File servers for network-based distributed systems
ACM Computing Surveys (CSUR)
Reliable object storage to support atomic actions
Proceedings of the tenth ACM symposium on Operating systems principles
Implementing atomic actions on decentralized data
ACM Transactions on Computer Systems (TOCS)
A reliable object-oriented data repository for a distributed computer system
SOSP '81 Proceedings of the eighth ACM symposium on Operating systems principles
Publishing: a reliable broadcast communication mechanism
SOSP '83 Proceedings of the ninth ACM symposium on Operating systems principles
Hi-index | 0.00 |
This thesis presents the design of a set of recovery mechanisms for the Swallow repository. Swallow is a distributed data storage system that supports highly reliable long term storage of arbitrary sized data objects with special mechanisms for implementing multi-site atomic actions. The Swallow repository is a data storage server that keeps permanent data in write-once stable storage such as optical disk. The recovery mechanisms provide on-line recovery for the repository''s internal data, as the repository proceeds with its normal operations. In this way users that wish to access any data that was not affected by the crash can do so while the damaged data is being recovered. Included in the repository''s recovery mechanisms are recovery epochs and checkpoint epochs, which facilitate the detection of damage to the data and minimize the amount of recovery that is necessary. Also included are specialized hash table algorithms that are immune to repository failures. In addition to describing these mechanisms, this thesis discusses how they support the global recovery mechanisms of Swallow and analyzes how they will affect the repository''s general performance.