Research On Cache Strategy For Content-centric Network

The traditional transmission protocol based on IP address incurs many duplicate content transmissions.To solve this problem,Content-Centric Network(CCN)is proposed.Its transmission is based on the name of content and routers cache the contents to reduce redundant transmissions,improving transmission efficiency.However,the cache capacity of the network is too small to contain all contents,so caching everything will result in severe cache redundancy and frequent cache replacement.How to efficiently use the cache resources is a fundamental problem in Content-Centric Network.In this thesis,we study in-networking cache decision strategies in CCN,considering cache system architecture,heterogeneous user requests and wireless access.The main contribution is as follows:1.We study the cache decision problem with heterogeneous of user requests.We propose a cache decision system framework based on SDCCN and formulate the cache decision problem with the objective of minimizing the total delay of users' requests.We prove that the problem is extremely hard unless P=NP.Then,we give an integer linear programming formulation(CDP-ILP)to find an optimal solution.For less running time,we propose three heuristic algorithms,AlgRelax,AlgGreedy,and AlgMix.Additionally,we design a simulation platform for in-networking cache system for Content-Centric Network,and implement various caching strategies.We compare the algorithms in terms of total delay of users' requests,cache hit counts and server hit counts.Simulation results show that the proposed solution can reduce the total delay,increase the cache hit counts and decrease the server hit counts,compared to existing cache decision strategies.2.We further consider the case when users can choose one router to associate.We consider user access and cache decision jointly to minimize the total delay of users' requests,provided that cache capacity at routers are satisfied.We establish a quadratic programming model for this case and propose an algorithm based on the similarity of requests,called AlgUACC.Finally,we extend the in-networking cache system platform,add user access module and implement AlgUACC.Simulation results show that our algorithm can reduce the request delay compared to RAND-Prob.