In-Depth

IBM: Once and future king?

• By Jack Vaughan, Michael W. Bucken
• June 15, 2001

Your executives are appearing in federal court day after day. Large portions of the industry are calling for your breakup. You are a big company at the top of your game, but are confronted with technology paradigm shifts that come with increasing frequency. Microsoft in 1998? Yes, but also IBM in the 1980s.

How things will work out for the young turks of Redmond remains to be seen. But IBM's path after an anti-trust suit can be traced. There were some years of stumbling, and some bloody noses. There were significant encroachments by upstarts like Sun, Lotus and the redoubtable Microsoft, as well as incursions by established forces like Hewlett-Packard and Digital Equipment, to name just a few. The company that can trace its lineage back to the 1896 census and Hollerith's tabulating machines seemed ready to join other mainframe-centric firms that did not fully make it out of the punch-card era.

Such talk has not been heard too much lately. Today, IBM is a new company with a new stature and luster. No firm has ever approached Big Blue in the breadth of its offerings. In recent years, its software portfolio -- which accounted for $12.8 billion or 16.3% of total 1997 revenue of $78.5 billion -- has been bolstered as the company absorbed some former competitive upstarts, notably Lotus and Tivoli. Under Chairman and CEO Louis Gerstner, the firm has fairly deftly "played the Web" for high-level impact -- which reached an initial apogee of sorts in 1997 with the Web show that was the "Deep Blue" computer's chess conquest of world champion Gary Kasparov.

As the '90s draw to a close, IBM's development tool strategy has largely recovered from the debacle that was its AD/Cycle strategy. Tools and object software tacks have gone through a sometimes dizzying series of changes, but it has been a case of constant refinement and increasing clarity. Integration is a software area in which the company, of necessity, has had to shine, and its MQSeries, in particular, seems to have found a home in big enterprises.

In software management, the acquisition of Tivoli has had a profound impact, rationalizing formerly disparate IBM solutions. The company has emphasized services for high-end customers, as might be expected from Gerstner, a former Mc-Kinsey & Co. management consultant, but seems prepared to roll out packaged solutions (built around the Internet) for less-affluent customers. The industry was also reminded that Bill Gates still lists IBM as his number one competitor after a recent speech to European I/T heavyweights.

Through it all, most viewers agree, IBM has emerged as a much different company. The selection of RJR Nabisco's Gerstner as chairman and CEO in April 1993 is usually cited as a turning point. Many thought his mostly marketing background would not translate effectively in the high-tech world. After initially misgauging the importance of the Internet, IBM under Gerstner has levered the Web to push forward all manner of IBM products. This is especially true of development tools.

Java evangelism

As the history of computing is so entwined with the history of IBM, it is little surprise that the company supports a full range of computer languages. [Jibes one IBM hand: "We are still supporting versions of Cobol in Japan that we stopped selling in 1975."] As the Gerstner era began, IBM was beginning to rationalize its tool offerings under the umbrella of the VisualAge family. Middleware software strategies centered on Unix flavors of CICS, DCE, and a little-known experimental messaging software that became known as MQSeries. Also promising as a distributed system middleware was Corba-based SOM and DSOM. Object-based tools, middleware, and, for a time, operating system strategy in those days also came to include Taligent technology derived from an application system environment effort launched with Apple. DSOM has given way to IBM Object Broker as the Corba ORB, Java as the language of choice, and Enterprise JavaBeans (EJBs) as a component standard. The constant goal has been to support heterogeneous computing.

"We have been trying to get a complete portfolio and to take it across platforms," said Steve Mills general manager of IBM Software Solutions. This on-going effort, he indicated, has led to a more cohesive set of messages that have, in turn, been rationalized under the Internet-savvy e-Business framework.

VisualAge has proven to be a useful vehicle for a variety of tool tacks. In the early '90s, IBM began to position VisualAge for Smalltalk as the way for typical (Cobol-intensive) enterprises to migrate to objects. Big Blue's moves quickly began to overwhelm the small handful of dedicated Smalltalk tools vendors. But that move paled in comparison to its full assault on object programming under the flag of Java.

Java-originator Sun Microsystems and Netscape Communications have tapped IBM's software building and marketing expertise. Last year, the three firms formed the Java Porting and Tuning Center to ensure consistent distribution of fast and reliable Java implementations. [Netscape subsequently reduced its role in that effort.] With Sun and Oracle, IBM has driven the Java Servlet and Enterprise JavaBean effort. If Java has been the lightning rod attracting anti-Microsoft/anti-Windows sentiment, IBM has been the dependable contractor building the house on which the lightning rod resides.

Said Mills: "We've done a good job with Java getting through a first level of adoption. The next phase is to build more significant apps and more server-centric apps. Performance has been a focus. We are the torch carrier for Java in the enterprise and on the server."

Mills was asked if there was any guarantee if this Java standardization would end up differently than once-hopeful predecessors such as OpenDoc, DCE, DME and Taligent. "I don't believe in guarantees," the pugnacious battler shot back. "But [Java] is not really new. All of this is a collection of computer science ideas that goes back three decades."

Still, there are differences he cites between EJB and OpenDoc. "OpenDoc was an attempt to serve in several directions. There were too many paths. EJB doesn't have to grapple with [several] languages. The language is Java," he noted.

"DCE is the grandfather; Corba took it forward and added object structure. We're very optimistic in this round; we have the will to perform up to expectations," added Mills. "But Java will not solve every problem. Java solves a big problem I was dealing with for years. It provides a high-performance, compiled environment for server-side infrastructure."

Mills described IBM's tools strategy as a process of refinement. "We've thrown away as much code as we currently ship," he said, referring to the ORB. "The lessons of the past have been fully absorbed.

"The problem with Corba, initially, was it demanded that developers know C++," he continued. "There was a need [for] a simpler abstraction so you didn't need to be an expert in infrastructure. True, Java is a C derivative. But it's a good match."

Said Mills: "It's a better match than we found in Smalltalk." Yet, indicating IBM's breadth of language support, he noted "we still have a nice business in Smalltalk."

Former IDC analyst Evan Quinn put IBM's Web positioning in bas-relief in a June IDCFlash when he wrote "With Microsoft carrying the DOJ burden, and Netscape struggling to redefine itself, the Web software market avails itself to a third major force. IBM always possessed the potential to truly challenge Microsoft and Netscape. That potential [has become] kinetic."