Packet inter-arrival time estimation using neural network models

This paper presents the estimation methods of packet inter-arrival times. The approach is based on neural network models for predicting the packet inter-arrival times by learning patterns and characteristics of observed traffics. The proposed methods are compared with the familiar exponentially weighted moving average (EWMA) estimator. Evaluation results demonstrate the effectiveness of the proposed methods. We also investigate the effect of classifying the observed traffics according to their protocol types. The classification helps find patterns of the traffics.

[1]  Victor C. M. Leung,et al.  A new traffic rate estimation and monitoring algorithm for the QoS-enabled Internet , 2003, GLOBECOM '03. IEEE Global Telecommunications Conference (IEEE Cat. No.03CH37489).

[2]  Suresh Singh,et al.  A feasibility study for power management in LAN switches , 2004, Proceedings of the 12th IEEE International Conference on Network Protocols, 2004. ICNP 2004..

[3]  David Tse,et al.  A framework for robust measurement-based admission control , 1999, TNET.

[4]  Van Jacobson,et al.  Link-sharing and resource management models for packet networks , 1995, TNET.

[5]  Victor C. M. Leung,et al.  Efficient fair queuing with decoupled delay-bandwidth guarantees , 2001, GLOBECOM'01. IEEE Global Telecommunications Conference (Cat. No.01CH37270).

[6]  Murray Smith,et al.  Neural Networks for Statistical Modeling , 1993 .

[7]  Yoshitaka Shibata,et al.  Empirical study of inter-arrival packet times and packet losses , 2002, Proceedings 22nd International Conference on Distributed Computing Systems Workshops.

[8]  George M. Mohay,et al.  The use of packet inter-arrival times for investigating unsolicited Internet traffic , 2005, First International Workshop on Systematic Approaches to Digital Forensic Engineering (SADFE'05).