Collaborative Query Optimization Across DBMS Servers
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing query optimization solutions in database management systems struggle to correlate query plans across different systems due to hardware and load differences, leading to inefficient resource allocation and increased licensing costs.
Innovation Solution
A system of interconnected DBMS servers with comparable hardware resources and load, utilizing a query cost manager component to build a deployment cost estimation function and share query-related data for collaborative optimization, allowing each server to identify and implement optimal query plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If query optimization is performed using conventional solutions, then query plans can be generated for individual systems, but the solutions cannot be correlated across different systems due to hardware and load differences
Solution Approach 1:
The patent segments the query optimization process into independent units (query plans) that can be generated and shared across multiple DBMS servers. Each server maintains its own query plan generation capabilities while exchanging optimized plans with peers, enabling cross-system correlation without requiring complete system homogeneity.
Solution Approach 2:
The patent implements copying by allowing DBMS servers to share and replicate optimized query plans across the network. Servers can copy successful query plans from peers with similar workloads and apply them locally, enabling knowledge transfer without direct system coupling.
2Quantity of substance
If rate-based licensing is used for query optimization, then system access is controlled, but users must acquire larger licenses than required for global deployment
Solution Approach 1:
The patent creates a universal query optimization framework where a single licensed server can serve multiple functions by sharing optimized query plans with other servers. The licensing model supports multi-server deployment where each server benefits from the collective optimization knowledge base, reducing the need for separate licenses on each node.
3Productivity
If query plans are generated without collaborative exchange, then each system operates independently, but optimization opportunities are lost due to lack of shared knowledge
Solution Approach 1:
The patent merges the query optimization capabilities of multiple DBMS servers into a collaborative network. Servers exchange query plans and optimization knowledge, combining their individual strengths to achieve better overall optimization effectiveness than any single server could achieve alone.
Solution Approach 2:
The patent implements feedback mechanisms where servers receive optimized query plans from peers and apply them to improve local query performance. The feedback loop continues as servers share their own optimization results, creating a continuously improving system-wide knowledge base.
Data Source
AI summary
A method, system, apparatus, and a computer program product are provided for optimizing database query plans generated from a received database query at a database management system by monitoring a global cost estimation function for one or more steps of the received database query to identify a first low-performing cost metric and collaboratively optimizing a query plan for the database query by comparing differences between the first low-performing cost metric and a second cost metric at an additional database management system, thereby generating an optimized query plan for the database management system that may be stored at the database management system along with corresponding information identifying a reason for rewriting the optimized query plan.