A survey of fault localization techniques in computer networks

Abstract

Fault localization, a central aspect of network fault management, is a process of deducing the exact source of a failure from a set of observed failure indications. It has been a focus of research activity since the advent of modern communication systems, which produced numerous fault localization techniques. However, as communication systems evolved becoming more complex and offering new capabilities, the requirements imposed on fault localization techniques have changed as well. It is fair to say that despite this research effort, fault localization in complex communication systems remains an open research problem. This paper discusses the challenges of fault localization in complex communication systems and presents an overview of solutions proposed in the course of the last ten years, while discussing their advantages and shortcomings. The survey is followed by the presentation of potential directions for future research in this area.

Network events in a large commercial network: What can we learn?

“…Since the presence of temporal correlations between emitted event times can provide insights into possible functional relationships between the devices that emit them [6], we used standard correlation techniques (see [18], for example) to gain further insights into the dataset. Namely, we assessed the presence of a functional relationship between two devices in terms of whether the Pearson's correlation coefficient between the time series of their respective recorded alert timestamps significantly differed from that expected under the null hypothesis that they were independent.…”

“…A lot of research has been done on algorithms performing Root Cause Analysis (RCA) [6]; i.e. identifying network events that escalate to actual network and service outages.…”

Deriving a unified fault taxonomy for event-based systems

“…Figure 5 depicts the six categories of fault sources, which have been extracted and compiled from earlier work on fault localization [3,63] and root cause analysis [38]. Where applicable, the fault source description refers back to elements of the model discussed in Section 2.…”

“…The work of Steinder and Sethi [63] surveys approaches and techniques for fault localization in computer networks, largely focusing on graph-theoretic fault propagation models like dependency networks and causality graphs. Their contribution provides much more general granularity than our work and parts of the faults discussed here are covered by their approach (e.g., detecting circular dependencies in EPNs).…”

Towards an approximate graph entropy measure for identifying incidents in network event data

“…Events occur at a high rate, and do not always directly indicate a problem. To illustrate, at a typical large enterprise network 1 , the event rate is 135 million events a day, generated by just a few hundred 'actionable incidents'.…”