A Novel Method to Estimate IP Traffic Matrix

Traffic flow is not only short-range dependence shown in traditional models, but also self-similarity and long-range dependence. The coexistence of these made it hard to estimate traffic matrix (TM) by using the modules based on temporal dimension. This paper avoids to establishing models for TM. Using the TM´s spatial self-similarity, we expressed TM as a weighted linear combination of the sample OD flows. Compared with previous methods, our method does not only hold the lower estimation errors but also is more robust.