Research On Content Caching Strategy In Information-centric Network Virtualization

Today,Internet has become an important infrastructure for social communication,affects people's work,living and learning profoundly,and advances the country's economic development and social progress continuously.However,with the explosive growth of network traffic,more and more users are concerned about content itself instead of the location of content.The existed TCP/IP-based network architecture in scalability,security,mobility and green and energy conservation has appeared constantly shortcomings.In order to solve these problems fundamentally,researchers have proposed information-centric network(ICN)architecture that content as center,and one of the most important features of ICN is in-network caching,which allow node to cache content that can service the future request.Meanwhile,in traditional network,the physical resources of different operators cannot be shared,which will increase capital expenses and operation expenses for content delivery.But using network virtualization(NV)technology can not only achieve shared infrastructure resources,but also shared content resources,among different service providers(SPs).In order to realize the cached content can be shared in different SPs and improve the end-to-end network performance,researchers propose an architecture of information-centric network virtualization(ICNV).In ICNV,we can use in-network caching technology to cahce content in node which can service the users and improve resources utilization,reduce the expenses of content delivery and network transmission redundancy and data center service load,etc.Therefore,studying caching strategy also become a hot topic in ICNV.This thesis mainly aims at studying the caching strategy in ICNV,the specific researching work is as follows:1.Because there are two actions for node to content: caching packet and dropping packet,we define the number of contents cached in node as caching state,and regard the transition of caching state as a Markov chain,then we built transition matrix of caching state with the transition of caching state.Through considering the response probability of node,we build average response hops as an objective function.Since the objective function is a centralized function,a distributed algorithm based on potential game is proposed in this paper.In the game,the node would choose caching strategy according to the situation of node itself,and the optimal caching strategy is obtained when node updates its caching strategy until reach Nash Equilibrium.2.Considering the network energy consumption,and taking network energy conservation and environmental protection as starting point and using energy consumption as the basic model,we proposed an energy-efficiency optimization with caching strategy based on Access-Frequency and Hops which considering the access frequency and the caching hops of content.Through the caching strategy,as the energy efficiency of the network is optimized,the users' experience in network is also improved.