The authors define the topology and the protocols of a fault-tolerant multichannel local area network. Specifically, a broadcast bus with passive taps, in which the channel as well as the stations can fail. A channel fails when it is severed; a station can fail in many different modes. In particular, a station may get stuck on transmit (i.e. jabbering station) thus preventing other stations from using the bus. To overcome the above failure situation and identify the faulty station, the authors propose to connect the stations with several parallel buses. These channels are interconnected by bridges, which in case of failure can isolate the faulty buses from the rest of the network, and provide connectivity among the healthy stations. The authors address the issue of optimal connection of stations to the buses. Distributed algorithms for failure identification and recovery are developed. A simple analytical model shows performance degradation caused by failures
Fault-Tolerance in Multichannel Local Area Networks / Camarda, P.; Gerla, M.. - STAMPA. - (1989), pp. 133-137. (Intervento presentato al convegno 8th Annual International Phoenix Conference on Computers and Communications tenutosi a Scottsdale, AZ nel March 22-24, 1989) [10.1109/PCCC.1989.37374].
Fault-Tolerance in Multichannel Local Area Networks
P. Camarda;
1989-01-01
Abstract
The authors define the topology and the protocols of a fault-tolerant multichannel local area network. Specifically, a broadcast bus with passive taps, in which the channel as well as the stations can fail. A channel fails when it is severed; a station can fail in many different modes. In particular, a station may get stuck on transmit (i.e. jabbering station) thus preventing other stations from using the bus. To overcome the above failure situation and identify the faulty station, the authors propose to connect the stations with several parallel buses. These channels are interconnected by bridges, which in case of failure can isolate the faulty buses from the rest of the network, and provide connectivity among the healthy stations. The authors address the issue of optimal connection of stations to the buses. Distributed algorithms for failure identification and recovery are developed. A simple analytical model shows performance degradation caused by failuresI documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.