Data-driven routing service composition via requirement chain

The current Internet should not only support simple packet forwarding, but also be able to provide diverse network functions to compose different routing services on the communication paths for network applications. Meanwhile, the user requirements for different applications are becoming more and more diversified, it is a challenge to distinguish the differences among these requirements in order to select appropriate functions for routing service composition. In this paper, we propose a dynamic routing service composition model to analyze the user communication requirements for different types of applications. A user requirement feature model and a complete and orderly requirement chain acquiring mechanism are devised to guide function selection and service composition. In addition, with the benefits of both user and Internet Service Provider (ISP) jointly considered, a method for achieving benefit equilibrium of user and ISP is presented. The simulation results and performance evaluation show that the proposed mechanism is feasible and effective.