Font Size: a A A

Research On Multi-Path Flow Allocation Method By Using Of Batched Sparse Codes

Posted on:2022-11-10Degree:MasterType:Thesis
Country:ChinaCandidate:D W SunFull Text:PDF
GTID:2518306764979059Subject:Automation Technology
Abstract/Summary:PDF Full Text Request
Recently,wireless multi-path transmission attracts much attention in wireless networks as a hotspot.However,existing multi-path protocols,such as MPTCP?CMTSCTP,suffer from weak cooperativity between sub-connections and large system overhead,resulting in low channel utilization.To enhenace the performance of multipath transmission,this thesis employs Batched Sparse Codes to improve the cooperativity between sub-connections and to suppress overhead.First,the designing goals are given for multi-path flow allocation using Batched Sparse Codes after the analysis of wireless multi-hop transmission model.In addition,a system model is proposed in conjunction with the definitions of some key parameters.Next,an ACK-based feedback mechanism is developed,which is composed of a periodical feedbackd and trigger feedbacks for sub-connections on each path.The periodical feedback fashion collects transmission quality metrics of each path periodically.The trigger feedback fashions are triggered by vigorous fluctuations of transmission quality parameters of each path to confirm the status of sub-connections.Sequentially,a multi-path flow allocation method using Batched Sparse Codes is devised.This method is composed of a master-flow allocation algorithm and a sub-path metrics updating algorithm.The master-flow allocation algorithm models the problem of primary connecting metrics configuration and the flow allocation as an integer optimization problem.Aiming at maximizing system throughput rate,this thesis gives a reasonable flow allocation strategy and corresponding coding parameter configuration by computing a sub-optimal solution using relaxation subproblem.The sub-path metrics updating algorithm firstly updates the end-to-end loss rate of information and the roundtrip time according to periodical feedbacks.This thesis then also proposes an inner coding rate updating algorithm to calculate and update the inner coding rate of sub-path.Lastly this thesis proposes transmission rate updating algorithm to calculate and update transmission rate of sub-path.Finally,the simulating program for multi-path transmission is conducted on MATLAB platform.Results show that,when the loss rate of all sub-paths is 20%,our proposed method achieves a significantly higher throughput than that of MPTCP by220% if three sub-paths are of same bandwidth and loss rate,by 285% if three sub-paths are of different bandwidth and loss rate,and by 235% if three sub-paths are of dynamic bandwidth and loss rate.
Keywords/Search Tags:Wireless Network, Multi-Path Transmission, Batched Sparse Codes, Periodical Feedback, Flow Allocation
PDF Full Text Request
Related items