Internet video traffic classification using QoS features

Abstract

This paper addresses the issue of effective classification of video traffic with the view of QoS guarantee, and presents a modified K-Singular Value Decomposition (K-SVD) classification framework based on the concept of QFAg (QoS based Flow Aggregation). By statistical analysis of video flows on large-scale real networks, we define 5 Quality of Service (QoS) categories with the features of downstream/upstream rates. To investigate the sparsity property of multimedia QoS feature, this paper utilizes modified K-SVD to train dictionary extracted from training samples. By learning feature-set to obtain sparse representation for video traffic, we propose a feature-based method to classify video traffic. Experimental results reveal that the proposed method can improve the classification performance significantly compared to previous methods.