QoS-Driven Service Matching Algorithm Based on User Requirements

摘要

Quality of Service (QoS) is an important factor which should be considered in service matching. There are two problems in most existing solutions. Firstly, most QoS models are static model described by determinate values or probability distributions, ignoring the impact of time factor. However, most QoS attributes are time-dependent, such as response time and reliability. Secondly, the service selection criteria of most QoS-driven service matching algorithms are based on service performance, but user requirements and the load of services are not considered. In this paper, we propose a Time-Segmented QoS Model (TSQM) to dynamically model QoS. Based on this model, a Service Matching algorithm based user QoS request and Priority (QPSM) is proposed. The priority of user requests is used to control the load of the services. Simulation results show that the algorithm can achieve a higher response rate and a better effect of load balancing.