Título: Temporal event correlation for fault localization in computer networks
Autores: Gu, Zhonghui.
Fecha: 1998
Publicador: McGill University - MCGILL
Fuente:
Tipo: Electronic Thesis or Dissertation
Tema: Computer Science.
Descripción: Alarm correlation is becoming one of the most central techniques in managing the high volume of alarm messages in fault management area, since the speed and accuracy of fault management are heavily dependent on it. The alarm correlation approach of this thesis helps operators to filter and condense the incoming alarms to meaningful high-level alarms, and diagnoses them automatically and efficiently. This research includes the comparison among existing alarm correlation approaches, and builds a class level network management system model. In managed system domain, dependency model among related objects for fault propagation is also built. In managing system, the alarm I correlation algorithm is represented by UML object oriented methodology, especially time constraints is integrated to this algorithm and nine temporal relations are built between a pair of alarms. Five alarm correlation methods are applied in combination or in sequence to determine causality relationships among a cluster of alarms, and efficiently filter redundant alarms. Part of the algorithm is also implemented in Java on Advent Web based network management platform, which is client/server architecture.
Idioma: en