NDN Mobility Optimization Method Based On Double Lead Search And Traffic Prediction

The internet is developing rapidly today.While the IP network is encountering various bottlenecks and challenges,some new networks are proposed.Among them,Named Data Networking(NDN),an implementation architecture of information center network(ICN)is a optimistic network solution.The distribution of data in NDN is based on the content name rather than the IP address,and NDN has the feature of taking content as the core.Nowadays,real-time multimedia mobile communication is becoming more and more common,and the amount of content is increasing.The content-centric nature of NDN makes it have the advantages of rapid content distribution and high resource utilization.As an emerging network,NDN still has some problems that need to be solved,including the mobility support of content producers.That is,when the communication occurs and the producer moves,the interest packet sent by the consumer cannot reach the producer's location correctly according to the routing table.The method of the native NDN is to rely on the Named-data Link State Routing protocol(NLSR)to globally update the routing table,and then resume communication.Because the global update of the routing table is time consuming,the communication in this mode is too high in time delay and inefficient in content distribution.In order to solve these problems,this paper proposes a mobility optimization scheme based on double-lead search and traffic prediction,which is dedicated to efficient and low-cost content transmission in the case of delay-sensitive conditions.Its main contributions are as follows:(1)Aiming at the problem of producer mobility in NDN,a double-lead search-based producer mobility support mechanism is proposed.That is,when the producer moves to the original access routing node's neighbor range,the batch forwards interest packets to actively search for the producer;when the producer moves to the original access routing node's neighbor range,the producer's current time is obtained in time through the notification of the proxy node position.Through the classification of the producer's movement and the different processing in different movement situations,the goal of quickly locating the producer and minimizing the traffic overhead is achieved.This mechanism retains the original characteristics of the NDN.After comparing with the producer mobility solutions(such as the solution based on the domain name resolution server and the solution based on the indirect node)in other related studies,various indicators such as delay and traffic can be improved.(2)There are two types of situations in the NDN two-way search mechanism.There is room for optimization:when the producer moves outside the range of the original access routing node's neighbors,the process of the original access routing node waiting for the proxy notification will cause additional delay;When the original access routing node is within the neighbor range,the interest packets at the original access routing node will be forwarded in batches to all nodes within its neighbor range,which will cause a certain amount of traffic overhead.Based on the above problems,this article further proposes an improvement scheme based on traffic prediction based on the first research results.The original access routing node where the producer is located will periodically predict the location to which the producer will switch,and modify the forwarding information table of the local routing node.After that,the interest packet of the requested content can be forwarded directly to the predicted location.This method not only removes the lag caused by waiting for proxy notifications when searching for producers passively,but also omits the traffic overhead of forwarding interest packets in batches when actively searching for producers.Experimental results show that the improved mechanism has better performance in terms of delay and overhead.