ScaleLFS: A Log-Structured File System with Scalable Garbage Collection for Commodity///

on YouTube ↗16m

" summary

ScaleLFS implements per-core dedicated GC, a scalable victim manager using atomic bitmaps and loose-sync counters, and a scalable victim protector that enables page-level GC with per-file concurrent hash tables. It is prototyped in the Linux kernel on F2FS and evaluated on an Intel Xeon machine with a Samsung 983 SSD. Results show up to 7.0× faster sustained execution time versus prior LFS designs, up to 19.6× higher submitted SSD bandwidth, 83.1% throughput gain on Filebench, and 70.4% faster MySQL (YCSB) writes.

AI-written summary · about

" chapters (8)

1. 0:00ScaleLFS introduction
2. 2:22Scalable GC obstacles
3. 4:08ScaleLFS design components
4. 8:48Data consistency in ScaleLFS
5. 9:30Victim protector and page GC
6. 10:55Evaluation setup results
7. 13:02Macrobenchmarking workloads
8. 15:16Conclusion