Research Of Service Migration Strategy For Service-Orientated Future Internet

Considering the features of Internet over which web services are widely distributed, internet expansion makes the Quality of Service (QoS) a challenge. When the network traffic is congested, it is costly and is difficult to ensure the QoS for the service which is popular and far from the source node at the same time. In this case, service migration is always a good choice to solve this issue. The research on service migration strategy has been already done in the field of cloud computing, CDN and mobile agent. So it’s very important to design a service migration strategy that reduce the communication cost, make internet scalable and improve the QoS in the complicated environment of future network.There are some common invisible problems in the discussion of previous research, that is, the strategy is passive, lack of local feature and costly. This paper proposes a new strategy in which non-source node initiates a migration request actively and provides local network information for the source node in order to make the selection decision goal-oriented and intelligent. We design a basic infrastructure and a new cost model to support our strategy and use shortest path tree (SPT) decide the final placement. The simulation results show that our approach gets satisfactory performance gains in the terms of communication cost and delay.The paper consists of four parts.Firstly, the paper briefly introduces the background of future network and the research of service migration strategy.Secondly, the paper discusses the structure and migration strategy on traditional migration research.Thirdly, the paper presents the issues of previous research and proposes a new service migration strategy based on local information and initiative request. Then the cost model and interaction way between source and non-source node are designed and the strategy based on SPT is presented.Finally, the simulation of service migration strategy proposed in this paper is designed by using NS2and the simulation results are analyzed.