Conclusion

Abstract

If computing systems could be designed and built so as to be free from faults throughout their operational life, then this book would be redundant. Or would it? Constructing a perfect system, even if this could be achieved in practice, is not necessarily the most cost-effective approach for producing high reliability. The law of diminishing returns (and its quantitative variant, the 80:20 rule) suggests that optimal solutions are rarely obtained by placing all one’s eggs in one basket. A well-engineered approach to building highly dependable systems is likely to be based on striving to attain perfection, but acknowledges that imperfections will remain and that fault tolerance will be needed to cope with them.