Keigo: Co-designing Log-Structured Merge Key-Value Stores with a Non-Volatile, Concurrency-aware Storage Hierarchy
Summary: Keigo: a concurrency- and workload-aware middleware that co-designs LSM KVs on heterogeneous non-volatile storage, placing files by device parallelism, bandwidth and capacity instead of fixed tiering. Combines concurrency-aware placement, persistent read-only caching and context-based I/O differentiation; portable to RocksDB/LevelDB/Speedb and yields up to 4× write and 18× read throughput improvements. (summarized by gpt-5-mini on Feb 09 2026)
Incoming Non-self Citations Over Time
No non-self incoming citations found for this paper in this database.
Authors
- 1. R uben Ad o
- 2. Zhongjie Wu
- 3. Changjun Zhou
- 4. Oana Balmau
- 5. Jo o Paulo
- 6. Ricardo Macedo
Incoming Citations (Sorted by Pagerank)
Showing 0 of 0 citing papers.
| Rank | Citing Paper | Year | Venue | Pagerank |
|---|
Previous
Page 1 / 1
Next
Outgoing Citations (Sorted by Pagerank)
Showing 5 of 5 cited papers.
Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.
| Rank | Cited Paper | Year | Venue | Pagerank |
|---|---|---|---|---|
| 569 | Optimizing Space Amplification in RocksDB | 2017 | CIDR | 0.00019924098 |
| 1,610 | MyRocks: LSM-Tree Database Storage Engine Serving Facebook's Social Graph | 2020 | VLDB | 0.00011148094 |
| 1,613 | Realtime Data Processing at Facebook | 2016 | SIGMOD | 0.00011140777 |
| 4,074 | Understanding the Idiosyncrasies of Real Persistent Memory | 2021 | VLDB | 6.4717302e-05 |
| 7,043 | NVM: Is it Not Very Meaningful for Databases? | 2023 | VLDB | 4.8539749e-05 |
Previous
Page 1 / 1
Next