Abstract
We have developed measures which reflect the interaction between the reliability and the performance characteristics of computing systems. These measures can be used to evaluate traditional computer architectures, such as uniprocessors and standby redundant systems; gracefully degrading systems, such as multiprocessors, which can react to a detected failure by reconfiguring to a state with a decreased level of performance; and distributed systems. This analysis method, which provides quantitative information about the tradeoffs between reliability and performance, is demonstrated in several examples.

This publication has 9 references indexed in Scilit: