Differentially Private Canary Optimization via Thompson Sampling for SQL Performance Fixes

Chiranjeevi Devi (), Pradeep Manivannan () and Radhakrishnan Pachyappan ()

Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023, 2024, vol. 3, issue 1, 532-550

Abstract: SQL performance tuning often involves deploying multiple query fixes in live systems to identify the most effective solution, a process that can risk data exposure and degrade service quality. This paper proposes a novel framework that integrates differential privacy with Thompson Sampling to optimize "canary deployments" of SQL fixes. Our method ensures that experimental testing across user groups maintains statistical efficiency while preserving user data privacy. By leveraging Bayesian exploration strategies, our approach identifies high-performing query modifications under strict privacy constraints, minimizing performance regression and exposure. Experimental evaluations on benchmark SQL workloads demonstrate that our approach achieves near-optimal performance improvements with provable differential privacy guarantees, offering a robust solution for safely and effectively deploying performance fixes in sensitive environments.

Keywords: Differential Privacy; Canary Deployment; SQL Performance Optimization; Thompson Sampling; Multi-Armed Bandits; Bayesian Optimization; Query Tuning (search for similar items in EconPapers)
Date: 2024
References: Add references at CitEc
Citations:

Downloads: (external link)
https://newjaigs.com/index.php/JAIGS/article/view/383 (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:das:njaigs:v:3:y:2024:i:1:p:532-550:id:383

Access Statistics for this article

Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023 is currently edited by Justyna Żywiołek

More articles in Journal of Artificial Intelligence General science (JAIGS) ISSN:3006-4023 from Open Knowledge
Bibliographic data for series maintained by Open Knowledge ().