一种基于缓存交互的命名数据网络拥塞控制算法  

A content store-based congestion control algorithm for named data networking

作　　者：石珊姗 任勇毛[1] 李俊[1] 李灵玲[1] 智江 

机构地区：[1]中国科学院计算机网络信息中心,北京100190 [2]中国科学院大学,北京100049

出　　处：《高技术通讯》2016年第4期359-366,共8页Chinese High Technology Letters

基　　金：973计划(2012CB315803);中国科学院计算机网络信息中心"一三五"计划(CNIC PY-1401)资助项目

摘　　要：研究了命名数据网络(NDN)的拥塞控制。为了解决突发流量问题和提高吞吐量及网络资源利用率,考虑了路由器缓冲区大小与拥塞控制机制的相互影响以及NDN内部署缓存这一重要特性,提出了一种基于缓存交互的NDN拥塞控制算法。该算法通过利用NDN中的路由器缓存,在逻辑上动态扩充缓冲区大小并控制Data包的发送速率,同时与现有的NDN拥塞控制算法相结合,动态调整Interest包发送速率阈值,以平滑突发流量,缓解网络拥塞。基于ndn SIM的仿真实验结果表明,该算法能有效提高NDN的传输效率、吞吐量和网络资源利用率。The congestion control of named data networking （NDN） was studied. To solve the bursty traffic problem, in- crease the throughput, and improve the network resource utilization, a content store-based congestion control algo- rithm for NDN was proposed under the considerations of the mutual influence between the buffer size of the router and the congestion control mechanism, as well as the in-network caching scheme of NDN. This algorithm logically performs the dynamical extension of the buffer size and controls the forwarding rate of Data packets by utilizing the router content store of NDN, and dynamically adjusts the sending rate threshold of Interest packets to smooth the bursty traffic and alleviate the network congestion by the combination with the existing NDN congestion control algo- rithms. The results of the experiment conducted based on ndnSIM indicate that this algorithm can effectively im- prove NDN＇s transmission efficiency, throughput and the network resource utilization.

关 键 词：命名数据网络(NDN) 拥塞控制 缓存 缓冲队列大小 动态阈值(DT) 突发流量 ' 

分 类 号：TP393.06[自动化与计算机技术—计算机应用技术]