Proactive network fault detection

Abstract

The increasing role of communication networks in today's society results in a demand for higher levels of network availability and reliability. At the same time, fault management is becoming more dificult due to the dynamic nature and heterogeneity of networks. We propose an intelligent monitoring system using adaptive statistical techniques. The system continually learns the normal behavior of the network and detects deviations from the norm. Within the monitoring system, the measurements are segmented, and features extracted from the segments are used to describe the normal behavior of the measurement variables. This information is combined in the structure of a Bayesian network. The proposed system is thereby able to detect unknown or unseen faults. Experimental results on real network data demonstrate that the proposed system can detect abnormal behavior before a fault actually occurs.

Automation of Cellular Network Faults

“…Managing complex hardware and software systems has always been a difficult task. The Internet and the proliferation of web-based services have increased the importance of this task, while aggravating the problem (faults) in at least four ways (Meira, 1997;Thottan & Ji, 1998;Hood & Ji, 1997a;Lazar et al, 1992):…”

“…The main reasons why Bayesian networks was chosen for cellular network faults prediction include (Hood et al, 1997a), (Hood et al, 1997b), (Hood et al, 1998):…”

A Traffic Decomposition and Prediction Method for Detecting and Tracing Network-Wide Anomalies

“…In the figure, at the point marked by the red circle, p = 7 indicates that there are 7 links experiencing anomalies at that point. Furthermore, q = (2545) 10 = (100111110001) 2 at that point indicates that #links 1,5,6,7,8,9, and 12 are experiencing anomalies. We mark these detected links with red lines on the shortest route tree, and get the anomaly propagation path shown in Fig.…”

Traffic Anomaly Detection at Fine Time Scales with Bayes Nets

“…Early efforts were focused on the practical problem of network fault detection [7], [8], and in using time series methods to detect traffic anomalies [9], [4]. Several studies have shown that entropy-based methods can be effective for anomaly detection [10], [11] including Xu et al who use entropy to classify traffic in packet traces taken from an ISP backbone [12].…”