A Research On Cache Mechanism Of Future Network

Use of Internet in today's world has evolved to be dominated by content distribution and retrieval,while currently used IP architecture can not meet this demand.To address this challenge,a variety of future network architectures emerge as the times require.They break the traditional host-to-host communication mode and put content in the core position.On this basis,they can distribute the content rapidly.As one of the future network architecture,CCN is the hotspot in research field.It decouples content from host and caches the content in every network node.Any node that contains the requested content can act as a server to response clients.Thus,CCN can deliver content efficiently.These advantages of CCN come from the fact that it supports in-network caching.It can be concluded that caching is the cornerstone of CCN.Accordingly,the performance of the caching strategies directly affects the performance of CCN.In this thesis,CCN is selected as the basic architecture to study the caching mechanism of future network.Firstly,several typical future network architectures are described and the architecture of CCN is analyzed emphatically.Then,the caching strategies in CCN are classified and their existing problems are pointed out.To solve these problems,a popularity-based collaboration caching strategy is proposed.It's core idea is dividing the network into several autonomous systems.Nodes in the autonomous system cooperate with each other and the cache is centrally controlled through the control node.It can be found that this strategy can improve the cache hit ratio and accelerate content retrieve speed through the simulation experiments.Finally,a popularity-based cache size allocation policy is proposed,it introduces popularity into the cache replacement strategy and proportionally allocates the cache space based on the popularity.The simulation results show that compared to the traditional caching replacement strategies,this strategy has a better performance in cache hit ratio,network traffic and energy consumption.