Research On Content Transmission And Caching Mechanism In Named Data Networking

With the rapid development of the Internet,multimedia traffic has become a ma?jor part of the existing network traffic.However,It is difficult for the traditional IP based network architecture which is designed for the end to end communications to meet this change.Many problems emerge in terms of transport efficiency,security,etc.Therefore,Named Data Networking(NDN)architecture has been proposed and received extensive attention.In NDN,each content is partitioned into multiple chunks and chunks are uniquely named,besides,NDN uses the chunk name as the index for routing and content matching so that NDN can realize the separation of content and its location.In addition,NDN implements in-network caching so that subsequent content requests can be directly serverd at the internal router,thereby reducing duplicate traffic in the network and decreasing the user's access delay.However,there are still a series of problems need to be solved in the NDN archi-tecture.The first one is the transmission problem caused by the content partition in the NDN.The content chunk of the existing NDN architecture is too small,which will cause the client to send a large amount of Interest packets to obtain a complete content.Besides,the NDN is a stateful routing architecture,which means the router needs to maintain the state of the Interest packets that passing through,and this will result in a waste of resources when dealing with multiple Interest packets.Increasing chunk size will solve this problem but still cause reliability issues due to packet loss probability increased and high delays due to hop-by-hop reassembly at the internal routers.The second is the caching redundancy problem.The cache space of the routers in the NDN is limited,hence,the default cache policy which leaves cache everywhere will cause re-dundancy problem.This not only wastes cache resources,but also results in low cache hit rate and high delay.Therefore,to solve the aforementioned problems,a low latency fragments reassembly scheme and a congestion avoidance caching mechanism are pro-posed.Specifically,the main contributions of this dissertation are as follows:(1)Aiming at the problem that the current NDN network content chunk is too small,we propose a low latency content chunk fragmentation and reassembly transmis-sion protocol which named FFRD.The key thoughts of FFRD is to realize the separa-tion of content reassembly and forwarding on the router side so as to avoid the delay accumulation due to router side reassembly.In addition,FFRD proposes a hop-by-hop reliable transmission mechanism so that the NDN can guarantee end-to-end reliability when transmitting large content chunks,thereby avoiding problems such as increasing time-delay due to end-to-end retransmissions.Finally,we implemented the proposed protocol FFRD and verified the performance through simulation.The results show that the FFRD can reduce the end-to-end delay effectively when transmitting large content chunks and provide better reliability.(2)Aiming at the cache performance problem existing in current NDN networks,we propose a Congestion Avoid Cache allocation and placement mechanism named CAC.Considering that if the content is cached at different nodes in the topology,the number of users they can serve is different,which resulting in different cache repetition rate.Therefore,this dissertation designs a cache allocation scheme based on the node location in the topylogy.In addition,by defining the potential value of the cache at this node and designing an "on-path" collaborative caching mechanism,the content can be cached at the downstream of the congested node,thereby reducing the number of subsequent requests routed to the congested node,and hence reducing the probability of the congestion loss.Finally,the simulation results verify that CAC has significant advantages in terms of cache repetition rate,latency,cache hit rate,and end-to-end retransmissions.