37th Annual IEEE Conference on Local Computer Networks (2012)
Clearwater Beach, FL, USA USA
Oct. 22, 2012 to Oct. 25, 2012
Sebastian Zander , Centre for Advanced Internet Architectures (CAIA), Swinburne University of Technology, Melbourne Australia
Thuy Nguyen , Centre for Advanced Internet Architectures (CAIA), Swinburne University of Technology, Melbourne Australia
Grenville Armitage , Centre for Advanced Internet Architectures (CAIA), Swinburne University of Technology, Melbourne Australia
Machine Learning (ML) classifiers have been shown to provide accurate, timely and continuous IP flow classification when evaluating sub-flows (short moving windows of packets within flows). They can be used to provide automated QoS management for interactive traffic, such as fast-paced multiplayer games or VoIP. As with other ML classification approaches, previous sub-flow techniques have assumed all packets in all flows are being observed and evaluated. This limits scalability and poses a problem for practical deployment in network core or edge routers. In this paper we propose and evaluate sub-flow packet sampling (SPS) to reduce an ML sub-flow classifier's resource requirements with minimal compromise of accuracy. While random packet sampling increases classification time from <1 second to over 30 seconds and can reduce accuracy from 98% to <90%, our tailored SPS technique retains classification times of <1 second while providing 98% accuracy.
Accuracy, Training, Niobium, Games, Arrays, Quality of service, IP networks, Packet Sampling, Traffic Classification, Machine Learning
S. Zander, T. Nguyen and G. Armitage, "Sub-flow packet sampling for scalable ML classification of interactive traffic," 37th Annual IEEE Conference on Local Computer Networks(LCN), Clearwater Beach, FL, USA USA, 2012, pp. 68-75.