Research On User Preference Based Routing And Caching Strategy In Content Centric Networking

With the development of communication network,content distribution and sharing will be a main application in the future networks,meanwhile,there are large numbers of terminals with Internet access available for each user.The current IP network can not handle the complexity of addressing and the limitation of capacity.Content centric networking focuses on the content itself instead of the content location,which can solve the problem that each content delivery of the end-to-end communication is indirectly mapped to the device where the content is located.CCN has potential advantages in terms of naming,security,routing,caching and transmission.Among them,routing and caching strategies are two key research directions in CCN.Designing and optimizing routing and caching strategies can effectively improve content retrieval efficiency and reduce network traffic load.The paper is based on the National Natural Science Foundation "The Research of Caching for Content Centric Mobile Social Networking(61502046)" and the Open Foundation of Beijing Key Laboratory of Network System Architecture and Convergence "The Research of Caching for Content Centric Wireless Networks based on User Preference".In this paper,we design a routing strategy and a caching strategy based on user preference for social-oriented content distribution requirements.The main work of this paper are as follows:(1)This paper outlines the current CCN architecture,node model,key technologies and related simulation platform.Then it introduces the routing and caching technologies of CCN in detail.Finally,the current research status of routing and caching strategies in CCN is summarized.(2)In order to make full use of the cached contents on the intermediate routing nodes and reduce control overhead,a user preference similarity based routing strategy is proposed.We first define the user preference similarity of directed link.Then it is integrated with link delay and link bandwidth to calculate the link loss.Finally,the forwarding loss of each face is calculated and a different best next-hop is produced for each content type.Simulation shows that the strategy has a better overall performance in terms of producer load,cache replacement rate,average retrieval delay and retrieval success rate after comparison.Also,reducing the cache redundancy can effectively improve the performance of cache replacement rate and average retrieval delay.(3)To reduce the statistical overhead of popularity and consider the timeliness of user preference,a content popularity prediction based caching strategy is proposed.The contents are divided into different content types using cluster analysis.Then,count the popularity of each content type periodically and predict the future popularity values.Based on prediction result,each node makes its own caching placement decision.Simulation shows that the strategy has a better performance in terms of cache replacement rate,cache hit rate and average hop count after comparison and the popularity prediction can effectively improve the performance of the latter two metrics.Meanwhile,simulation validated the performance gain of UPSR and CPPC proposed in this paper for routing and caching collaboration.