At International Business Machines Corp., a company more used to setting the pace than following the leader, the acceptance of client-server technology has been a slow and arduous process.
A study released earlier this week by the IBM Consulting Group suggests the need for an initiative in this area. But, unsurprisingly for the huge mainframe vendor, IBM cautions its customers about the potential pitfalls of PC-based systems in its report.
The study -- based on a survey of 24 major companies in industries ranging from financial to transportation, insurance to utilities -- revealed that the honeymoon seems to be over for most client-server users, but the marriage won't end. The majority of survey participants said client-server technology proved useful in many business applications, but it could not completely replace mainframe computing -- especially in some of the more basic, data-intensive back-office operations.
Admittedly Slow on the Uptake
Nevertheless, banks and other companies have welcomed client-server technology with open arms, and IBM executives have openly admitted to the company's failure to react fast enough to this trend.
"As Mr. [Lou] Gerstner has said, IBM missed this one," said Michael J. Sinneck, the vice president for IBM's Consulting Group.
But Mr. Sinneck was quick to point of that "the mainframe is not dead" and, in fact, remains the bread and butter of largescale computing.
"Most companies didn't do a very good job of judging value," he said. "The cost of these [client-server] solutions was not understood, and the benefits were felt rather than measured."
The study focused on the companies' deployment, cost assessment, adaption, and skills in their use of client-server hardware and software. According to Mr. Sinneck, one of the key factors that influenced these changes was the increased involvement of business units, rather than the information technology specialists, in making decisions about operations.
In general, respondents said their client-server applications were not "mission critical" to their business, and the systems cost more and took longer to implement that they had expected. Although almost nine out of 10 of the survey participants said their systems had to be redesigned from the original plan, about an equal number said that they were satisfied with their results.
The survey -- started in August 1993 -- was the brainchild of Lou Gerstner, IBM's chairman and CEO. It stemmed from a meeting he led last May, in which many customers expressed mixed interest and confusion about client-server usage. It was hoped that IBM's ensuing study would demysticize the new technology and identify better ways to manage it and marry it to existing mainframe systems.
Although this may not offer much help to companies that have already installed new systems, it did help to "reconfirm the issues" for banks like Chase Manhattan, according to Craig Goldman, Chase's chief information officer.
"Client-server is really not the cheap alternative, and a lot of us have learned that the hard way," Mr. Goldman said. "The real challenge has been, and still is, that it's not all integrated together yet."
Although the bank has been implementing client-server applications since 1988, Mr. Goldman said the technology still has not fully matured and there is a lack of experts in the field.
And the increased number of vendors offering client-server products is a two-edged sword. On the one hand, it increases the expertise in client-server technology. But not all of the products are compatible, which can cause big problems.
A few years ago, Chase did business with "one or two vendors ... now we deal with hundreds," Mr. Goldman added.
"All the products don't necessarily play well in the sandbox together," Mr. Goldman said.
The IBM report suggests that most companies discovered a happy medium by blending their already existing mainframe systems, which hold the bulk of their data, with the newer distributed technology.
Common Fallacy
"There is a concept that client-server technology means without mainframe." said Robert Howe, a vice president at IBM and general manager of the consulting group.
The IBM study concludes that this perception is a fallacy, and should be corrected.
Back-office operations in particular, such as item processing, are a "lousy fit" for client-server computing, Mr. Sinneck said.
IBM sees the best match for client-server in customer service, decision support, and workflow communications applications -- like the popular groupware applications such as Lotus Notes.
Although IBM is now coming around to embrace the idea of client-server -- as one of the company's six key initiatives under Mr. Gerstner -- critics still hold that the computer giant has lost too much ground and is still doesn't see the handwriting on the wall.
Like the Auto Industry
"They are fighting the aggressive movement of PC suppliers moving up the data processing food chain," said Richard Crone, a senior manager at KPMG Peat Marwick Consulting.
Mr. Crone attributed the growing presence of PC-based systems and IBM's hesitant acceptance of this trend with the relatively recent availability of "mainframe-class computing on desktop." He compared IBM to the major Detroit carmakers of the late 1960s and early 1970s that refused to realize the threat posed by their foreign counterparts.
"These PC-based systems are the small Japanese cars that are taking over the market," he said.
