Estimating SDN Traffic Matrix Based on Online Informative Flow Measurement Method
Dongyang Li, Ningyun Dai, Li Feng, Changyou Xing, Fei Dai · 2017
Traffic Matrix (TM) estimation is important to traffic engineering in data center networks. However, current estimation methods are insufficient in estimation accuracy and measurement cost. In this paper, by using the advantage of SDN, we propose an online informative flow measurement based traffic matrix estimation method IFTME, which uses an iterative method to select these most informative flows to measure, so as to decrease the system uncertainty. Besides, we also propose two different measurement flow selection strategies: Flow Intensity First (FIF) and Flow Variation First(FVF). The experiment results on two Internet measurement datasets show that our proposed mechanism can improve the accuracy of traffic matrix estimation effectively, and can also dynamically adjust the measurement flow selection choice according to the time-varying characteristics of network traffic.