Reliable Broadcast in the Presence of Process Crash Failures

Abstract

This chapter focuses on the uniform reliable broadcast (URB) communication abstraction and its implementation in an asynchronous message-passing system prone to process crashes. This communication abstraction is central in the design and implementation of fault-tolerant distributed systems, as many non-trivial fault-tolerant distributed applications require communication with provable guarantees on message deliveries. After defining the URB abstraction, the chapter presents a construction of it in an asynchronous message passing system prone to process crashes but with reliable channels (i.e., in the system model CAMP n,t [∅]). The chapter then considers two properties (related to the quality of service) that can be added to URB without requiring enrichment of the system model with additional assumptions. These properties concern the message delivery order, namely "first in first out" (FIFO) message delivery and "causal order" (CO) message delivery. The broadcast operation "broadcast (m)", introduced in the previous chapter, was a simple macro-operation which expands in the statement for each j ∈ {1,. .. , n} do send m to p j end for. In the system model CAMP n,t [∅], this operation has best effort semantics in the following sense. If the sender p i is correct, a copy of the message m is sent to every process, and, as the channels are reliable, every process (that has not crashed) receives a copy of the message. As the channels are asynchronous, these copies can be received at distinct independent time instants. Whereas if the sender crashes while executing broadcast m, an arbitrary subset of the processes receives the message m. Hence, in the presence of process crash failures, the specification of "broadcast m" provides no indication which processes will actually receive the message m. The aim of this section is to introduce a broadcast operation that provides the processes with stronger message delivery guarantees. © Springer Nature Switzerland AG 2018 M. Raynal, Fault-Tolerant Message-Passing Distributed Systems, https://doi.