DumpKV: Learning based lifetime aware garbage collection for key value separation in LSM-tree

Summary: DumpKV uses key-access histories and a lightweight learned model to predict value lifetimes in KV-separated LSM-trees, replacing static GC policies. It collects features during L0–L1 compaction and cuts GC writes by 19–65% with low overhead. (summarized by gpt-5.6-luna on Jul 24 2026)

Paper ID: 13980
Venue: VLDB
Year: 2025
Pagerank: 5.1814573e-05
Overall Rank: 9,995 | 31.43%
DOI: 10.14778/3717755.3717778

Incoming Non-self Citations Over Time

Authors

1. Zhutao Zhuang (Microsoft; Sun Yat-Sen University)
2. Xinqi Zeng (Sun Yat-Sen University)
3. Zhiguang Chen (Sun Yat-Sen University)

BibTeX Citation

@article{zhuang_vldb25,
        title = {{DumpKV: Learning based lifetime aware garbage collection for key value separation in LSM-tree}},
        author = {Zhuang, Zhutao and Zeng, Xinqi and Chen, Zhiguang},
        journal = {PVLDB},
        series = {{VLDB} '25},
        volume = {18},
        number = {4},
        pages = {1223--1236},
        doi = {10.14778/3717755.3717778},
        url = {https://doi.org/10.14778/3717755.3717778},
        year = {2025}
}

Incoming Citations (Sorted by Pagerank)

Showing 1 of 1 citing papers.

Rank	Citing Paper	Year	Venue	Pagerank
10,955	DobLIX: A Dual-Objective Learned Index for Log-Structured Merge Trees	2025	VLDB	5.093636e-05

Outgoing Citations (Sorted by Pagerank)

Showing 8 of 8 cited papers.

Citations counted here include only citations to other VLDB/SIGMOD/CIDR/PODS papers in this database.

Rank	Cited Paper	Year	Venue	Pagerank
86	Automatic Database Management System Tuning Through Large-scale Machine Learning	2017	SIGMOD	0.00035316107
271	TiDB: A Raft-based HTAP Database	2020	VLDB	0.00022703024
447	ALEX: An Updatable Adaptive Learned Index	2020	SIGMOD	0.00018322593
461	Query-based Workload Forecasting for Self-Driving Database Management Systems	2018	SIGMOD	0.00018068441
498	QTune: A Query-Aware Database Tuning System with Deep Reinforcement Learning	2019	VLDB	0.00017440583
568	SageDB: A Learned Database System	2019	CIDR	0.0001641553
831	Dostoevsky: Better Space-Time Trade-Offs for LSM-Tree Based Key-Value Stores via Adaptive Removal of Superfluous Merging	2018	SIGMOD	0.00013748512
3,802	Leaper: A Learned Prefetcher for Cache Invalidation in LSM-tree based Storage Engines	2020	VLDB	7.1115502e-05

Semantically Similar Papers

#	Overall Rank	Paper	Year	Venue
1	831	Dostoevsky: Better Space-Time Trade-Offs for LSM-Tree Based Key-Value Stores via Adaptive Removal of Superfluous Merging	2018	SIGMOD
2	8,919	MirrorKV: An Efficient Key-Value Store on Hybrid Cloud Storage with Balanced Performance of Compaction and Querying	2023	SIGMOD
3	7,274	LSM-Trees and B-Trees: The Best of Both Worlds	2019	SIGMOD
4	9,457	Rethinking The Compaction Policies in LSM-trees	2025	SIGMOD
5	857	SlimDB: A Space-Efficient Key-Value Storage Engine For Semi-Sorted Data	2017	VLDB
6	11,908	Splaying Log-Structured Merge-Trees	2018	SIGMOD
7	8,811	ArceKV: Towards Workload-driven LSM-compactions for Key-Value Store Under Dynamic Workloads	2026	VLDB
8	10,616	Terark-DS: A High-Performance and Storage-Efficient Key-Value Separation Storage Engine on Disaggregated Storage	2026	VLDB
9	6,843	Learning to Optimize LSM-trees: Towards A Reinforcement Learning based Key-Value Store for Dynamic Workloads	2023	SIGMOD
10	10,480	PartitionKV: Redesigning LSM-tree KV Stores on NVMs with Adaptive Partitioning for Reducing Write Stalls and Amplification	2026	SIGMOD