This paper describes distributed fault-tolerant(FT) and auto-healing(AH) algorithms for dual-ring netowrks. The algorithms are based on the use of intercommunications and hand-shaking processes of adjacent nodes to exchange information in case of network faults. Some of their remarkable features are: (1) allow fast network restoration, (2) use short restoration message (2-bytes) to alleviate the network loading during restoration, (3) maintain high network availability to users during restoration and (4) enable hot replacement of faulty network components. The validity of the algorithms was tested on a high-speed multimedia network prototype designed at the Chinese University of Hong Kong - CUMLAUDE NET~1. Failure recovery time on the order of hundreds of micro-seconds is achieved. Finally, measures including communicative probability, survivability and average reachability are used to quantify the network reliability under the algorithms.
展开▼
