An Adaptive Service Composition Performance Guarantee Mechanism in Peer-to-Peer Network

Service composition is becoming a hot research point in recent years, especially in distributed peer-to-peer (P2P) network. However, how to guarantee the high-quality composition service performance becomes the key technology and difficulty in P2P network, which draws lots of attention in computer research field. According to the characteristics of distributed network, an adaptive service composition performance guarantee mechanism in P2P network is presented. Firstly, the composition services are divided into preferred service and backup one; secondly, BP neural network is introduced to detect and evaluate the service performance exactly; finally, the backup service will adaptively replace the preferred one when the preferred composition service performance is less than the threshold value. Except that, the half-time window punishment mechanism is introduced to increase the punishment to the low-quality service and encourage users to provide more high-quality one, which further improves the service composition performance in P2P network. The simulations verify our method to be effective and enrich the proposed theory.