Research On Cooperative Caching Strategy Oriented To Information-centric Networks

With the increase of network data traffic,the traditional network architecture for the host centered is faced with a series of challenges.On one hand,the overload of network traffic can easily lead to large time delay in the process of accessing network;on the other hand,there is a lot of redundant data and only a small part can be reused.In order to meet the massive and heterogeneous information access,the Informationcentric Networks(ICN)uses network communication mode centered on information to replace the traditional network communication mode centered on host.The ubiquitous and transparent network cache can store content forwarded,thus it can improve content access and transmission efficiency and the utilization of network resources.In ICN,the good or bad of a caching policy is a direct impact on the transmission performance of network.The existing caching schemes have problems of high cache redundancy,low cache utilization ratio and low content diversity ratio.Thus,a popularity and gain based caching scheme(PGBCS)is proposed based on fine-grained cache management,it promote the cache performance to an extent.The main research content and innovation place of this thesis is as follows:1.Based on fine-grained cache management,this thesis proposes a placement method which finding the best place to cache by comparing the comprehensive gain of nodes.First,the object of study of existing caching strategy is mostly content files,this does not meet the characteristics of fine-grained cache of ICN,so this thesis refines the study object to the chunk which realizes fine-grained cache requirements;Secondly,studying the popularity based on the content chunk,in this thesis we set a value index for each content chunk,the greater the popularity of the content chunk the larger its value;Finally,comprehensively measuring several factors that affect the caching gain when looking for the best node to cache content chunk,thus can maximize the whole caching gain of nodes.2.This thesis proposes a cache replacement method based the comparison of content chunk value.The value of the content chunk in cache space is changing dynamically as the user's access frequency,in order to ensure that the high popularity content chunks have high caching value,the value of content chunks increases exponentially with the requested counts,and decays exponentially with time.The identification of content chunk is marked as “replaceable” when the content chunkvalue is below the threshold value set,and the occupied space is regarded as free.When the cache space of the node is full,the new arrival content chunk will replace the content chunk with the lowest value in the node by comparing their value.In this thesis,first of all,the study of ICN and network cache is reviewed,then on this basis,the PGBCS caching mechanism is introduced in datail.To verify the advantages of PGBCS performance,four representative caching schemes are selected in this thesis,they are LCE,RCOne,Prob Cache and Betw.Making the comparison on aspects of caching hit ratio,the hop that user access to content and content diversity ratio.The simulation results show that compared with the above four kinds of caching scheme,PGBCS can have a larger increase of cache performance,when introducing additional small flow.