Space-efficient page-level incremental checkpointing

Abstract

Incremental checkpointing, which is intended to minimize checkpointing overhead, saves only the modified pages of a process. However, the cumulative size of incremental checkpoints increases at a steady rate over time because many updated values may be saved for the same page. In this paper, we present a comprehensive overview of Pickpt, which is a page-level incremental checkpointing facility. Pickpt provides space-efficient techniques for minimizing the use of disk space. For our experiments, the results show that the use of disk space of Pickpt was significantly reduced compared with existing incremental checkpointing.

Understanding the Impact of BPRAM on Incremental Checkpoint

“…One is page-protection-based; the other is hash-based. Pageprotection-based technique requires hardware and operating system supports to identify dirty pages [4], [6], [10], [11]. The main drawback of page-protection-based approaches is that they fix the granularity to be a page, which reduces the flexibility of checkpoint implementation.…”

Hybrid Checkpointing for MPI Jobs in HPC Environments

“…Incremental Checkpointing: Recent studies focus on incremental checkpointing [20], [21], [22]. TICK (Transparent Incremental Checkpointer at Kernel Level) [20] is a systemlevel checkpointer implemented as a kernel thread.…”

“…In contrast, our solution transparently supports hybrid full/incremental checkpoints for an MPI job, which is unprecedented. Pickpt [21] is a page-level incremental checkpointing facility. It provides space-efficient techniques for automatically removing useless checkpoints aiming to minimizing the use of disk space that differ from our garbage collection thread technique.…”

Assessing HPC Failure Detectors for MPI Jobs

“…Our resulting implementation can easily be combined with reactive checkpoint/restart frameworks to trigger restarts after components have failed [2,[5][6][7][8][9][10][12][13][14]17,[17][18][19][20]22,23,26,27,29,[32][33][34][35][36].…”