Abstract
Agreement problems are at the heart of the design of dependable and reliable distributed services. Distributed systems that run such services may experience unpredictable processing and communication delays, and some of their components can fail in various ways. It has been proved that in such settings, the consensus problem, the most popular and fundamental of the agreement problems has no deterministic solution.
Therefore, researchers started investigating ways of circumventing the impossibility result. Two main directions were explored: relaxing the requirements of the consensus problem, and strengthening the assumptions on the system.At least two ways of relaxing the consensus requirements have been investigated: randomization (termination is achieved only with high probability) and approximate agreement. Also, at least two ways of strengthening the assumptions on the system have been considered: adding synchrony assumptions to the system and abstracting the details of how a processor suspects a failure has occurred, without referring to particular synchrony assumptions by the mean of the Unreliable Failure Detectors that provides processes with a list of processes suspected to have crashed.
Chapter PDF
Similar content being viewed by others
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2009 Springer-Verlag Berlin Heidelberg
About this paper
Cite this paper
Mostefaoui, A. (2009). What Agreement Problems Owe Michel. In: Keidar, I. (eds) Distributed Computing. DISC 2009. Lecture Notes in Computer Science, vol 5805. Springer, Berlin, Heidelberg. https://doi.org/10.1007/978-3-642-04355-0_3
Download citation
DOI: https://doi.org/10.1007/978-3-642-04355-0_3
Publisher Name: Springer, Berlin, Heidelberg
Print ISBN: 978-3-642-04354-3
Online ISBN: 978-3-642-04355-0
eBook Packages: Computer ScienceComputer Science (R0)