Statistic-Based Dynamic Complexity Measurement for Web Service System

Abstract

The existing research mainly concerns on the static complexity measurement for service-based system, but the dynamic features like execution behavior have been ignored. In this paper, we proposed a hierarchical measurement framework for evaluating the complexity of Web services from the dynamic aspect. At the level of single service, fluctuation rate is used to represent the QoS (Quality of Service) change during service invocation. Then, a cumulative distribution function is used to measure the dynamic complexity of service performance. At the system level, execution vectors and the corresponding probabilities can be counted according to the trace set of system dynamic executions. Subsequently, the complexity of dynamic execution behaviors can be calculated by the usage of entropy value. In addition, the rationality of above metrics has been validated by the studies on two real applications.