Efficient probe selection algorithms for fault diagnosis

Abstract

Increase in the network usage for more and more performance critical applications has caused a demand for tools that can monitor network health with minimum management traffic. Adaptive probing has the potential to provide effective tools for end-to-end monitoring and fault diagnosis over a network. Adaptive probing based algorithms adapt the probe set to localize faults in the network by sending less probes in healthy areas and more probes in the suspected areas of failure. In this paper we present adaptive probing tools that meet the requirements to provide an effective and efficient solution for fault diagnosis for modern communication systems. We present a system architecture for adaptive probing based fault diagnosis tool and propose algorithms for probe selection to perform failure detection and fault localization. We compare the performance and efficiency of the proposed algorithms through simulation results.

A multi-factor integrated fault diagnosis method for distributed applications

“…In this paper, various properties of the probes are taken into account, and the objective weighting and fuzzy evaluation method is used to evaluate the probes. Pseudo-code is as follows: PSFL algorithm: probes selection in the fault diagnosis The cost properties of the probe includes detection time, detection flow, the complexity .The efficiency properties of the probe include the detection quality and the information entropy [6]. Give priority to probes covering fewer faults, when the probe fails, it can greatly reduce the space of suspicious failures.…”

“…For the problem, some use the minimum and maximum search algorithm [6]and information entropy-based algorithm [2], They are all adopt the number of detection sets as the evaluation standard, the solutions are approximate, and can only give one detection set. However, in the actual environment, the costs and efficiency of different probes are different, there are also clear differences in multiple aspects, such as detection time, the required flow rate, probe complexity.…”

Probe generation for active probing

“…The target probe set should be able to diagnose the same collection of faults as the CPS and should ideally be much smaller in size. However, the selection of best possible target probe set from the CPS is proved to be NP‐Complete in Natu et al Several heuristic‐based algorithms are proposed in several studies …”

“…Several approaches have been proposed in the literature for monitoring networks for anomalies and faults. One of these approaches is active probing, [1][2][3][4][5][6][7][8][9][10][11][12] in which one or more dedicated nodes in the network, called probing stations, send out packets called probes to monitor the state of the network. Examples of probes may be ping or traceroute, which can obtain performance statistics such as latency, loss, and throughput of network links.…”

Recent Advances in Fault Localization in Computer Networks

“…The fault localization techniques that are based on a model representation of the network are faced with the challenges of obtaining and maintaining accurate information [15], [29], [36]. The accuracy of a technique directly depends upon the accuracy of the information in the model.…”