Towards Distributed and Adaptive Detection and Localisation of Network Faults

Abstract

Abstract-We present a statistical probing-approach to distributed fault-detection in networked systems, based on autonomous configuration of algorithm parameters. Statistical modelling is used for detection and localisation of network faults. A detected fault is isolated to a node or link by collaborative fault-localisation. From local measurements obtained through probing between nodes, probe response delay and packet drop are modelled via parameter estimation for each link. Estimated model parameters are used for autonomous configuration of algorithm parameters, related to probe intervals and detection mechanisms. Expected fault-detection performance is formulated as a cost instead of specific parameter values, significantly reducing configuration efforts in a distributed system. The benefit offered by using our algorithm is fault-detection with increased certainty based on local measurements, compared to other methods not taking observed network conditions into account. We investigate the algorithm performance for varying user parameters and failure conditions. The simulation results indicate that more than 95% of the generated faults can be detected with few false alarms. At least 80% of the link faults and 65% of the node faults are correctly localised. The performance can be improved by parameter adjustments and by using alternative paths for communication of algorithm control messages.

Long-Term Adaptation and Distributed Detection of Local Network Changes

“…This model of interarrival times is motivated by the assumption that the probe response delays are sums of exponential transmission delays, caused by the queueing times in processing nodes. The assumption is supported by previous work [11], [12] and empirical network latency tests [2].…”

“…This allows for probing intervals that are adapted to local network behaviour, which can reduce the link load caused by probing [2].…”

“…Detection of temporal changes is often motivated by the interest of performing early or preventive actions to fault symptoms and anomalous behaviour (such as re-configuration of equipment), to maintain quality of service. Further, adaptation to various aspects of varying network behaviour is of particular interest, as it can be used to autonomously configure algorithm parameters, thereby reducing configuration efforts and improving algorithm performance [2].…”

A Survey of Fast-Recovery Mechanisms in Packet-Switched Networks

“…On the other hand, passive schemes only make use of data packet delivery to confirm correct operation of nodes and links (however, if data packets are not sent frequently enough, passive methods are hardly useful). The status of a given node can then be determined by other nodes either independently, or collaboratively [79]. It is worth noting that, failure detection in multilayer networks is often one of the most redundant tasks, as it is commonly performed at multiple layers [76].…”

A Survey of Fast Recovery Mechanisms in the Data Plane

“…On the other hand, passive schemes only make use of data packet delivery to confirm correct operation of nodes and links (however, if data packets are not sent frequently enough, passive schemes are hardly useful). The status of a given node can then be determined by other nodes either independently, or collaboratively [72].…”