Home > Research > Publications & Outputs > Roll-forward recovery

Links

Text available via DOI:

View graph of relations

Roll-forward recovery: the bidirectional cache approach

Research output: Contribution in Book/Report/Proceedings - With ISBN/ISSNConference contribution/Paperpeer-review

Published
Close
Publication date12/06/1996
Host publicationProceedings of IEEE Workshop on Fault-Tolerant Parallel and Distributed Systems
PublisherIEEE
Pages59-68
Number of pages10
ISBN (print)0818668075
<mark>Original language</mark>English
EventIEEE Workshop on Fault-Tolerant Parallel and Distributed Systems - College Station, United States
Duration: 12/06/199614/06/1996

Conference

ConferenceIEEE Workshop on Fault-Tolerant Parallel and Distributed Systems
Country/TerritoryUnited States
CityCollege Station
Period12/06/9614/06/96

Conference

ConferenceIEEE Workshop on Fault-Tolerant Parallel and Distributed Systems
Country/TerritoryUnited States
CityCollege Station
Period12/06/9614/06/96

Abstract

A split-cache memory architecture is presented which provides efficient architectural support for checkpointing and roll-forward recovery mechanisms in distributed systems. Unlike existing techniques, the approach does not require the use of a discrete stable storage unit or explicit synchronization among the processors. A nonscheduled checkpointing mechanism is presented based on a cache-line replacement policy, instead of the conventionally used periodic checkpoint establishment protocols.