ABSTRACT
As communication networks continue to increase in size and complexity to support new applications and large numbers of new users, understanding and managing network behavior becomes increasingly difficult. The goal of intelligent network monitoring is to identify problems at a stage when they can be addressed and catastrophic failures can be avoided. One of the most common techniques used for network fault management can be considered as a change detection approach. The measurements collected were all related to network protocols so it was natural to organize the information into a hierarchy. Bayesian networks have been widely used for medical diagnosis and troubleshooting. A learning window is used to continually adjust the view or baseline of normal behavior. By detecting subtle changes from normal and then correlating the information in a Bayesian network it is possible to do proactive fault detection. Both the information processing and information fusion areas need more work.
