Performance Analysis of Service Composition Based on Fuzzy Differential Equations

In this paper, a new method is proposed to measure the performance of service composition. Service composition described with Business Process Execution Language (BPEL) is modeled by a group of fuzzy differential equations, where each equation describes the state change of the service composition. Each service state is measured by a time-dependent fuzzy number that indicates the extent to which the state can be reached in execution. This information of measure can help us to conduct performance analysis, such as estimating response time, throughput, efficiency, and maximum load. This method has the following advantages: 1) It can handle fuzziness such as nondeterministic response time; 2) it treats the system as a box and displays a global picture of execution states to the users so that users know exactly where to improve the performance; 3) it can entirely avoid the state explosion problem, which may be hit by the existing performance analysis methods. A case study demonstrates the advantages of our method.