Stopping times of distributed consensus protocols: a probabilistic analysis

Given a model where each processor remains correct for an exponentially distributed random time and then fails independently of the others, we characterize system executions that permit the processors to reach consensus. We show that with non-zero probability, a protocol can achieve consensus even d...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
1. Verfasser: Babaoğlu, Özalp (VerfasserIn)
Format: Buch
Sprache:English
Veröffentlicht: Ithaca, New York 1986
Schriftenreihe:Cornell University <Ithaca, NY> / Department of Computer Science: Technical report 756
Schlagworte:
Zusammenfassung:Given a model where each processor remains correct for an exponentially distributed random time and then fails independently of the others, we characterize system executions that permit the processors to reach consensus. We show that with non-zero probability, a protocol can achieve consensus even during executions where the number of actual processors to fail exceeds its resiliency.
Beschreibung:11 S.