Research On Forwarding Strategy For Named Data Networking

With the rapid development of network applications,people's demand for Internet has gradually changed from end-to-end connection to content acquisition.The current TCP/IP architecture adopted by the Internet can no longer meet this demand change,so researchers have proposed a new information-centric network architecture to meet the future network development needs by taking information as the core of the network.Among them,Named Data Networking(NDN)represents content by name,which enables users to obtain content by name without concern for content sources.It has achieved great success and has become a typical representative of Information Center network.However,the effect of NDN relying only on Forwarding interest tables is not good,so this paper studies the forwarding strategy of NDN.The main work of this paper is as follows:1.In view of the low efficiency of NDN data transmission,this paper proposes a forwarding strategy based on recommendation algorithm.According to the hierarchical naming rules of NDN,this strategy mines the similarity of interest packages and nodes forwarding the same interest packages,and proposes a forwarding algorithm based on package recommendation and a forwarding algorithm based on node recommendation.The forwarding efficiency is improved by recommending the most similar forwarding situation of interest packages and the closest node to the current node to the received unknown interest packages.2.In view of the shortage of quality of service in NDN support,this paper proposes a quality of service(QOS)perceived forwarding strategy based on Markov Decision Processes(MDP).According to the mechanism of NDN data transmission,this strategy models the process of forwarding interest packages by routing nodes in MDP,uses network overhead as system revenue,and achieves QoS-aware forwarding by maximizing revenue.3.In this paper,the above two forwarding strategies are validated on ndnSIM 2.5 platform,and several current mainstream and typical forwarding strategies are selected as comparative experiments to evaluate the performance of the proposed forwarding strategies.The simulation results show that the forwarding strategy based on recommendation algorithm has higher interest satisfaction rate and cache hit rate than the four forwarding strategies such as SAF,and the QoS-aware forwarding strategy based on MDP has shorter delay and higher throughput than the two forwarding strategies such as MULticast.