摘要
由于网络数据传输时常出现乱序问题,异常数据无法及时提取,导致数据接收端易出现缓存阻塞问题,降低了带宽利用率。为此提出基于时序预测的网络数据传输阻塞控制方法。对网络数据离散化处理,通过字符描述不同时间段数据的范围,采用时序预测模型获取下一阶段的网络数据传输量。分析网络数据传输量的空间相关性,对全部网络数据完成奇异值分解降维处理,消除冗余数据。通过网络数据相关特性,采用时间序列组建置信区间,检测网络异常数据。将节点间路径的抵达时间作为衡量指标,分别给出不同链路的尝试概率,获取最佳数据传输速率,实现网络数据传输控制。实验结果表明,所提方法缩短了网络数据阻塞时间,节点受限次数以及传输能耗也得以有效降低,全面提升了网络带宽利用率。
Due to the disorder during network data transmission,abnormal data is unable to be extracted in time,resulting in low bandwidth utilization.Therefore,this paper presented a method of controlling network data transmis-sion congestion based on time series prediction.Firstly,network data was discretized.And then,the range of data in different periods was described by characters.Secondly,a time-series prediction model was used to calculate the a-mount of network data transmission in the next stage.Thirdly,the spatial correlation of the amount was analyzed.After that,singular value decomposition and dimensionality reduction are implemented on all network data,thus eliminating redundant data.Based on the correlative characteristics of network data,the time series was used to construct a confi-dence interval detecting network abnormal data.Moreover,the arrival time of the path between nodes was taken as the measurement index.Respectively,the tentative probabilities of different links were given,so that the best transmission rate could be obtained.Finally,we achieved the network data transmission control.Experimental results show that the proposed method shortens the network data blocking time and effectively reduces the number of node constraints and transmission energy consumption.In addition,the method also comprehensively improves the utilization of network bandwidth.
作者
刘松旭
符太东
LIU Song-xu;FU Tai-dong(Management Center of Big Data and Network,Jilin University,Changchun Jilin 130000,China)
出处
《计算机仿真》
北大核心
2023年第11期379-383,共5页
Computer Simulation
关键词
时序预测
冗余数据
数据传输
阻塞控制
置信区间
Time-series prediction
Redundant data
Data transmission
Blocking control
Confidence interval
