Software Engineering for Fault Tolerant Systems

摘要

The objectives of this study are to (1) assess the current state of the art offault tolerant software schemes, (2) evaluate the status of various software engineering issues in this context, (3) identify critical gaps in the currently available technology and, (4) provide recommendations for research and development efforts to enhance the technological base of fault tolerant software engineering. Towards these objectives, we have discussed several software fault tolerance schemes, studied the available experimental and analytical evidence about their usefulness and assessed the current status of fault tolerant software engineering for sequential and parallel computers. Based on the studies reported here, we feel that the current state-of-the-art of fault tolerant software is mature enough to tolerate design faults in specific circumstances with appropriate provisions of redundancy and allied supporting mechanisms. However, no known fault tolerance technique can guarantee failure-free system operation. Further, it is questionable whether the current approaches are cost-effective in achieving the desired gain in operational software reliability. We feel that what is needed is a systematic, cost effective approach to software development which explicitly addresses the fault tolerance issues throughout the development life-cycle.