A Metrics Framework for Evaluating SOA Service Granularity

Service-Oriented Architecture (SOA) is intended to improve software interoperability by exposing dynamic applications as services. To evaluate the design of services in service-based systems, quality measurements are essential to decide tradeoffs between SOA quality attributes. Current SOA quality metrics pay little attention to service granularity as an important key design feature that impacts other internal SOA quality attributes. In this paper we introduce the structural attribute of service granularity for the analysis of other internal structural software attributes: complexity, cohesion and coupling. Consequently, metrics are proposed for measuring SOA internal attributes using syntax code. These metrics will assist in development of optimal service design by considering appropriate trade-offs. An example case study is included to demonstrate proposed metrics.