Facing Fault Management as It Is, Aiming for What You Would Like It to Be
Soft-Ware 2002 Proceedings of the First International Conference on Computing in an Imperfect World
Towards Autonomic Computing: Effective Event Management
SEW '02 Proceedings of the 27th Annual NASA Goddard Software Engineering Workshop (SEW-27'02)
Proceedings of the 44th annual Southeast regional conference
Autonomic networks: engineering the self-healing property
Engineering Applications of Artificial Intelligence
Hi-index | 0.00 |
As computer networks increase in size, heterogeneity, complexity and pervasiveness, effective management of such networks simultaneously becomes more important and more difficult. This paper explores in detail one aspect of network management, {\em fault identification}. Fault identification is the process whereby the existence and nature of network faults are ascertained. Characteristics of the fault identification problem are explored and existing approaches are surveyed. Interestingly, much of the work in this area makes use of techniques from Artificial Intelligence, especially expert systems. Features of the fault identification problem that make it is amenable to AI approaches and resistant to more traditional algorithmic solutions are examined. Finally, a new approach to fault identification is proposed that employs an algorithm for finding dependencies among values in multiple streams of data over time. The new approach is compared to existing approaches, and it advantages and disadvantages are weighed.