ABSTRACT
One of the outcomes of the client/server computing revolution was going to be the elimination of these expensive, proprietary, dated platforms by a new generation of low-cost servers based on industry standards or
de facto
standards. Client/server never achieved this lofty promise. In large part, this failure was due to the lack of a solid business case for porting the legacy applications to new platforms. The legacy platforms provided fault tolerance and 24
×
7 operation that were unavailable on new servers. In addition, the cost of the legacy platforms began to fall and the platforms became much more adept at supporting the new open technologies. The cost of the IBM mainframe in terms of dollar per MIP, for example, has fallen drastically in the past decade, and the standard mainframe operating system now includes support for TCP/IP as a no-charge feature.
