Near-Data Compaction for LSM Tree on Rack-Scale Disaggregated Storage
IEEE Transactions on Parallel and Distributed Systems · 2026-08 · 37 (8) · 1758-1771
DOI: 10.1109/tpds.2026.3696772 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Parallel and Distributed Systems, 2026-08. Verfasst von Sungho Moon und 6 weiteren Autor:innen. Die Arbeit nennt 45 Literaturangaben. Gefördert von: Samsung, National Research Foundation of Korea, Institute of Information.
Abstract (Originalsprache)
In LSM trees, background compaction tasks contend with foreground queries for CPU cycles, cache space, and also SAN bandwidth, if deployed in a disaggregated storage architecture. This study proposes Near-Data Compaction (NDC), which executes compaction on the storage node to utilize its underutilized computing resources. However, enabling NDC introduces several challenges. First, it has to support concurrent file access from both compute and storage nodes. In addition, it has to decide which compaction tasks to be executed on the storage node since the computing resources of a storage node are not unlimited. This study presents TetherDB, an LSM tree for disaggregated storage architecture that addresses these challenges through lightweight dual-node coordination and selective NDC admission policies. Our evaluation demonstrates that TetherDB improves throughput by up to 2.1× compared to RocksDB in write-heavy workloads.
Förderung
- Samsung
- National Research Foundation of Korea
- Institute of Information
- Communications Technology Planning
- Korean Government
Zitieren
APA
Moon, S., Han, D., Koo, H., Chae, S., Bae, D., Seo, E., & Nam, B. (2026). Near-Data Compaction for LSM Tree on Rack-Scale Disaggregated Storage. IEEE Transactions on Parallel and Distributed Systems, 37(8), 1758-1771. https://doi.org/10.1109/tpds.2026.3696772
Vancouver
Moon S, Han D, Koo H, Chae S, Bae D, Seo E, et al. Near-Data Compaction for LSM Tree on Rack-Scale Disaggregated Storage. IEEE Transactions on Parallel and Distributed Systems. 2026;37(8):1758-1771. doi:10.1109/tpds.2026.3696772
BibTeX
@article{Moon2026,
author = {Moon, Sungho and Han, Daegyu and Koo, Hera and Chae, Sangeun and Bae, Duck-Ho and Seo, Euiseong and Nam, Beomseok},
title = {Near-Data Compaction for LSM Tree on Rack-Scale Disaggregated Storage},
journal = {IEEE Transactions on Parallel and Distributed Systems},
year = {2026},
volume = {37},
number = {8},
pages = {1758--1771},
doi = {10.1109/tpds.2026.3696772}
}
EUVIMED