Using Time Instead of Timeout for Fault-Tolerant Distributed Systems.
ACM Transactions on Programming Languages and Systems (TOPLAS)
Checkpointing and Rollback-Recovery for Distributed Systems
IEEE Transactions on Software Engineering - Special issue on distributed systems
Manetho: Transparent Roll Back-Recovery with Low Overhead, Limited Rollback, and Fast Output Commit
IEEE Transactions on Computers - Special issue on fault-tolerant computing
A survey of rollback-recovery protocols in message-passing systems
ACM Computing Surveys (CSUR)
Garbage collection in message passing distributed systems
PAS '95 Proceedings of the First Aizu International Symposium on Parallel Algorithms/Architecture Synthesis
Sender-based message logging for reducing rollback propagation
SPDP '95 Proceedings of the 7th IEEE Symposium on Parallel and Distributeed Processing
A low-overhead recovery technique using quasi-synchronous checkpointing
ICDCS '96 Proceedings of the 16th International Conference on Distributed Computing Systems (ICDCS '96)
Completely Asynchronous Optimistic Recovery with Minimal Rollbacks
FTCS '95 Proceedings of the Twenty-Fifth International Symposium on Fault-Tolerant Computing
Message logging: pessimistic, optimistic, and causal
ICDCS '95 Proceedings of the 15th International Conference on Distributed Computing Systems
| Hi-index | 0.00 |
This paper presents a garbage collection protocol for message content logs and message determinant logs which are saved on a stable storage. Previous works of garbage collections in a causal message logging protocol try to solve the garbage collection of message determinant log and force additional checkpoints[5,6,7]. In order to avoid the sympathetic rollback, we classify the fault tolerance information into message determinants logs and message contents logs. Then we propose new definitions for garbage collections conditions for message determinant logs and message content logs and present a garbage collection algorithm. To represent determinants of messages, a data structure called MAG (Modified Antecedence Graph) is proposed. MAG is an extension of Antecedence Graph of Manetho system [7] and it is used for garbage collections conditions of message determinant logs and message content logs. Unlike Manetho system that needs additional messages for garbage collection of message content logs, our algorithm does not need additional messages. The proposed garbage collection algorithm makes ’the lazy garbage collection effect’ because it relies on the piggybacked checkpoint information in send/receive message. ‘The lazy garbage collection effect’ provides the whole system with an efficient and simple recovery protocol.