Intelligent fault management using cost sensitive fault remedial approach

Abstract

Organizations and individuals are becoming more dependent on computer networks to accomplish their daily tasks. As networks continuously evolve in size, heterogeneity, complexity and pervasiveness the risk of faults increases. Faults and downtimes are costly. Therefore, there is a need to automate fault detection and recovery process. The aim of this research is to develop an intelligent fault management system that automates the detection, identification and recovery of a network fault. A fault controller was proposed to identify the source of a fault after which a corresponding repair action is executed to resolve the fault. The methodology used to develop the fault controller is the Cost-sensitive Fault Remediation (CSFR) algorithm. The CSFR-based fault controller provides an intelligent and timely response upon the detection of an IP address conflict problem. The output of the framework is a policy consisting of an ordered list of test actions that can be executed to identify the fault source of an IP address conflict event after which an appropriate repair action is executed to solve the problem. The CSFR-based fault controller is a partially automated system that successfully identifies the source of fault of an IP address conflict problem in the shortest time duration.