A framework for monitoring and runtime recovery of web service-based applications

Abstract

Service provisioning is a challenging research area for the design and implementation of autonomic service-oriented software systems. It includes automated QoS management for such systems and their applications. Monitoring, Diagnosis and Repair are three key features of QoS management. This work presents a self-healing Web service-based framework that manages QoS degradation at runtime. Our approach is based on proxies. Proxies act on meta-level communications and extend the HTTP envelope of the exchanged messages with QoS-related parameter values. QoS Data are filtered over time and analysed using statistical functions and the Hidden Markov Model. Detected QoS degradations are handled with proxies. We experienced our framework using an orchestrated electronic shop application (FoodShop).