An Efficient Δ-Causal Distributed Algorithm for Synchronous Cooperative Systems in Unreliable Networks

Abstract

Abstract. In cooperative systems causal ordering delivery has been used to resolve problems of coherency of type producer-consumer. Causal order delivery is important for distributed systems since it allows an asynchronous execution to participants. When time delivery constraints are considered, ensuring causal delivery becomes more complex, as is the case for synchronous cooperative systems, such as Telemedicine and Teleimmersion. In these systems, the messages (units of data of continuous and discrete media) have an associated lifetime that determines the period of useful time in which the messages must be delivered. On the other hand, generally in these systems there is no time for retransmit them when messages are lost. Causal order with time constraints has previously been addressed, and it is called causal−Δ order. In this paper, we present an efficient causal−Δ distributed algorithm for unreliable networks that is characterized by the use of a forward error correction (FEC) scheme and a distributed method to calculate the message lifetime based on relative time points (i.e. no global time is used). We show the efficiency of our causal−Δ algorithm in terms of the control information attached per message.