| 摘要: |
| 提出了一种多周期检查点设置方法.它允许各个进程采用不同周期进行检查点设置.为了保证一致全局检查点的向前推进,检查点周期可以根据一个P模式进行调整.在所提出的方法中,进程可以进行组划分处理,从而用于检查点周期调整的依赖跟踪可被限定在组内,同时也将使基于时间的多周期检查点设置具有较好的可扩展性. |
| 关键词: 容错 检查点 依赖跟踪 |
| DOI： |
| 分类号: |
| 基金项目:Supported by the National High-Tech Research and Development Plan of China under Grant Nos.2008AA01A204, 2009AA01A404 (国家高技术研究发展计划(863)); the State Key Laboratory of High-End Server & Storage Technology of China under Grant No.2009HSSA07 (高效能服务器和存储技术国家重点实验室开放基金项目) |
|
| Scalable Time-Based Multi-Cycle Checkpointing |
|
CI Yi-Wei,ZHANG Zhan,ZUO De-Cheng,WU Zhi-Bo,YANG Xiao-Zong
|
| Abstract: |
| In this paper, a time-based multi-cycle checkpointing approach, allowing each process to take checkpoints with its own checkpoint cycle, is proposed. To ensure the advancement of consistent global checkpoint, checkpoint cycles can be adjusted according to a “P-pattern”. In the proposed approach, processes will be divided into zones, so that dependency tracking required for checkpoint cycle adjustment can be restricted in the zone scope. It makes the time-based multi-cycle checkpointing more scalable. |
| Key words: fault-tolerance checkpoint dependency tracking |
