Markov Model of Web Services for Their Performance Based on Phase-Type Expansion
Yanjie Liu, Zheng Zheng, Jiantao Zhang · 2019
Developers usually combine various existing atomic services into composite services that satisfy complex needs. There are mainly four basic structures in composite Web service which include sequence, branch, parallelism and loop. Besides functional issues, non-functional issues such as performance play an important role in evaluating services. Therefore, before the service deployment, it is necessary to check whether performance will meet its service-level agreement (SLA). In this paper, we use a novel approach to model an expanded homogeneous Continuous Time Markov Chain (HCTMC) of basic Web services structures by Phase-Type (PH) fitting. It is based on observations of the execution times of atomic service. Analysis of experiments results on real Web services verifies the validity of our method.