Bivariate Non-parametric Anomaly Detection
Christian Callegari, Stefano Giordano, Michele Pagano · 2014
Detecting anomalous traffic with low false alarm rates is of primary interest in IP networks management. In this paper we propose a novel anomaly detection system, based on a combined use of sketches and of a novel bivariate non-parametric detection method. The latter allows us to simultaneously analyse two different traffic features so as to improve the performance of the "classical" detection systems, in terms of both detection rate and false alarm rate. The preliminary performance analysis, presented in this paper, demonstrates the effectiveness of the proposed system.