A taxonomy is presented of knowledge-based approaches to fault management for telecommunications networks. Fault management systems are first characterized as multistage process systems. Using this characterization, a discussion is presented of a variety of current systems reported in the open literature that implement automated fault management and which span all processing stages (from alarm receipt to fault correction). The authors then focus on the diagnostic stage of each system and categorize the methods in terms of the major artificial intelligence methods employed. They conclude by considering an apparent trend in automated fault management systems toward an increased use of semantic knowledge and other approaches to describe network behavior. The emphasis throughout is on the use of information technology for automating fault management, rather than on engineering-level details of any specific subsystem.
展开▼
