| Abstract: | Global telecommunication systems have built-in redundancy to ensure robustness. Unfortunately, this means that each fault that occurs can trigger a cascade of alarm events as individual parts of the system discover and report failure. It is then difficult to locate the origin of the fault. One contribution to effective fault management is through practical intelligent assistance. The general goal is to find a rich set of strategies and rules to help locate faults from alarm event data. This paper presents a three-tier rule discovery framework that combines computer-assisted human discovery with human-assisted computer discovery techniques. The approach is illustrated with a small experimental case study performed on a test high-speed network at Nortel Networks. |
