A case-based reasoning approach to the management of faults in communications networks

Most fault resolution systems for communications networks represent fault resolution expertise using a rule-based reasoning (RBR) paradigm. Although the RBR paradigm is appropriate for problem-solving tasks that are confined and well-understood, its limitations are: an inability to learn from experience; an inability to deal with novel problems; and the difficulty of updating the systems to keep up with rapidly changing domains such as expanding heterogeneous networks. The author describes a case-based reasoning (CBR) application that is less constrained by these limitations. The approach is to enhance a standard fault management system with a CBR component. CBR techniques are reviewed for retrieving, adapting, and embedding knowledge in a case library, and CRITTER, a CBR trouble ticketing system for managing and resolving network faults, is described