Research On Congestion Control In Content Centric Network

With the continuous development of multimedia technology and network technology,the scale of network users,the number of network equipment and the network video traffic are all showing an explosive growth trend.More and more web applications tend to focus on content.They care more about the content itself and do not care about the location of the content.Driven by the new application mode,the traditional TCP/IP network architecture based on host-to-host communication mode faces a series of problems,such as scalability,mobility,security and so on.Information Centric Networking emerges,which is looked as a new content-centric network architecture.It decouples the content from the storage address,uses the content name to identify and route the data and uses ubiquitous cache and flexible forwarding engine to transmit content.ICN uses this method to solve the problem in TCP/IP network well.Content Centric Networking is one of the most representative network architectures in ICN.It has the characteristics of receiver driven,hop-to-hop sending and universal caching,so the traditional end-to-end resource-sharing mode has been fundamentally changed,which makes TCP congestion control for sliding window cannot be carried out.Most of the existing congestion control mechanisms control the content request rate or the forwarding rate.They seldom consider the fairness problem caused by the different popularity content in the congestion control process and cannot effectively use the network idle link to avoid congestion.From the point of fairness of congestion control and congestion avoidance combined with forwarding strategy,this paper makes a deep research on congestion control mechanism of CCN network.The main contributions are as follows:Firstly,we deeply analyze the reason why TCP congestion control is no longer applicable to ICN,study and compare the existing congestion control mechanisms of ICN.In order to solve the problem of fairness in congestion control of CCN,we propose a quantitative measure of fairness-fairness factor and design a fair congestion control strategy based on flow(Fair Flow Congestion Control,FFCCON).FFCCON can detect congestion by monitoring queue length on routers and use different marking ratios to mark packets to notify congestion according to the popularity of different flows.At the same time,the request rates of different flows are controlled on the client side differently according to the number of tags and the popularity of the messages.The strategy guarantee the fairness of the flows while doing the congestion control.Secondly,we take advantage of the characteristics of ubiquitous cache and multipath forwarding in CCN,design a congestion avoidance strategy based on multipath forwarding(Multipath Forwarding based Congestion Avoidance,MFCA).MFCA monitors the queue length,forewarns the congestion status of the network in advance.In MFCA,routers can transfer different proportions of traffic to other available paths according to the popularity of the flow before congestion occurs in order to avoid congestion.At the same time,by analyzing,we improves the forwarding time of multipath forwarding to single path forwarding after the network state changes,which further improve the efficiency of congestion control and network utilization.Finally,we extend the CCN simulation program ccnSim on simulation platform OMNet++,achieve a fair congestion control strategy based on flows(FFCCON)and two congestion avoidance strategies based on multipath forwarding(MFCA and IMFCA),and verify their correctness,fairness and other performance.