An Analytic Approach for Resiliency Quantification of Systems
Resiliency is becoming an important service primitive for large scale distributed systems and networks. A key problem in quantifying the resiliency of computer systems is the lack of consensus on its definition and a systematic approach for performing resiliency analysis. Traditionally, resiliency has been interpreted as the fault-tolerant capability of systems. Laprie and Simoncini defined resiliency as the persistence of service delivery that can justifiably be trusted, when facing changes. Our definition of resiliency includes this notion of change(s). In this paper, we analytically quantify the resiliency of a system in terms of effects of such changes. We outline the general approach for systematic resiliency quantification. We quantify the resiliency of performance and availability measures using stochastic analytic models through an example of a telephone switching system.
Availability Continuous time Markov chain Performance Resiliency
KISHOR S. TRJVEDI RAHUL GHOSH DONG SEONG KIM
Dept. of Electrical and Computer Eng., Duke University, NC 27708, USA
国际会议
北京
英文
882-888
2011-06-20（万方平台首次上网日期，不代表论文的发表时间）