Accession Number : ADA310228
Title : Using Time to Improve the Performance of Coordinated Checkpointing,
Corporate Author : ILLINOIS UNIV AT URBANA
Personal Author(s) : Neves, Nuno ; Fuchs, W. K.
PDF Url : ADA310228
Report Date : 1996
Pagination or Media Count : 11
Abstract : This paper describes and evaluates a coordinated checkpoint protocol that uses time to eliminate several peformance overheads that are present in traditional protocols. The time-based protocol does not have to exchange coordination messages, does not need to add information to the processes' messages, and only accesses stable storage when checkpoints are saved. This protocol uses a simple initialization procedure to set checkpoint timers at the dlfferent processes. After the initialization, each process saves its state independently from the other processes. By disallowing processes from sending messages during an interval before the checkpoint time, the protocol prevents in-transit messages from occurring. Two coordinated checkpoint protocols were implemented on a CM5, and their performance was compared using several applications. Results showed that the time-based protocol outperforms the two-phase protocol in all applications.
Descriptors : *DISTRIBUTED DATA PROCESSING, *VALIDATION, *MESSAGE PROCESSING, STABILITY, DATA STORAGE SYSTEMS, STORAGE, INTERVALS.
Subject Categories : Computer Systems
Distribution Statement : APPROVED FOR PUBLIC RELEASE