Markov Decision Process Model for Path Selection Algorithm on Multi-Business System with Services Composition

Abstract

Services composition can reorganize a variety of heterogeneous network services to form new businesses on the multi-service streaming media system, which can speed up the development and deployment of new services. In this paper, a three layer model is used to describe the service composition in a streaming media system. In order to improve the real-time request response performance, path selection is divided into two steps, the initial paths planning and dynamic path selection. The dynamic path selection behavior is described by using Markov decision processes model. When QoS is guarded, the policy iteration algorithm is adopted to improve the system of the total receipts and to reduce the refuse rate. The simulation results show that the proposed method of service composition improves the performance of the streaming media system.