A Service Identification Framework for Legacy System Migration into SOA

Abstract

Service-Oriented Architecture (SOA) enables the re-engineering and migration of legacy software systems into loosely-coupled and interoperable sets of services. Existing approaches focus mainly on defining coarse-grained services corresponding to the key business requirements. An improved migration of legacy systems onto SOA-based systems requires identifying the `optimal' services with an appropriate level of granularity. This paper proposes a novel framework which identifies the key services effectively. The framework approach focuses on defining these services based on a Model-Driven Architecture approach supported by a SOA meta-model. Effective guidelines are proposed for identifying the optimal service granularity over a wide range of possible service types.