ABSTRACT
Despite a large body of literature and methods devoted to the Traffic Matrix (TM) estimation problem, the inference of traffic flows volume from aggregated data still represents a major issue for network operators. Directly and frequently measuring a complete TM in a large-scale network is costly and difficult to perform due to routers limited capacities. In this paper we introduce and evaluate a new method to estimate a TM from easily available link load measurements. The method uses a novel statistical learning technique to unveil the relation between links traffic volume and origin-destination flows volume. By training a system based on Random Neural Networks, we provide a fast and accurate TM estimation tool that attains proper results without assuming any traffic model or particular behavior. Using real data from an operational backbone network, we compare this new method to the most well known and accepted TM estimation techniques, including in the evaluation some more accurate and up-to-date methods developed in recent works. Results show that current TM estimation techniques can still be improved.
- Y. Vardi, "Network tomography: estimating source-destination traffic intensities from link data", in J. Amer. Statist. Assoc, 91, pp. 365--377, 1996.Google ScholarCross Ref
- J. Cao et al, "Time-varying network tomography", in J. Amer. Statist. Assoc, 95, pp. 1063--1075, 2000.Google ScholarCross Ref
- A. Medina et al, "Traffic Matrix Estimation: Existing Techniques and New Directions", in ACM Sigcomm, 2002. Google ScholarDigital Library
- M. Roughan et al, "Experience in Measuring Backbone Traffic Variability: Models, Metrics, Measurements and Meaning", in ACM Sigcomm IMW, 2002. Google ScholarDigital Library
- Y. Zhang et al, "Fast Accurate Computation of Large-Scale IP Traffic Matrices from Link Load Measurements", in ACM Sigmetrics, 2003. Google ScholarDigital Library
- A. Lakhina et al, "Structural Analysis of Network Traffic Flows", in ACM Sigmetrics, 2004. Google ScholarDigital Library
- A. Soule et al, "Traffic Matrix Tracking using Kalman Filters", in LSNI, 2005. Google ScholarDigital Library
- A. Soule et al, "Traffic Matrices: Balancing Measurements, Inference and Modeling", in ACM Sigmetrics, 2005. Google ScholarDigital Library
- H. Ringberg et al, "Sensitivity of PCA for Traffic Anomaly Detection", in ACM Sigmetrics, 2007. Google ScholarDigital Library
- P. Casas et al, "Efficient Methods for Traffic Matrix Modeling and On-Line Estimation In IP Networks", in ITC21, 2009.Google Scholar
- D. Jiang et al, "Large-Scale IP Traffic Matrix Estimation Based in Backpropagation Neural Network", in IEEE ICINIS, 2008. Google ScholarDigital Library
- H. Bakircioglu et al, "Survey of Random Neural Network Applications", in E. J.O.R., v. 126 (2), pp. 319--330, 2000.Google ScholarCross Ref
- S. Mohamed et al, "A Study of Real-Time Packet Video Quality Using Random Neural Networks", in Trans. Circ. Syst. Video Tech., v. 12, pp. 1071--1083, 2000. Google ScholarDigital Library
- C. Bishop, "Neural Networks for Pattern Recognition", Oxford Uni. Press, 1995. Google ScholarDigital Library
- R. Duda et al, "Pattern Classification", Wiley-Inter, 2000. Google ScholarDigital Library
- E. Gelenbe, "Random neural networks with negative and positive signals and product form solution", in Neural Computation, v. 1, pp. 502--511, 1989. Google ScholarDigital Library
- E. Gelenbe, "Learning in the Recurrent Random Neural Networks", in Neural Computation, v. 5, pp. 154--164, 1993. Google ScholarDigital Library
- S. Uhlig et al, "Providing Public Intradomain Traffic Matrices to the Research Community", in ACM Sigcomm CCR, 2006. Google ScholarDigital Library
- Y. Zhang, "Abilene Dataset 2004", http://www.cs.utexas.edu/~yzhang, 2004.Google Scholar
Index Terms
- On the use of random neural networks for traffic matrix estimation in large-scale IP networks
Recommendations
Traffic matrix estimation on a large IP backbone: a comparison on real data
IMC '04: Proceedings of the 4th ACM SIGCOMM conference on Internet measurementThis paper considers the problem of estimating the point-to-point traffic matrix in an operational IP backbone. Contrary to previous studies, that have used a partial traffic matrix or demands estimated from aggregated Netflow traces, we use a unique ...
Large-Scale IP Traffic Matrix Estimation Based on Backpropagation Neural Network
ICINIS '08: Proceedings of the 2008 First International Conference on Intelligent Networks and Intelligent SystemsTraffic matrix estimation is significantly important for operators. However, it is difficult to estimate accurately traffic matrix. This paper proposes a novel method of large-scale IP traffic matrix estimation, termed the backpropagation neural network ...
GARCH model-based large-scale IP traffic matrix estimation
This letter proposes a novel method to estimate large-scale IP traffic matrix (TM). By using the Generalized Autoregressive Conditional Heteroscedasticity (GARCH) to model the Origin-Destination (OD) flows, we can easily get rid of the ill-posed problem ...
Comments