Cluster-Based Failure Detection Service for Large-Scale Ad Hoc Wireless Network Applications

  • Authors:
  • Ann T. Tai;Kam S. Tso;William H. Sanders

  • Affiliations:
  • IA Tech, Inc., Los Angeles, CA;IA Tech, Inc., Los Angeles, CA;University of Illinois, Urbana, IL

  • Venue:
  • DSN '04 Proceedings of the 2004 International Conference on Dependable Systems and Networks
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

The growing interest in ad hoc wireless network applicationsthat are made of large and dense populations oflightweight system resources calls for scalable approachesto fault tolerance. Moreover, the nature of these systemscreates significant challenges for the development of failuredetection services (FDSs), because their quality often dependsheavily on reliable communication. In particular, adhoc wireless networks are notoriously vulnerable to messageloss, which precludes deterministic guarantees for thecompleteness and accuracy properties of FDSs. To meetthe challenges, we propose an FDS based on the notion ofclustering. Specifically, we use a cluster-based communicationarchitecture to permit the FDS to be implementedin a distributed manner via intra-cluster heartbeat diffusionand to allow a failure report to be forwarded across clustersthrough the upper layer of the communication hierarchy.In doing so, we extensively exploit the message redundancythat is inherent in ad hoc wireless settings to mitigate theeffects of message loss on the accuracy and completenessproperties of failure detection. As shown by our mathematicalanalysis, the resulting FDS is able to provide satisfactoryprobabilistic guarantees for the desired properties.