Multi-Cluster Adaptive Distributed System-Level Diagnosis Algorithms

概要

論文の詳細を見る
System-level diagnosis is used to determine the state of a system consisting of N units which may be faulty or fault-free. Each unit tests a subset of all others, and fault-free units perform tests and report test results reliably. In adaptive diagnosis the tests each unit performs are adaptively selected based on previous test results; and in distributed diagnosis the units themselves perform the diagnosis, instead of a centralized observer. In this paper we present new adaptive distributed system-level diagnosis algorithms, that by grouping units in logical clusters improve the diagnosis latency of current algorithms, while still requiring the same order of diagnostic messages. Two algorithms, ADSD with Intersections and Hierarchical ADSD are presented and analyzed. Applications of these algorithms, including network fault management are considered.
社団法人電子情報通信学会の論文
1995-12-12