MCTuner: Spatial Decomposition-Enhanced Database Tuning via LLM-Guided Exploration

2025-09-07

AI Summary

Database configuration tuning faces challenges of high-dimensional, large-scale parameter spaces and prohibitively expensive exploration costs: existing approaches either neglect domain expertise or fail to search efficiently. This paper proposes MCTuner, a novel framework featuring (i) an LLM-guided Mixture-of-Experts (MoE) mechanism to identify performance-sensitive parameters by leveraging domain knowledge, and (ii) a recursive subspace decomposition algorithm that hierarchically reduces the dimensionality of the configuration space. Building upon this, MCTuner integrates hierarchical subspace Bayesian optimization to synergistically combine domain knowledge with data-driven learning. Evaluated on OLAP, OLTP, and HTAP benchmarks, MCTuner achieves up to 19.2% higher performance and 1.4× faster configuration discovery compared to state-of-the-art methods, significantly improving both tuning efficiency and accuracy.

Technology Category

• Search and Optimization: Algorithm Configuration
• Machine Learning: Auto ML and Hyperparameter Tuning
• Data Mining & Knowledge Management: Scalability, Parallel & Distributed Systems

Application Category

• Search and Retrieval-Augmented AI: Vertical and domain-specific search
• User Modeling, Personalization and Recommendation: ML for personalized search and recommendations
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

Database knob tuning is essential for optimizing the performance of modern database management systems, which often expose hundreds of knobs with continuous or categorical values. However, the large number of knobs and the vast configuration space make it difficult to identify optimal settings efficiently. Although learning-based tuning has shown promise, existing approaches either ignore domain knowledge by relying solely on benchmark feedback or struggle to explore the high-dimensional knob space, resulting in high tuning costs and suboptimal performance. To address these challenges, we propose MCTuner, an adaptive knob tuning framework that minimizes exploration in ineffective regions of the configuration space. MCTuner employs a Mixture-of-Experts (MoE) mechanism with specialized LLMs to identify performance-critical knobs. In further, MCTuner introduces the first spatial decomposition algorithm that recursively partitions the space into hierarchical subspaces, on which Bayesian Optimization is performed to efficiently search for near-optimal configurations. Evaluated on different benchmarks (OLAP, OLTP, and HTAP), MCTuner achieves up to 19.2% performance gains and 1.4x faster configuration discovery per iteration compared to state-of-the-art methods.

Problem

• Optimizing database performance via high-dimensional knob tuning
• Reducing exploration costs in vast configuration spaces
• Overcoming suboptimal performance in learning-based tuning methods

Innovation

• LLM-guided Mixture-of-Experts knob identification
• Spatial decomposition algorithm for hierarchical partitioning
• Bayesian Optimization in subspaces for efficient search