Issue No. 02 - April (1996 vol. 16)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/40.491463
Fault-tolerant group communication is highly desirable for maintaining a consistent state in distributed systems. Many existing protocols are quite expensive and of limited benefit for distributed systems in terms of efficiency. This article presents a novel group communication protocol including dynamic assessment of membership. It achieves fault tolerant group communication based on a logical token ring approach. The protocol ensures total ordering of messages and atomicity of delivery in the presence of communication faults and site faults, guaranteeing that all operational members belonging to the same group observe a consistent view of ordered events. The dynamic membership and failure recovery algorithms can handle site faults and restart, group partitions and merges, dynamic members join and leave. The efficiency of the protocol is shown with some performance figures.
Fault tolerance, distributed systems, communications, token rings
J. Kaiser, E. Nett and W. Jia, "RMP: Fault-Tolerant Group Communication," in IEEE Micro, vol. 16, no. , pp. 59-67, 1996.