Lazy Checkpointing: Exploiting Temporal Locality in Failures to Mitigate Checkpointing Overheads on Extreme-Scale Systems
Lazy Checkpointing: Exploiting Temporal Locality in Failures to Mitigate Checkpointing Overheads on Extreme-Scale Systems
复制标题
DOI：
10.1109/dsn.2014.101
复制
发表时间：
2014-06
期刊：
影响因子：
--
通讯作者：
Devesh Tiwari;Saurabh Gupta;Sudharshan S. Vazhkudai
中科院分区：
文献类型：
--
作者：
Devesh Tiwari;Saurabh Gupta;Sudharshan S. Vazhkudai
超级计算机计算能力的不断提高，使得天体物理学、核聚变、气候和燃烧等领域的大规模科学应用能够运行更大规模、更长期的模拟，从而促进了更深入的科学见解。然而，这些长时间运行的模拟经常被多个系统故障中断。因此，这些应用程序依赖于“检查点”作为一种弹性机制，将应用程序状态存储到永久存储并从故障中恢复。不幸的是，检查点会在超级计算机上产生过多的I/O开销，因为检查点的大小很大，从而导致性能和资源利用率不理想。在本文中，我们设计了新的机制来展示如何通过利用系统故障的时间特征来显著减轻检查指向开销。我们提供新的见解和详细的定量理解检查指向开销和大型机器的权衡。我们的原型实现显示了我们的方法在极端规模机器上的可行性。
Continuing increase in the computational power of supercomputers has enabled large-scale scientific applications in the areas of astrophysics, fusion, climate and combustion to run larger and longer-running simulations, facilitating deeper scientific insights. However, these long-running simulations are often interrupted by multiple system failures. Therefore, these applications rely on "check pointing'" as a resilience mechanism to store application state to permanent storage and recover from failures. Unfortunately, check pointing incurs excessive I/O overhead on supercomputers due to large size of checkpoints, resulting in a sub-optimal performance and resource utilization. In this paper, we devise novel mechanisms to show how check pointing overhead can be mitigated significantly by exploiting the temporal characteristics of system failures. We provide new insights and detailed quantitative understanding of the check pointing overheads and trade-offs on large-scale machines. Our prototype implementation shows the viability of our approach on extreme-scale machines.