Abstract
This paper presents a novel approach to deal with dynamic and highly uncertain traffic in dynamic network scenarios. The Reactive Robust Routing (RRR) approach is introduced, a combination of proactive and reactive techniques to improve network efficiency and robustness, simplifying network operation. RRR optimizes routing for normal-operation traffic, using a time-varying extension of the already established Robust Routing technique that outperforms the stable approach. To deal with anomalous and unexpected traffic variations, RRR uses a fast anomaly detection and localization algorithm that rapidly detects and localizes abrupt changes in traffic flows, permitting an accurate routing adaptation. This algorithm presents well-established optimality properties in terms of detection/localization rates and localization delay, which allows for generalization of results, independently of particular evaluations. The algorithm is based on a novel parsimonious model for traffic demands which allows for detection of anomalies using easily available aggregated-traffic measurements, reducing the overheads of data collection.
Similar content being viewed by others
References
Cisco Systems: Global IP Traffic Forecast and Methodology, 2006–2011. http://www.hbtf.org/files/cisco_IPforecast.pdf. Accessed 05 Aug 2010
Lakhina, A., Crovella, M., Diot, C.: Diagnosing Network-Wide Traffic Anomalies. Proceedings of the SIGCOMM’04 (2004). doi:10.1145/1015467.1015492
Ben-Ameur, W., Kerivin, H.: Routing of uncertain traffic demands. Optim. Eng. 6, 283–313 (2005)
Applegate, D., Cohen, E.: Making Intra-Domain Routing Robust to Changing and Uncertain Traffic Demands: Understanding Fundamental Tradeoffs. Proceedings of the SIGCOMM’03 (2003). doi:10.1145/863955.863991
Wang, H., Xie, H., Qiu, L., Yang, Y., Zhang, Y., Greenberg, A.: COPE: Traffic Engineering in Dynamic Networks. Proceedings of the SIGCOMM’06 (2006). doi:10.1145/1151659.1159926
Roughan, M., Thorup, M., Zhang, Y.: Traffic Engineering with Estimated Traffic Matrices. Proceedings of the IMC’03 (2003). doi:10.1145/948205.948237
Zhang, C., Liu, Y., Gong, W., Kurose, J., Moll, R., Towsley, D.: On Optimal Routing with Multiple Traffic Matrices. Proceedings of the INFOCOM’05 (2005). doi:10.1109/INFCOM.2005.1497927
Casas, P., Vaton, S.: An Adaptive Multi Temporal Approach for Robust Routing. Euro-FGI Workshop on IP QoS and Traffic Control, IST Press, Portugal (2007)
Kandula, S., Katabi, D., Davie, B., Charny, A.: Walking the Tightrope: Responsive yet Stable Traffic Engineering. Proceedings of the SIGCOMM’05 (2005). doi:10.1145/1090191.1080122
Elwalid, A., Jin, C., Low, S., Widjaja, I.: MATE: MPLS Adaptive Traffic Engineering. Proceedings of the INFOCOM’01 (2001). doi:10.1109/INFCOM.2001.916625
Zhang, Y., Ge, Z., Greenberg, A., Roughan, M.: Network Anomography. Proceedings of the IMC’05, 317–330 (2005)
Soule, A., Salamatian, K., Taft, N,: Combining Filtering and Statistical Methods for Anomaly Detection. Proceedings of the IMC’05, 331–344 (2005)
Ringberg, H., Soule, A., Rexford, J., Diot, C.: Sensitivity of PCA for Traffic Anomaly Detection. Proceedings of the SIGMETRICS’07 (2007). doi:10.1145/1254882.1254895
Casas, P., Fillatre, L., Vaton, S., Chonavel, T.: Efficient Methods for Traffic Matrix Modeling and On-line Estimation in Large-Scale IP Networks. Proceedings of the ITC 21 (2009)
Casas, P., Fillatre, L., Vaton, S.: Robust and Reactive Traffic Engineering for Dynamic Traffic Demands. Proc. NGI’08 (2008). doi:10.1109/NGI.2008.16
Casas, P., Fillatre, L., Vaton, S.: Multi Hour Robust Routing and Fast Load Change Detection for Traffic Engineering. Proceedings of the IEEE ICC’08 (2008) doi:10.1109/ICC.2008.1081
Ben-Ameur, W.: Between Fully Dynamic Routing and Robust Stable Routing. Proceedings of the DRCN’07 (2007). doi:10.1109/DRCN.2007.4762277
Pioro, M., Medhi, D.: Routing, Flow, and Capacity Design in Communication and Computer Networks. Elsevier/Morgan Kaufmann, Amsterdam (2004)
Zhang, Y.: Abilene Dataset (2004). http://userweb.cs.utexas.edu/~yzhang/. Accessed 12 Aug 2010
Abilene Obs.: http://www.internet2.edu/observatory. Accessed 12 Aug 2010
Coates, M., Hero, A., Nowak, R., Yu, B.: Internet tomography. IEEE Signal Process. Mag. 19(3), 47–65 (2002)
Medina, A., Salamatian, K., Bhattacharyya, S., Diot, C.: Traffic Matrix Estimation: Existing Techniques and New Directions. Proceedings of the SIGCOMM’02 (2002). doi:10.1145/964725.633041
Zhang, Y., Roughan, M., Duffield, N., Greenberg, A.: Fast Accurate Computation of Large-Scale IP Traffic Matrices from Link Load Measurements. Proceedings of the SIGMETRICS’03 (2003). doi:10.1145/781027.781053
Nürnberger, G.: Approximation by Spline Functions. Springer, New York (1989)
TOTEM: http://totem.run.montefiore.ulg.ac.be/. Accessed 12 Aug 2010
Soule, A., Lakhina, A., Taft, N., Papagiannaki, K., Salamatian, K., Nucci, A., Crovella, M., Diot, C.: Traffic Matrices: Balancing Measurements, Inference and Modeling. Proceedings of the SIGMETRICS’05 (2005). doi:10.1145/1071690.1064259
Lakhina, A., Crovella, M., Diot, C.: Characterization of Network-Wide Anomalies in Traffic Flows. Proceedings of the IMC’04 (2004). doi:10.1145/1028788.1028813
Nikiforov, I.: A lower bound for the detection/isolation delay in a class of sequential tests. IEEE Trans. Inf. Theor. 49(11), 3037–3047 (2003)
Nikiforov, I.: A simple recursive algorithm for diagnosis of abrupt changes in random signals. IEEE Trans. Inf. Theor. 46(7), 2740–2746 (2000)
Basseville, M., Nikiforov, I.: Detection of Abrupt Changes: Theory and Applications. Prentice Hall, NJ (1993)
Author information
Authors and Affiliations
Corresponding author
Appendix: Elimination of the Anomaly-free Traffic
Appendix: Elimination of the Anomaly-free Traffic
The anomaly-free traffic \(H\varvec{\mu}_t\) is removed by projecting the whitened measurements vector \({\bf z}_t = \Upphi^{-\frac{1}{2}}{\bf y}_t\) onto the left null space of H (i.e. WH = 0). Let us define the matrix W T = (w 1, .., w r-q) of size r × (r − q), composed of eigenvectors w 1, .., w r-q of the projection matrix \(P_H^{\bot} = I_r - H{\left(H^TH\right)}^{-1}H^T\), corresponding to eigenvalue 1. The matrix W satisfies the following conditions: \(WH = 0, W^TW = P_H^{\bot}\), and WW T = I r-q. The matrix W can be considered as a linear rejector that eliminates the anomaly-free traffic. In the presence of an anomaly in OD flow j, traffic residuals u t = W z t can be modeled as \({\bf u}_t = \theta W \Upphi^{-\frac{1}{2}} {\bf r}_j + W\varvec{\epsilon}_t\). By defining the vectors \({\bf v}_j=W\Upphi^{-\frac{1}{2}} {\bf r}_j\), we get the following distribution for traffic residuals in the presence of an anomaly in OD flow \(j: {\bf u}_t \sim {\mathcal N}(\theta {\bf v}_j,I_{r-q})\).
Rights and permissions
About this article
Cite this article
Casas, P., Fillatre, L., Vaton, S. et al. Reactive Robust Routing: Anomaly Localization and Routing Reconfiguration for Dynamic Networks. J Netw Syst Manage 19, 58–83 (2011). https://doi.org/10.1007/s10922-010-9182-9
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10922-010-9182-9