Title: Behavioural Proximity Discovery: an adaptive approach for root cause analysis

Authors: M-Tahar Kechadi, Jacques Henry Bellec, AbdelKamel Tari

Addresses: School of Computer Science and Informatics, University College Dublin, Belfield, Dublin 4, Ireland. ' School of Computer Science and Informatics, University College Dublin, Belfield, Dublin 4, Ireland. ' Department of Computer Science, University A. MIRA of Bejaia, Targua Ouzemour, 06000 Bejaia, Algeria

Abstract: This paper introduces a new approach to deal with the fault recognition problem in modern telecommunication networks, called Behavioural Proximity Discovery (BPD). BPD is a framework for root cause analysis that consists of three complementary clustering algorithms based on alarm behaviours. It provides a complete and adaptive solution to the network manager who can then decide of the repair actions to restore the network back to a normal state. An evaluation of this approach on both simulated and real data from 3G mobile-networks is presented and discussed.

Keywords: data mining; temporal data sets; fault management; network management; alarm correlation; fault recognition; telecommunications networks; root cause analysis; clustering algorithms; repair; 3G mobile networks.

DOI: 10.1504/IJBIDM.2011.041958

International Journal of Business Intelligence and Data Mining, 2011 Vol.6 No.3, pp.259 - 282

Published online: 22 Apr 2015 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article