An important problem in the construction of fault-tolerant distributed database systems is the design of nonblocking transaction commit protocols. This problem has been extensively studied for synchronous systems (i.e., systems where no messages ever arrive late). In this paper, the synchrony assump...

Ausführliche Beschreibung

Bibliographische Detailangaben

Hauptverfasser:	Coan, Brian A. (VerfasserIn), Welch, Jennifer Lundelius (VerfasserIn)
Format:	Buch
Sprache:	English
Veröffentlicht:	Cambridge, Mass. Mass. Inst. of Technology, Laboratory for Computer Science 1988
Schlagworte:	Computer Hardware Data bases Degradation Distributed data processing Fault tolerant computing Models Processing equipment Synchronism Systems engineering Time Verteiltes Datenbanksystem Fehlertoleranz
Zusammenfassung:	An important problem in the construction of fault-tolerant distributed database systems is the design of nonblocking transaction commit protocols. This problem has been extensively studied for synchronous systems (i.e., systems where no messages ever arrive late). In this paper, the synchrony assumption is relaxed. A new partially synchronous timing model is given. In this model, a new nonblocking randomized transaction commit protocol is given, based on a Byzantine agreement protocol of Ben-Or. The new protocol works as long as fewer than half the processor fail. A lower bound is proved, showing that the number of processor faults tolerated is optimal. The protocol exhibits a graceful degradation property: when more than half the processors fail, the protocol blocks, but not processor produces a wrong answer. A notion of asynchronous round is defined and the protocol is shown to terminate in a small constant expected number of asynchronous rounds. The final result is that no protocol in this model can terminate in a bounded expected number of steps, even if processors are synchronous. (KR).
Beschreibung:	38 S.