Dependability Concepts and Taxonomy

Summary

Reliable computing and fault tolerance, or, to use a more current term, dependable computing, is a longstanding area of research and practical implementation. This chapter discusses the recent challenges and opportunities in the reliability domain. Four important metrics measure system dependability: reliability, maintainability, availability, and safety. The chapter provides a brief, somewhat generic discussion of broad classes of faults/errors. The classes of software faults identified using Orthogonal Defect Classification were used as a basis for emulation of software faults using fault injection. As technology matures, the user set changes, the degree of product sophistication increases, and new sources of failures become prominent. One of the main trends driving the design of modern dependable systems has been the change in failure rates as well as the change in the dominant sources of failures. The chapter also presents an overview of the key concepts discussed in the subsequent chapters of this book.