Reproduce “DiaLSM: Towards Write-Stall-Free Performance via Shard-based LSM-tree” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Hongsu Byun, Safdar Jamil, Honghyeon Yoo, Sungyong Park, Myungcheol Lee, Xubin He, Zhichao Cao, Youngjae Kim
DiaLSM reduces write stalls in RocksDB by splitting the LSM-tree into shards and redirecting writes from stalled primary shards to non-stalled fallbacks. Clue entries help reads locate redirected values, but add work. On 8 KB db_bench FillRandom, DiaLSM reports 94.3% fewer stalls, up to 2.4× throughput, and 50% lower latency than ADOC; these benchmark results do not preclude all shards stalling together.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.