A Novel Approach in Adaptive Traffic Prediction in Self-sizing Networks Using Wavelets

Abstract

In this paper we propose a traffic predictor based on multiresolution decomposition for the adaptive bandwidth control in locally controlled selfsizing networks. A selfsizing network can provide quantitative packet-level QoS to aggregate traffic by allocating link/switch capacity automatically and adaptively using online traffic data. We show that wavelet based adaptive bandwidth control method performs better than other classical methods in the case of average queue size and maximum buffer size. We have compared the performance of different orthonormal wavelets and found that all the other wavelets do far better than Haar with respect to bandwidth utilization factor but Haar shows a very good queue performance. We have studied the effect of other wavelet parameters such as window size, number of decomposition levels and number of filter coefficients. We also introduce a novel adaptive wavelet predictor which can adapt very well to the changes of incoming bursty traffic based on different window sizes and decomposition levels.