摘要
网络流量预测是网络性能管理的一个重要组成部分,较好的流量预测能够提高网络管理的效果和网络带宽的利用率.以C/S模式对校园网的流量进行采集,然后以预测误差的平方最小为目标,建立动态指数平滑模型,增强了指数平滑模型对时间序列的适应能力,较好地解决了指数平滑模型中平滑参数α为静态而导致预测偏差等问题.通过测试,此模型能够较准确地预测出校园网的流量,从而实现了对校园网络流量的临控,提高了网络服务的质量.
The prediction of the network traffic is an important part of the network performance management.An accurate prediction of network traffic can improve the effect of network management and network bandwidth utilization.The campus network traffic by C/S mode was collected and the optimal model of the dynamic exponential smoothing model was established aiming at the square of forecast errors,through which the corresponding dynamic parameters can be obtained impersonally.It will enhance the adaptability of exponential smoothing model on time sequence and solves the problems that smoothing parameter is static so to generate prediction deviation and etc.The test result shows this model can predict the network traffic more accurately,achieving the monitoring of campus network traffic and improving the quality of network services.
出处
《内蒙古科技大学学报》
CAS
2009年第4期330-332,共3页
Journal of Inner Mongolia University of Science and Technology
基金
内蒙古教育科研基金资助项目(NJzy08077)
关键词
网络流量预测
校园网络
监控
动态指数平滑
network traffic prediction
campus network
monitor
dynamic exponential smoothing
