Читать книгу From Traditional Fault Tolerance to Blockchain - Wenbing Zhao - Страница 44

2.2.1 Uncoordinated Checkpointing

Оглавление

Uncoordinated checkpointing, where each process in the distributed system enjoys full autonomy and can decide when to checkpoints, even though appears to be attractive, is not recommended for two primary reasons.

First, the checkpoints taken by the processes might not be useful to reconstruct a consistent global state. In the worst case, the system might have to do a cascading rollback to the initial system state (often referred to as the domino effect [16]), which completely defeats the purpose of doing checkpointing. Consider the following example.

From Traditional Fault Tolerance to Blockchain

Подняться наверх