Portfolio
SAA REVOLUTION NEEDS TIME FOR EVOLUTION
Source: SOFTWARE MAGAZINE Published: Oct 1990
In the early 90’s, IBM made a major effort to create Systems Application Architecture (SAA) as an industry standard for software development. Ultimately, SAA was a failure due to the expense of underlying software and hardware platforms and the emergence of PC centric client/server computing as a dominant application architecture in the 90’s. In this early article, George Schussel offers some intelligent commentary on potential pitfalls of the SAA approach.
A few comments from that article follow:
The jury is still out on IBM‘s plan for providing application standards across platforms. But before the industry can fairly judge the SAA blueprint, it must have time to evolve. Current drawbacks, such as expense and lack of products, may not be an issue over time. With IBM‘s track record as an industry leader, organizations committed to Big Blue would be well advised to prepare the way for SAA migration.
IBM‘s thinking about distributed applications has centered on data systems architecture and distributed SQL DBMS. While the concepts of distributed SQL are elegant and sophisticated, the delivery of these capabilities is now planned over the next five years. Client/server SQL solutions are simpler and might be deliverable earlier.
IBM’s workstation standards are built around OS/2EE. This environment requires a large 80386 PS/2, which costs around $10,000. Since few existing PCs have enough power to run OS2/EE, this means substantial added outlays for most potential SAA users. DOS/Windows 3.0 has a GUI that is almost identical to OS/2’s Presentation Manager. Support for Windows 3.0 on clients would not require a migration away from DOS and would require less investment in hardware.
