Work-in-Progress Reports (WiPs)

on YouTube ↗31m

" summary

Seven short systems results: an application‑local user‑space library filesystem for persistent memory shares VM pages, omits ACLs, and in single‑process multiversion mode beats kernel PM filesystems on small IOs; automated I/O tuning captures traces with blktrace and replays with fio‑style jobs to pick parallelism and request size; WAL work uses ioprio+BFQ/CFQ, direct I/O and thread‑level priorities to regain near‑dedicated WAL latency; SSP shadow paging boosts ext4 throughput by 52% at 40 threads; RAID2.0++ cuts reconstruction read imbalance (λ_read +41%, λ_write +74%); update‑friendly erasure coding lowers update I/O by 77% and cross‑rack traffic by 76%; smart storage nodes prototype FPGA offloads implementing randomized 3D rotations and differential privacy for low‑latency enforcement.

AI-written summary · about

" chapters (8)

1. 0:00Start
2. 0:16Keiichi Matsuzawa, Hitachi, Ltd.; Takahiro Shinagawa, The University of Tokyo
3. 3:29Azat Nurgaliev and Marcus Paradies, German Aerospace Center
4. 7:17Lalitha Donga, Ben Reed, and Kayla Walton, San Jose State University
5. 11:51Joontaek Oh, Hojin Nam, Kyoungho Koo, and Youjip Won, KAIST
6. 16:40Qiliang Li, Yinlong Xu, and Min Lyu, University of Science and Technology of China
7. 20:49Wei Wang, Min Lyu, and Yinlong Xu, University of Science and Technology of China
8. 25:20Zsolt István, IT University, Copenhagen