Reconciliation And Analysis Of Service Composition

In open, dynamic, and uncontrollable environments, software as a service has become a rising innovative mode of software applications with the development of service computing and cloud computing in the 21st century. As a mainstream technology system to achieve software services, Web Services provide a general mechanism for business process integration. According to some certain application logic and combination strategies, services composition technology combines existing, relatively simple Web Services to provide value-added services which can meet the meet the increasingly complex business requirements of users. However, actual execution of a process instance often deviate from its original process definition. To ensure that the instance is able to provide the expected service, how to measure services composition conformance in a suitable way has become an important research content.On the one hand, the existing methods only focus on the control flow of services. They are not sensitive to the data flow deviations and can not identify the decrease of conformance degree which caused by the change of data flow, composition and analyze and measure the consistency focusing on the control flow. On the other hand, the conformance metrics of these approaches are too strict which overemphasize sequence orders between activities, causing the conformance degree values is often lower than the user’s experience. To address these problems, this paper proposes a new metric and related analysis methods to measure conformance. The main work is as follows:1. Based on the essential relations between activities, we proposed a relaxed notion of trace equivalence. By quantifying this notion of trace equivalence, we can further define the consistency degree between two aligned traces.2. To measure the conformance of services composition, we need to find a reference trace "best fitting" to the sequence orders of the instance such that the conformance checking problem can be turned into the problem of trace consistency analysis. Since it can be proved that finding the reference trace is NP-hard, we proposed heuristics based on process decomposition and trace replaying to find the reference trace.3. In order to verify the accuracy and effectiveness of the proposed method, we designed and implemented a proof-of-concept tool ProM-C to measure and analysis conformance. This tool could not only return a quantitative conformance degree of services composition, but also provide diagnostic information for the deviations of the process instance from the process definition, which facilitates process debugging, refactoring, and adaptation.4. We proved that our approach can effectively identify both control flow and data flow deviations and can avoid false negatives. We verified the accuracy and validity of our approach to search the reference trace by experiments using real-life WS-BPEL data set.