Copyright 1994 by William F. Zachmann, President, Canopus Research.
Downsizing, Client/Server Fads Fade, Are Replaced By Reality
Distributed microprocessor based systems unquestionably do provide tremendous new opportunities for organizations that use them wisely. Any notion, however, that traditional mainframe and minicomputer systems are thereby rendered obsolete and irrelevent is as foolish as it is naive. The relevant criteria by which to pick the best technologies to use to build the sort of industrial strength systems required to run the contemporary enterprise are defined by what provides the most cost-effective, reliable, practical solution to its information systems requirements, not to whether they are traditional mainframe/minicomputer vs. client/server solutions.Those who religiously assert that downsized client/server approaches are always to be preferred to traditional mainframe and minicomputer-based approaches quite literally do not know what they are talking about. Sometimes downsized C/S options are preferable. Sometimes they are not. All the time, however, the sensible course for any organization is to base technology decisions upon a careful examination of the range of options actually available to it, not on the opinions of some wild-eyed zealots who claim to know the correct answer without actually studying the specific problems and especially not those all-too-numerous apostles of downsizing, who, in fact, have no real experience with the real problems of the deployment of real world information systems of significant size and scope.
I invented the use of the word downsizing in the present context in an International Data Corporation Office of Technology Assessment research memorandum I wrote in the spring of 1985. I promoted the concept (and the term) in articles and letters for nearly five years before anyone else even bothered to use the word (except, occasionally, to criticize me for having such a silly idea about the future of the industry). I've researched and studied downsizing and its potential for nearly a decade now. And I have not the slightest double that the majority of those who have jumped on the downsizing bandwagon over the past few years are little different from the famous primitive tribe who, upon first hearing of the benefits of electricity, could not wait to put their fingers into the nearest light sockets in order personally to realize its benefits except insofar as that it is their unfortunate clients whose fingers they propose to be so placed.
There currently are many cases in which downsizing made sense. However, there are also many cases where the most sensible, reliable, cost-effective way to meet the information system requirements of an organization will be to use traditional mainframe and minicomputer systems for the job. What's more, even when downsizing or client/server approaches have a substantial contribution to make, traditional mainframe and minicomputer systems still have not just an important but an essential role to play as part of the system as a whole.
Those are facts that are beginning to dawn, not only upon the awareness of those who are responsible for the development and deployment of information systems (most of which have known better all along anyway). but on the general corporate executives who were more likely to be drawn by the siren call of the downsizing cultists with their totally unrealistic promises of dramatically slashed IS expenses via downsizing. The fantasies of running your typical Fortune 2000 organization on a network of personal computers with systems built by end user in Visual Basic (or whatever) are rapidly giving way to a more realistic sense of what downsizing is really able to accomplish.