Client profiling for QoS-based Web service recommendation

Abstract

Quality of service (QoS) is one of the important factors when choosing a Web service (WS) provided by a particular WS provider. In this paper we focus on performance, which is one of the important QoS attributes. Performance can be evaluated from the client-side as well as the server-side. As clients are connected through heterogeneous network environments, different clients can experience different performance although they are connected to the same WS provider. Through client grouping and profiling, performance experienced by prospective clients can be estimated based on other clients' historical performance. In this paper we present the common factors for client grouping and profiling, and present the results of the experiments we carried out to understand the effect of these factors on the performance experienced by WS clients. The experimental results show the importance of client grouping and profiling for WS recommendation and composition.