White-box micro-adaptive query processing
File(s)
Accepted version
Author(s)
Pearce, Jack
Mohr-Daurat, Hubert
Pirk, Holger
Type
Conference Paper
Abstract
Operator performance in in-memory data management systems (DMS) often suffers from micro-architectural hazards such as cache misses and branch mispredictions. While many operators have alternative implementations that are robust against such hazards, these generally perform worse when no hazards are encountered. Unfortunately, hazards are caused by order-dependent data characteristics that query optimizers struggle to capture (e.g., sortedness, clusteredness) making a prior hazard-conscious optimization difficult. Additionally, statically optimized plans fail to adapt when data characteristics vary within a table. To address these problems, we propose a hazard-adaptive approach to query execution. Through hardware-assisted runtime profiling of low-level metrics, operators dynamically adapt to "hazardous" data. We propose an architecture for hazard-adaptive operators and integrate our approach into a DMS. We demonstrate that using hazard-adaptive operators provides a ~2-20× speedup across several TPC-H queries.
Date Issued
2025-05-19
Date Acceptance
2025-03-26
Citation
Proceedings of the 2025 IEEE International Conference on Data Engineering (ICDE), 2025, pp.2880-2893
Publisher
IEEE Computer Society
Start Page
2880
End Page
2893
Journal / Book Title
Proceedings of the 2025 IEEE International Conference on Data Engineering (ICDE)
Copyright Statement
© 2025 IEEE. This is the author’s accepted manuscript made available under a CC-BY licence in accordance with Imperial’s Research Publications Open Access policy (www.imperial.ac.uk/oa-policy)
License URL
Source
2025 IEEE International Conference on Data Engineering (ICDE)
Publication Status
Published
Start Date
2025-05-19
Finish Date
2025-05-23
Coverage Spatial
Hong Kong SAR, China