Defining a Self-Healing QoS-based Infrastructure for Web Services Applications

Web services (WS) technology proposes tools to implement complex distributed systems. It allows creating networks of heterogeneous and cooperative WS on applications. Due to the complexities related to the dynamics of this kind of system (new WS could come in and come out of the system), applications are exposed to several failure points in both components and connections. Current efforts in WS literature offer a supporting infrastructure for this problem; most of them are oriented to repairing, mainly on application's workflow level. In this paper, we expose our ideas to define an infrastructure based on quality of service (QoS) analysis of WS-based applications. Our approach considers a complementary strategy, supported by the adaptability of the application's architecture. This strategy is planned to prevent and respond to QoS degradation in WS-based applications, trying to enhance their performance. Our solution is oriented to discover new approaches for the three main stages, identified in the related literature for self-healing infrastructure: monitoring, diagnosis and recovering. In order to show our strategy effectiveness, we are going to implement it in the access to a Digital Library based on WS.