Z-Journal: Scalable Per-Core Journaling
summary
Z-Journal implements per-core journal stacks in the Linux kernel (replacing JBD2) and modifies ext4 to record a system-wide transaction order graph via order-preserving transaction chaining so cores can commit independently while checkpoint enforces cross-core write order. The prototype is evaluated on an 80-core server with FxMark, SysBench and Filebench and claims large gains—examples in the paper include ~315% at 20 cores, ~58% at 80 cores, and an overall reported speedup up to ≈4000%—and the design works end-to-end.
AI-written summary · about
chapters (16)
- 0:00Intro
- 0:23Conventional Journaling Scheme
- 1:12Journal - A Scalability Bottleneck
- 1:50Related work
- 2:48Intuitive Solution
- 3:04Journal Coherence Problem
- 4:12Our Approach: Z-Journal
- 4:34Order-preserving Transaction Chaining
- 5:38Construction of Transaction Chain Graphs
- 7:29Conventional Checkpoint Process
- 8:14Journal Coherence Checkpoint
- 9:01Transaction Chain Graph Traversing
- 11:11Evaluation
- 11:51Under Various Sharing Conditions
- 13:01Overall File System Performance
- 14:24Z-Journal Conclusion
