An Approach for QoS-aware Selection of Shared Services for Multiple Service Choreographies

Abstract

Choreographies represent a particular approach for service composition in which coordination is performed in a distributed way. This approach is becoming one of the de facto standards for service composition in the Future Internet. However, QoS-aware deployment and management of service choreographies is usually performed without explicitly considering service sharing. This simplifying assumption makes choreography deployment less feasible in real scenarios, in which individual services are actually shared by several choreographies. This paper presents an approach for QoS-aware service selection for multiple choreographies considering the sharing of services among them. In particular, we argue that sharing-aware service selection represents a more realistic perspective since unsupervised service sharing may degrade the overall QoS provided for the choreographies. Our approach for service selection avoids this degradation by taking into account, besides the QoS requirements of service consumers and the QoS properties of services, the aggregate service load that results from the sharing.