SRIndex: Bridging the Last Mile between Learned-Index Scalability and Workload Dynamics for Memory Pooling Systems
ACM Transactions on Storage · 30.12.2026 · 22 (4) · 1-23
DOI: 10.1145/3806051 ↗Kurz gefasst
Fachartikel in der Zeitschrift ACM Transactions on Storage, online erschienen am 14.08.2026. Verfasst von Zhiwei Yang und 8 weiteren Autor:innen. Die Arbeit nennt 46 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0). Gefördert von: The National Key R&D Program of China, The Major Research Plan of the National Natural Science Foundation of China, the National Natural Science Foundation of China.
Abstract (Originalsprache)
As data volumes continue to escalate, conventional index structures struggle to keep pace with high-performance demands. Learned indexing is a data structure that leverages machine learning models to predict the locations of data, which significantly improves search performance compared to traditional indexing methods. However, existing learned index schemes often focus solely on optimizing read and write operations, neglecting scalability and robustness in their index design. This oversight leads to significant performance variations across workloads with different characteristics. In this article, we present SRIndex , a scalable and robust learned index, to address the abovementioned problems. SRIndex consists of the following design primitives: (i) It implements a model layer , which is optimized with a select number of keys to reduce asynchronous retraining time and accelerate query processes; (ii) It incorporates a transition layer to achieve the scalability of the index; and (iii) It introduces a leaf layer to provide efficient storage and high access efficiency. Extensive experimental results demonstrate that SRIndex outperforms state-of-the-art learned indexes by improving 1.27× of write throughput under dynamic workloads.
Förderung
- The National Key R&D Program of China
- The Major Research Plan of the National Natural Science Foundation of China
- the National Natural Science Foundation of China
- The Alibaba PolarDB
Zitieren
APA
Yang, Z., Zhou, S., Shen, Z., Zhou, Y., Wu, R., Shu, J., Cheng, L., Chen, Z., & Kaplun, D. (2026). SRIndex: Bridging the Last Mile between Learned-Index Scalability and Workload Dynamics for Memory Pooling Systems. ACM Transactions on Storage, 22(4), 1-23. https://doi.org/10.1145/3806051
Vancouver
Yang Z, Zhou S, Shen Z, Zhou Y, Wu R, Shu J, et al. SRIndex: Bridging the Last Mile between Learned-Index Scalability and Workload Dynamics for Memory Pooling Systems. ACM Transactions on Storage. 2026;22(4):1-23. doi:10.1145/3806051
BibTeX
@article{Yang2026,
author = {Yang, Zhiwei and Zhou, Shuyue and Shen, Zhirong and Zhou, Yixiu and Wu, Ronglong and Shu, Jiwu and Cheng, Li and Chen, Zheng and Kaplun, Dmitrii},
title = {SRIndex: Bridging the Last Mile between Learned-Index Scalability and Workload Dynamics for Memory Pooling Systems},
journal = {ACM Transactions on Storage},
year = {2026},
volume = {22},
number = {4},
pages = {1--23},
doi = {10.1145/3806051}
}
EUVIMED