Research On Congestion Control Strategy For Content-Centric Networking

CCN(Content-Centric Networking) is a new network architecture.It is one of the hotspots of the future network. Congestion is still a basic problem in CCN due to the limited network resources and sudden traffic burstiness.With the features of one-to-one flow balance and multihoming,CCN requires a congestion control scheme that is significantly different from TCP.Congestion control is crucial for improving network performance and quality of service.Therefore research about congestion control for CCN has important significance.The current CCN congestion control researches focus on adjusting the interest rate in the receiver or CCN routers,without taking into account the type of content. Such regulatory mechanisms could affect the quality of service,especially for services that have high-demand for delay.On the other hand,congestion judgment in a lot of congestion control strategies are based on RTT prediction. However,mutihoming of content can lead to instable RTT prediction. In addition,differences between upstream link capacity and downstream link capacity are less considered in the existing congestion control strategies based on hop-by-hop.For the problem that congestion control strategies currently proposed haven’t considered types of services,this thesis propose a service-oriented congestion control strategy with explicit feedback.The CCN router proactively detect pending Interest packet queue,judging the network status according to the number of different services.In order to ensure the quality of more services,the router feedbacks the network status to some of receivers according to the algorithm.Moreover,given the CCN router is more open,this thesis propose a service-oriented congestion control strategy based on hop-by-hop,shaping Interest rate depending on flow in the queue.Both strategies can solve the congestion problem while more services’QoS are ensured.As for the instable RTT prediction,this thesis propose an improved RTT prediction mechaniam. Identifiers of subsequent data packets it intends to request are added in each Interest packet. In the path to the content,the routers indicate whether they have cached such data packets.Thus,we can calculate related RTTs according to the information in the packets.Combined with the improved RTT prediction mechanism,this thesis carries out research on a new joint hop-by-hop and receiver-driven congestion control strategy.The router shaping Interest rate in consideration of link capacity of upstream and downstream,feedbacking the congestion message to receivers according to RTT prediction.And then receivers adjust congestion windows smoothly.The evaluation shows it effectively solve the congestion problem. At the same time,it takes full advantage of bottleneck bandwidth and reduces the flow completion time.This thesis propose a service-oriented congestion control strategy and a new joint congestion control strategy.The relevant work in this thesis provides a new perspective for the research on this direction.