Nowadays it is hard to think of any applications in modern society in which electronic systems do not play a significant role. In aerospace and aviation, defence, telecommunication and healthcare, to name a few, computers have already assumed the most life-critical tasks. Unlike most human beings, computers seem to do their job pretty well, at most times and under all environmental conditions. Sometimes, however, for some reason, the fresh water supply in a city stops, the mainframe computer of an international security exchange centre malfunctions, or the satellite television goes out abruptly. Possible sources for such dissatisfactory performances are physical deterioration or design faults in hardware components. Fortunately in the 1950s and 1960s quite a number of reliability models were developed for hardware. Another major source for malfunctioning of computer systems is the presence of bugs in the software that controls the system. The modelling of software reliability was only begun in the early 1970s. This chapter presents a comprehensive approach to the development of a reliable PACS, which are capable of meeting the high-quality level required of mission-critical medical devices. To develop a preliminary design, the PACS team would begin with a system description and reliability evaluation of a baseline system, includ ing implementation of hardware redundancy, software provisions, and acceptance test. Through detailed system analyses and electrical, electronic and mechanical reliability studies, a final preliminary design can be derived. In this chapter the essential mathematical and statistical aspects of hardware and software reliability predictions are first presented, followed by a spreadsheet-based approach to model hardware and software reliabilities. A method of designing higher system reliability through parallel and cross-linked configurations is then given. Finally a brief case on the acceptance test of a PACS software is illustrated.
Automatically generated acceptance test: a software reliability experiment