EUVIMEDEUVIMED entdecken
Crossref

Turbocharging Vector Databases Using Modern SSDs

Joobo Shim, Jaewon Oh, Hongchan Roh, Jaeyoung Do, Sang-Won Lee

Proceedings of the VLDB Endowment · 2025-07 · 18 (11) · 4710-4722

DOI: 10.14778/3749646.3749724 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the VLDB Endowment, online erschienen am 04.09.2025. Verfasst von Joobo Shim und 4 weiteren Autor:innen. Die Arbeit wurde bisher 11-mal zitiert (Crossref) und nennt 48 Literaturangaben.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Efficient and scalable vector search is critical for modern AI applications, particularly in retrieval-augmented generation (RAG) and large-scale semantic search. However, disk-based vector databases often suffer from significant I/O bottlenecks due to suboptimal cache hit ratios and inefficient use of modern SSD architectures. In this work, we introduce a suite of optimizations to enhance the performance of disk-resident Approximate Nearest Neighbor (ANN) indices, specifically focusing on hierarchical graph-based indexing such as HNSW. Our approach leverages three key strategies: (1) Parallel I/O leveraging io_uring to exploit SSD concurrency and reduce retrieval latency, (2) Spatially-aware insertion reordering to improve cache efficiency by dynamically adjusting insert execution order based on locality, and (3) Locality-preserving colocation to restructure index layouts and minimize costly random disk accesses. We implement these techniques within pgvector, a PostgreSQL extension for vector search, and conduct extensive evaluations using real-world datasets. Our optimizations yield up to 11.1× improvement in query throughput, a 3.23× increase in cache hit ratio, and a 98.4% reduction in index build time. Moreover, our findings underscore the importance of SSD-aware indexing strategies for scalable vector retrieval. By integrating hardware-aware I/O optimizations with intelligent data placement techniques, this work paves the way for more efficient, high-performance disk-based vector search engines that could fully leverage modern SSD's high parallelism.

Zitieren

APA

Shim, J., Oh, J., Roh, H., Do, J., & Lee, S. (2025). Turbocharging Vector Databases Using Modern SSDs. Proceedings of the VLDB Endowment, 18(11), 4710-4722. https://doi.org/10.14778/3749646.3749724

Vancouver

Shim J, Oh J, Roh H, Do J, Lee S. Turbocharging Vector Databases Using Modern SSDs. Proceedings of the VLDB Endowment. 2025;18(11):4710-4722. doi:10.14778/3749646.3749724

BibTeX
@article{Shim2025,
  author = {Shim, Joobo and Oh, Jaewon and Roh, Hongchan and Do, Jaeyoung and Lee, Sang-Won},
  title = {Turbocharging Vector Databases Using Modern SSDs},
  journal = {Proceedings of the VLDB Endowment},
  year = {2025},
  volume = {18},
  number = {11},
  pages = {4710--4722},
  doi = {10.14778/3749646.3749724}
}
Kostenlos starten →