A Distributed and Reliable Platform for Adaptive Anomaly Detection in IP Networks

  • Authors:
  • L. Lawrence Ho;Christopher J. Macey;Ronald Hiller

  • Affiliations:
  • -;-;-

  • Venue:
  • DSOM '99 Proceedings of the 10th IFIP/IEEE International Workshop on Distributed Systems: Operations and Management: Active Technologies for Network and Service Management
  • Year:
  • 1999

Quantified Score

Hi-index 0.00

Visualization

Abstract

Algorithms for anomaly detection in IP networks have been developed and a real-time distributed platform for anomaly detection has been implemented. These algorithms automatically and adaptively detect "soft" network faults (performance degradations) in IP networks. These algorithms are implemented as a reliable and fully distributed real-time software platform called NSAD (Network/Service Anomaly Detector). IP NSAD has the following novel features. First, it provides a flexible platform upon which preconstructed components can be mixed/matched and distributed (to different machines) to form a wide range of application specific and fully distributed anomaly detectors. Second, anomaly detection is performed on raw network observables (e.g., performance data such as MIB2 and RMON1/2 variables) and algebraic functions of the observables (objective functions), making NSAD an objective driven anomaly detection system of wide detection range and high detection sensitivity. Third, controlled testing demonstrates that NSAD is capable of detecting network anomalies reliably in IP networks.