bryon. "Unless a company needs a specific type of high-transaction monitoring or high-performance MOM, the trend will be convergence of traditional and modern middleware."

For CIOs that are really "locked into" MOM, Byron contends vendors will offer continued support. "They will invest new ideas into it if it really matters. However, if it proves too costly for vendors to support legacy MOM, they may push for change when big IT overhauls are slated, as when manufacturing companies add new lines to production." By going to something more modern, but that still includes a message queue, there will be asynchronous and real-time capabilities. "In the meantime, IBM, which owns 74.9 percent of the MOM market, will 'milk' the maintenance of MOM until companies swap out MQ Series for newer products," Byron says.

IBM spokesperson Ron Fravoli says, "MQ is not going anywhere, but like other WebSphere messaging middleware, MQ will also be evolved to support ESB as it evolves in the next year." He adds that IBM expects to make announcements in October regarding additional middleware products supporting ESBs.

Some may choose to swap out in favor of open source software in n-tier architectures. "We believe the demand for open source could be heavy, but limiting factors may be cost (despite the perception by many that it is free) and support," says Byron. (For an in-depth exploration of ESB architecture, see "In Pursuit of Agility" by Fred Cummins)

Susana Schwartz is a New York-based freelance writer specializing in emerging technologies and their impact on IT infrastructure.