基于SDN具有缓存感知的NDN节点拥塞控制策略  

NDN Node Congestion Control Strategy with Cache-Awareness Based on SDN

作　　者：刘开华 侯雅俊 李卓 宫霄霖 罗蓬 Liu Kaihua;Hou Yajun;Li Zhuo;Gong Xiaolin;Luo Peng(School of Microelectronics,Tianjin University,Tianjin 300072,China;Tianjin Key Laboratory of Imaging and Sensing Microelectronic Technology,Tianjin 3000072,China;Peng Cheng Laboratory,Shenzhen 518000,China;Electric Power Research Institute,State Grid Hebei Electric Power Co.,Ltd.,Shijiazhuang 050021,China)

机构地区：[1]天津大学微电子学院,天津300072 [2]天津市成像与感知微电子技术重点实验室,天津300072 [3]鹏城实验室,深圳518000 [4]国网河北省电力有限公司电力科学研究院,石家庄050021

出　　处：《天津大学学报（自然科学与工程技术版）》2023年第2期214-220,共7页Journal of Tianjin University：Science and Technology

基　　金：河北省省级科技计划资助项目(20314301D);天津市科技计划资助项目(20JCQNJC01490);鹏城实验室资助项目(PCL2021A02).

摘　　要：命名数据网(named data networking,NDN)作为一种新型的互联网架构,旨在应对日益增长的数据流量.然而,随着用户需求进一步增长,拥塞控制对于多路径传输的命名数据网仍然是一个具有挑战性的问题,亟需一种能够快速地检测网络拥塞和有效地管理网络资源的拥塞控制机制.针对这一问题,提出了一种基于软件定义网络技术的、具有缓存感知功能的命名数据网节点拥塞控制策略,称BCMCC.首先,介绍了BCMCC的新型网络架构.利用软件定义网络控制平面与数据平面解耦合技术,该架构将拥塞控制功能集中于SDN控制器中,以实现集中式节点拥塞控制、降低节点运行负荷的目的.其次,基于新型网络架构,研究了BCMCC的缓存感知算法和多路径选择拥塞控制算法.其中,缓存感知算法实现了网络缓存内容的感知以及缓存内容全局流行度的计算,能够利用命名数据网的节点缓存特性,降低缓存内容对拥塞控制的影响;多路径选择拥塞控制算法协同节点更新转发端口信息以实现流量迁移、智能化管理多路径容量,达到提高网络资源利用率、有效避免和缓解网络拥塞的目的.最后,在ndnSIM仿真平台进行BCMCC的性能测试.实验结果表明,BCMCC在丢包量、网络吞吐量以及传输时延等方面优于Best-Route算法、NCC算法和Random算法.在网络流量相同的情况下,BCMCC平均传输时延比Random低14.16%,比NCC低22.68%,比Best-Route低17.24%,同时相比Best-Route具有更稳定的网络吞吐量.BCMCC可以有效解决命名数据网拥塞问题.Named data networking(NDN)is a novel internet architecture,which aims to cope with the increasing data traffic.However,as user demand increases,congestion control remains a challenging problem for NDN,which enables multi-path content retrieval.Node congestion control based on the multi-path forwarding strategy in NDN is required to detect congestion rapidly and manage network resources efficiently.Thus,a congestion control strategy for NDN with cache-awareness based on software-defined network(SDN)technology,called BCMCC,was proposed.First,the network architecture of BCMCC,which uses the feature of decoupling the control and data planes of the SDN,centralizes congestion control in the controller to achieve centralized node control and reduce the node operation load.Second,based on the novel network architecture,the cache-awareness algorithm and multipath selection congestion control algorithm of BCMCC were proposed.The cache-awareness algorithm realized network cache content awareness and the calculation of the global prevalence of cache content to effectively utilize the node caching characteristics of the NDN and reduce the impact of cache content on congestion control.Meanwhile,the multi-path selection congestion control algorithm utilized the nodes to update forwarding port information for traffic migration to intelligently manage multi-path capacity,improve network resource utilization,and effectively prevent network congestion.Finally,the performance of BCMCC was tested in the simulation platform of ndnSIM.The experimental results showed that the BCMCC was superior to other schemes,including the Best-Route,NCC,and Random algorithms,in terms of packet loss,throughput,and transmission delay.The transmission delay of BCMCC is 14.16%lower than that of Random,22.68%lower than that of NCC,and 17.24%lower than that of BestRoute,with the same network traffic while having a more stable network throughput.BCMCC can effectively solve the NDN congestion problem.

关 键 词：命名数据网 缓存感知 多路径选择 拥塞控制 软件定义网络 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]