Packet Size Modeling And Optimization For Network Coding Enabled Wireless Networks

The network coding enables the intermediate network node to encode and combine the received multiple packets before transmitting,which can significantly improve the performance of the network.Studies have shown that compared with traditional methods,network coding can significantly improve network throughput,improve transmission reliability,and reduce delivery delay.In the wireless network environment,network coding can make full use of the broadcast characteristics of the wireless channel,which can further improve the performance of the wireless network,and has received extensive attention in wireless network unicast and multicast communication.However,the lossy nature of wireless channel greatly affects the efficiency and performance of wireless network coding.In the case of poor channel conditions,the longer a packet,the greater the probability of packet loss;In the case of better channel conditions,the shorter a packet is,the bigger the overhead due to fixed packet header will be.Therefore,consider wireless channel conditions,how to optimize the packet size to maximize the throughput of a wireless network is a key problem.In this paper,considering unicast and multicast application scenarios respectively,the throughputs of different scenarios are modeled for inter-flow network coding and intra-flow network coding,so as to solve the optimal packet size while maximizing network throughput.Simulations are conducted for performance validation.The main research work of this thesis is as follows:(1)Consider two-hop lossy wireless networks supporting IEEE 802.11,this thesis studies the optimal packet size in the context of inter-flow network coding.First of all,according to different packet receptions,analyze the corresponding coding gain,broadcast gain,and saturation channel access gains,and also the success probability of different packet reception scenarios,on this basis,the network throughput expressions are given,and then use the dichotomy solving the maximum throughput of the optimal packet size,finally,simulation results based on NS-2 validates the high accuracy of our analysis results.(2)For wireless one-hop multicast scenarios using random linear combination network coding for batch forwarding,consider uniform random independent loss link model,this thesis establishes the mathematical relationship between packet size and multicast throughput,and derive the optimal packet size,the simulation results of NS-3,verifies the accuracy of the analysis results.