Causal order multicast protocol using minimal message history information

Abstract

Many applications of cloud-based P/S systems require communication services providing reliable message delivery and high scalability with flexible consistency simultaneously. Gossip communication is becoming one of the promising solutions for addressing P/S scalability problems in providing information propagation functionality by exploiting a mixture of diverse consistency options in the cloud-based P/S. In particular, despite the importance of both guaranteeing message delivery order required and supporting managing performance in cloud-based P/S systems, there exist little research works on development of gossip-style dissemination protocols to satisfy all these requirements. In this paper, we present a causal message order protocol with the information of only the predecessors immediately before the messages to subscribers from their chosen brokers. In the protocol, brokers might guarantee consistently causally ordered message delivery among themselves by aggregating the information of all ancestors before the messages. On the other hand, only the information of the predecessors immediately before the messages piggybacked on each multicast message is transmitted from brokers to subscribers through gossip-style disseminations. Its feature might be highly scalable and suitable for the area of the applications requiring only the minimum causal information of message delivery with flexible consistency. © 2012 Springer-Verlag.