Research On Issues Of Delay Optimization For Information-centric Networking

With the increasing demand for Internet users, the speed of network traffic grows faster than the performance improvement of the hardware.Meanwhile, users are inclined to share contents among them, which brings about lots of repeated traffic on the Internet. TCP/IP shows more and more shortcomings in the mobility, scalability and security. As a promising network architecture, Information-Centric Networking utilize caching on the Internet and can solve these problems effectively.In this paper, we focus on delay optimization in Information-Centric Networking and propose a new caching strategy: Socially-aware NodeRank-based Caching Strategy for Information-Centric Networking(SNCS). SNCS is a caching strategy based on the importance of nodes to community users. We place cache copy in the proper position to improve cache hit ratio and reduce the content acquisition delay. The current network gradually becomes a social-oriented network and the social link has a direct influence on the content distribution. In order to select proper caching nodes, we include the social attributes into the network traffic model.SNCS considers physical and social attributes and divides the network into social Autonomous Sections (sociAS) by the community detection algorithm. In each sociAS, we find out the key nodes with great influence. In order to measure the importance of nodes, this paper is based on the idea "if neighbor node is with great influence, node itself will also be important" and it emphasizes the value of nodes in the whole network. This paper proposes a new algorithm for ranking nodes. The algorithm constructs a matrix based on the physical importance of neighbor nodes and the social attributes of node itself. By calculating the matrix, we get the NodeRank eigenvector. Each element in the eigenvector represents the score of nodes. The node with highest score is selected as the key node. The key node will push popular contents to neighbor nodes. Meanwhile, cache copy will be placed on the key node in each sociAS off-path and the nodes with higher scores in-path. So the distribution of cache copy is reasonable in terms of spatial distribution.In order to realize the communication process of SNCS, this paper designs new packets, the process of request and response. And we simulate the SNCS on the SocialCCNSim platform. The simulation results demonstrate our scheme achieves better performance, including:cache hit ratio, cache hit distance, delay of download and server load saving. Different from other caching schemes, SNCS includes social characteristics and designs a new algorithm for node ranking, which makes caching strategy socially-aware and content-aware and reduces content acquisition delay.