Issue No.02 - February (2001 vol.12)
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/71.910870
<p><b>Abstract</b>—While <it>Total Order Broadcast</it> (or <it>Atomic Broadcast</it> ) primitives have received a lot of attention, this paper concentrates on <it>Total Order Multicast to Multiple Groups</it> in the context of asynchronous distributed systems in which processes may suffer crash failures. “<it>Multicast to Multiple Groups</it>” means that each message is sent to a subset of the process groups composing the system, distinct messages possibly having distinct destination groups. “<it>Total Order</it>” means that all message deliveries must be totally ordered. This paper investigates a consensus-based approach to solve this problem and proposes a corresponding protocol to implement this multicast primitive. This protocol is based on two underlying building blocks, namely, <it>Uniform Reliable Multicast</it> and <it>Uniform Consensus</it>. Its design characteristics lie in the two following properties: The first one is a <it>Minimality</it> property, more precisely, only the sender of a message and processes of its destination groups have to participate in the total order multicast of the message. The second property is a <it>Locality</it> property: No execution of a consensus has to involve processes belonging to distinct groups (i.e., consensus is executed on a “per group” basis). This <it>Locality</it> property is particularly useful when one is interested in using the Total Order Multicast primitive in large-scale distributed systems. In addition to a correctness proof, an improvement that reduces the cost of the protocol is also suggested.</p>
Asynchronous systems, consensus, groups, reliable multicast, total order, group multicast.
Philippe Ingels, Achour Mostefaoui, Michel Raynal, "Consensus-Based Fault-Tolerant Total Order Multicast", IEEE Transactions on Parallel & Distributed Systems, vol.12, no. 2, pp. 147-156, February 2001, doi:10.1109/71.910870