Enhancing Availability through Dynamic Monitoring and Management in a Self-Adaptive SOA Platform

Abstract—The availability of Service-oriented Architectures (SOA) depends on two factors. These are firstly the availability of the services that provide a certain business functionality and, secondly, the availability of the components or services that make up the underlying SOA platform. For platforms that are supposed to form the core of mission-critical service-oriented applications, this implicates the need for mechanisms that can regulate the availability levels of the core services in changing conditions. In this paper, we handle open issues about the kind of monitoring functionalities and adaptation mechanisms that should be integrated in SOA infrastructures. In our proposed solution, we integrate concepts of event-based systems to enhance the dynamicity of the SOA platform monitoring, as well as concepts from peer-to-peer computing to achieve an efficient distribution of the SOA platform core. By prototypically implementing the concepts as extensions of Apache Tuscany, which is a realization of the Service Component Architecture standard, we show in an experiment-based evaluation how the availability of the core services of SOA infrastructures has been improved. Additionally, we explain further benefits that can be achieved with adaptation mechanisms other than replication, which are also enabled by our extensions