Trends in computer system structure and architecture
Management of Complexity is the overriding concern in the development of modern Data Processing Systems. We observe a split into systems which are used in a computing center, and systems which are used by an individual department. The first class of systems is characterized by general purpose attributes of its structure and features for adaption to a wide spectrum of individual users. The second class is characterized by a tayloring of hardware and software features to individual applications.
Decentralized departmental systems can achieve significant efficiency improvements through application tayloring, if and when their applications can be treated as isolated from each other. Where this is not possible we observe a development trend towards computing networks with distributed intelligence, very often with a powerful central host system. In this case, issues of architecture integrity, compatibility, program and data portability are a major concern, and will impact future developments to a significant extent.
KeywordsMini System Data Processing System Main Store Data Base System Instruction Execution
- M.V. Wilkes, "The best way to design an automatic calculating machine", presented at the Manchester University Comp. Inaugural Conference, Manchester, England, 1951, p. 16Google Scholar
- C. Schuenemann, "Micro-and Picoprogram Stores", Proceedings of the IBM Informatik Symposium on Rechnerstrukturen, R. Oldenbourg, 1974, p. 36–74Google Scholar
- E.I. Organick, "The Multics System", MIT Press, 1972Google Scholar