A flow analysis approach for service-oriented architectures

Abstract

The discipline of SOA (Service-oriented Architecture) provides concepts for designing the structural and behavioral aspects of application landscapes that rely on the interaction of self-contained services. To assess an architecture’s quality and validate its conformance to behavioral requirements, those models must be subjected to sophisticated static analyses. We propose a comprehensive methodology that relies on data flow analysis for a context-sensitive evaluation of serviceoriented system designs. The approach employs a model-based format for SOA artifacts which acts as a uniform basis for the specification and execution of various analyses. Using this methodology, we implement two analyses which reveal blocking calls and assess performance metrics. These applications are evaluated in the context of two case studies that have been developed in the SENSORIA and the ENVIROFI projects.