Distributed Checkpointing and Rollback

概要

論文の詳細を見る
In distributed systems, transaction processing is an essential part. While processing the transaction on multiple database systems, the transaction suffers from various failure. In order to minimize the recovery time, the checkpointing is required. In the distributed system, we have to take a consistent global checkpoint among multiple processes. Also, it is very expensive to take the global checkpoint due to the communication overhead. In this paper, we present checkpoint and rollback algorithm by using reliable broadcast communication. In our algorithms, a minimal number of processes take checkpoints and rollback to their checkpoints.
一般社団法人情報処理学会の論文
1990-03-14