Research On Congestion Control Algorithm In Named Data Networking

With the rapid growth of network size and application types,the current Internet is facing unprecedented challenges in mobility,scalability and security.InformationCentric Networking(ICN),which is designed to overcome the limitations of IP networks,comes into being.Named Data Networking(NDN)is the most representative one.NDN as a revolutionary network architecture,there are many issues need to be solved,congestion control is the one of research focuses.This thesis is funded by the Chongqing Key Project of Research on Fundamental Science and Advanced Technology and undertakes the research on NDN congestion control.In NDN,there are two typical types of congestion: one is caused by burst traffic,which also exists in the traditional network;the other is caused by redundant data due to interest flooding and in-network caching,this thesis proposes Name-based Weighted Fair Queuing(NWFQ)and ELSA-based Redundant Control(ELSA-RC)based on Extended Link State Advertisements(ELSA)respectively.The primary contributions of this thesis are summarized as follows:For the first kind of congestion,due to the fact that the existing solutions do not fully consider the fairness between the data streams,a congestion control algorithm based on rate limiting is proposed,which is called NWFQ.The algorithm detects congestion at intermediate nodes and limits the forwarding rate of interest packets to prevent further congestion;In addition,since it is difficult for consumers to obtain the congestion feedback directly,an explicit feedback mechanism based on data packet is proposed.Simulation results based on ndnSIM show that NWFQ can effectively improve the bottleneck link utilization to 95%,while maintaining a low packet loss rate and average flow completion time,compared with other algorithms.For the second type of congestion,in order to reduce the number of redundant interest and data packets,this thesis proposes ELSA-RC.The algorithm adds a hop count database(hopDB)in the NDN nodes to record the latest hop count of interest,and then reduces the interest forwarding by sending ELSA to enhance the hop DB update of neighbor node;In addition,ELSA-RC prevents redundant data from returning based on the incoming interface of pending interest table and received ELSA.The simulation results show that compared with the traditional NDN,the number of interest and data packets decreased by about 15% and 19% respectively,and the average delay is reduced by about 14% with ELSA-RC.In short,for the two representative types of NDN congestion,this thesis proposes two solutions respectively,and through simulation experiments to verify that the proposed schemes can effectively control the burst traffic and reduce the redundant data.The relevant work in this thesis provides a new perspective for the NDN congestion control research.