Data Engineering and Information Systems
This article discusses how the data requirements of information systems influence the structure of the distributed computer systems upon which they are often built.
Abstract
T he c ter has made it possible to me anize much of the infor-matio terchange and processing that §stitute the nervous system of ouro ty. The computerized sys-temsprovide this mechanization, co only called information sys-s, will be responsible for computer technology's long-term impact on society. We need to understand how the information requirements of the different segments of our society should influence the design and architecture of the information systems serving those segments. I We have for some time been in a distributed computer era. This article discusses how the data requirements of information systems influence the structure of the distributed computer systems upon which they are often built. Many of the concepts, carried over into centralized systems, enable those systems to evolve in a distributed world. Unfortunately, our current understanding of computer science does not iwnclude precise definitions of, or per-iMsharp boundaries among, data, in-for on, and knowledge.2 In this articleXorientation is on data; concepts toipg on information and knowledge are introduced only as necessary. Consequently, the reader will find situations in which the term "data" is used in a context where it might be more appropriate to use the terms "information" or "knowledge." Our two principal premises are 1. that a specific information system may involve logically and physically distributed computers and databases , and 2. that the individual functional components of an information system are driven by information from companion components and in turn provide information to other components through the transfer of data. The term "component" here includes the interfaces as well as the internal parts of the systems, namely humans, displays, sensors, actuators, computers , and storage devices. Both of these premises are likely self-evident. However, not so self-evident is that this interchange of data is far more than a data communication problem. For example, we can describe the services provided by the post office from a pure data communication viewpoint. In the post office system, an envelope is delivered to the ad