MPLS网络流量具有不确定性和突发性的特征,针对该特征分析了现有网络流量预测模型的不足,并提出了一种更具针对性的基于小波和卡尔曼滤波的流量预测模型。利用卡尔曼滤波目标的小波变换系数,通过与多尺度分析方法相结合,设计和实现了具备实时性和递归性,同时具有多尺度分析能力的小波—卡尔曼滤波混合流量预测模型。通过仿真模拟实验,结果表明MPLS流量预测算法相对于传统流量预测模型具有较好的实时性和预测精度,而且算法的复杂度也比较低。
The MPLS network traffic is of uncertainties and abruptness.For these characteristics,analyzed insufficiency of the current prediction models of the network traffic,and proposed a traffic prediction model based on wavelet and Kalman filter,which presented more efficiently.Using the wavelet coefficients of the Kalman filter target state and combining with the method of multi-scale analysis,designed and implemented a method which had not only real-time and recursiveness,but also the wavelet with multi-scale analysis capabilities-Kalman filter mixed traffic prediction model.Did the simulation experiments,and the results show the MPLS traffic prediction algorithm has better prediction accuracy and real-time compared to the traditional traffic prediction model,and the algorithm complexity is lower.