WebSphere has been a remarkable market success story for IBM. Not too long ago, Big Blue was being urged to break up into separate companies. Although it spun off some of its properties, IBM remained as one entity, and began working to connect the vast number of its offerings. WebSphere was key in making those connections.
When WebSphere first came out, many critics said it was more sizzle than steak. Suddenly, messaging products and integration brokers shared the same name as the company's new J2EE server. But programming them required different tool sets, and they said that true integration was a work in progress. But over the years, more things worked together more effectively, and the WebSphere name came to mean something. Today, developers from across the development spectrum are delving into Java to get a better handle on WebSphere as a universal means for application integration.
Numbers seem to bear out IBM's WebSphere marketing strategy. Gartner this week estimated that IBM holds a 37.2 percent share of the 2004 market for worldwide new license revenue for application integration and middleware.
Meanwhile, IBM continues to expand its WebSphere focus, most recently placing its DB2 Information Integrator under the IBM WebSphere umbrella. Thus, at least one part of DB2 is getting the "WebSphere" treatment.
While responsibility for sales still remains with the massive IBM DB2 group, the name of the product going forward is IBM WebSphere Information Integrator. In true Big Blue fashion, a number of once separate products today dwell under the WebSphere Information Integrator banner, including content integrators and federation integrator products. Also in the WebSphere Information Integrator mix are "plain old" WebSphere J2EE application servers. Depending on the complexity of the integration chore, skills required here resemble more a team than a person, since J2EE (and EJB), DB2 SQL and XML are just a few of the development talents such projects demand.
SearchDomino recently talked about the changes to WebSphere with Eric Sall, IBM program director, Information Integration and Information Management. He told us that IBM's new moves in data integration revolve around three areas: handling structured and unstructured data; providing autonomic and metadata capabilities that better explain data sources to developers; and enterprise search technology.
"WebSphere Information Integrator is middleware architecture, a set of services," said Sall. "And it's delivered in a number of editions. Different editions focus on data federation, replication, enterprise search and content management." These are all needed, Sall said, because a "full relationship to a customer is broader than just structured data."
[Ed Note: You may recall an earlier interview here with Sall, who discussed Notes DB integration issues with us. A former Lotus hand, he came to IBM by way of Venetica, a middleware provider that offered a link between DB2 and digital assets stored in Lotus Notes, Documentum, FileNet, and other data repositories.]
Last year, the Integrator added search capabilities that further bring the data base world of queries in line with the Web world of searches.
"People gravitate toward search," explained Sall. "Search is becoming the universal access method for all kinds of information, and people are very familiar with the search interface."
"What we have seen from our customer base is that people want to get that same easy access within a company that they get with search engines outside the company," said Sall. With the enterprise search capabilities now available through WebSphere Information Integrator," he said, "you search your Word docs, your applications' data, and your in-house databases" as you do the Web.
All and all, it appears that data integration will be the next country for WebSphere to conquer, as the software seeks to solve the age-old problem of heterogeneous data. Clearly, Java and J2EE will be pitched by IBM as a standard programming tool for helping to make the heterogeneous look homogenized.