Steve Hamilton of Comdisco UK reckons that talk of swappable technology between 3090s and the forthcoming Summit machine is not unreasonable. He forecasts two versions of Summit, the first of which will have 3090-based technology, and the second will be an upgrade. Speaking at the UK Computer Management Group meeting in Glasgow last week, Hamilton told users that the growth of second user marketplace and plug compatible manufacturers’ success means there are good opportunities to save money. He strongly advocates that users keep their requirements in mind when looking at upgrade prices. It’s difficult in the UK and Australia where upgrade lists aren’t available, but he says that price should be justified by a commensurate improvement in performance.
Expensive
A 600E to S upgrade used to be $3.5m, the S to J is now $4.7m. Yet, that only brings a 7% to 14% incremental improvement in performance, and it is largely dependent on expensive upper level memory. The pressure that Hitachi and Amdahl are exerting on IBM also means that users are in a fairly strong position, and if the Amdahl machine based on Fujitsu’s EP2000 is announced in early 1991, then the competitive pressure looks set to continue for a considerable time. Also looking to the future was Tony Clifford-Winters of Amdahl (UK) Ltd. His address was entitled the Long Term Significance of the IBM Repository, and he believes that will be most pronounced for operations departments, which will lose immediate control of mainframe operations. Clifford-Winters says he approves of the AD/Cycle philosophy since it is a total life cycle approach, and also that IBM has realised the potential users of AD/Cycle are going to need help in adopting it. He believes that IBM is attempting to get most of its large systems user base into Systems Application Archi-tecture, and that the degree of acceptance will strongly influence subsequent SAA application announcements. Another effect of AD/Cycle is the depletion of the third party vendor market, and according to Clifford-Winters, most of the third party software suppliers will be locked into the IBM. He says that this places potential users in a quandary. Should they ignore the dangers of lock-in, or should they assert their independence and see competitors use IBM’s assistance to gain commercial advantage?
What’s next in the IBM mainframe hardware world and what should users do in the meantime?
What is the strategy behind IBM’s Systems Application Architecture, AD/Cycle and the Repository, and why is the whole trend IBM is forcing in its software initiatives bad news for those involved in IBM mainframe operations? Janice McGinn got some of the answers at the UK Computer Measurement Group conference in Glasgow last week. -That has still to be resolved, but Clifford-Winters reckons that IBM is staking all on SAA, and that it won’t be allowed to fail. He ties in the Repository with IBM’s Computer Integrated Manufacturing Repository, the CIM Communications and Data Facility Version 1, saying that the industrial solutions were important for two reasons. They represent the first vertical industry application suite for SAA, and they illustrate the eventual intention for Repository Manager/MVS for the application programming interfaces that will underlie SAA’s Common Programming Interfaces. He reckons that when IBM acquired Lockheed Corp’s subsidiary, Cadam Inc, the combinbation of its software with Copics and Maapics gave IBM an important lead in large scale Computer Integrated Manufacturing. Clifford-Winters believes that IBM will attempt to mirror this success in other vertical markets. He acknowleges that there are differences between CIM-CDF and Repository Manager/MVS, specifically that CIM-CDF offers both data and process definition, and execution facilities. However, he also claims that to separate build and run time is an artificial distinction since production applications developed using a computer aided software engineering tool can access the Repository at run time if required. Clifford-Winters believes that the next six m
onths will see more detail about Repository Manager/MVS and of course, AD/Cycle will be available. OfficeVision and CIM Architecture, the two SAA-compliant packages, have been announced and he expects more vertically oriented packages this year. But, he emphasises that this is only the start of the Repository, and as more roles become apparent, they will be significant to those working in Installation Management. Development of the operational support repository facility is lagging Repository Manager/MVS by about 18 months, but IBM has already said that NetView will be the control interface for automated operations.
Cede control
Clifford-Winters thinnks it likely that the operational repository will become the central point for operational activities. These include automated operations, system configuration, system generation, network design and generation, network control, and system managed storage. Hence his assertion that operations departments will cede control of mainframe operations. The operating system will disappear from view as will control of the location of data. Clifford-Winters finds an irony in this style of mainframe operating system and compares it with the System 38 and AS/400, which he claims were spawned by the Future Systems Project scrapped nearly two decades ago.
