L. Rodrigues and P. Veríssimo
From INESC Journal of R & D, 1(1), June 1990.
The problem of reaching distributed agreement among processors has deserved an increasing interest in the last years. Recent developments in distributed computing systems have shown that the implementation of reliable communication protocols satisfying a set of desirable requirements, such as the achievement of interactive consistency, may simplify the design of higher level distributed algorithms.
The description of many protocols which solve this problem, offering different services under diverse fault assumptions and communication media, can be found in the literature. These are generally called reliable broadcast protocols, and this paper presents a survey of different approaches to the problem.
Special attention is given to the relationship between the services provided by the protocols, the assumptions made in their design and their cost both in time and traffic. The relation with other closely related subjects, such as group management protocols and failure handling mechanisms, is analyzed.
Not available in postscript.