Bell Labs Seminars in Network and Service Management


Statistical Methods for Fault Management

Cindy Hood
Dept. of Computer Science and Engineering
Illinois Institute of Technology, Chicago

11-12noon, Friday August 1

Room 4C-511 (Note: room change), Holmdel

Abstract

The increasing role of communication networks in today's society results in a demand for higher levels of network availability and reliability. As our dependency on networks increases, faults and downtime become very costly. At the same time, fault management is becoming more difficult due to the dynamics and heterogeneity of the network. Current fault management systems rely heavily on the availability of network expertise to construct fault and/or network models. These systems must be customized to a particular network, and as such cannot be easily generalized from one network to another.
To reduce the need for prior network expertise, we have investigated the use of adaptive statistical methods for fault management. We have developed a monitoring approach that continually learns the behavior of the network and detects deviations from the norm. By detecting anomalies instead of specific faults, we are able to detect problematic behavior at an early stage. This talk will describe the approaches and methods studied, results obtained and ongoing research.

Contact: Pratyush Moghé at pmoghe@bell-labs.com