Hybrid Database Query Routing for Latency Reduction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current database systems optimized for online transaction processing (OLTP) are not suitable for handling online analytical processing (OLAP) workloads, leading to performance issues and infrastructure complexities when complex analytics and reporting are required on transactional data.
Innovation Solution
A hybrid system is proposed, where a primary database system optimized for OLTP is extended with OLAP capabilities through the use of accelerator database systems, featuring different configurations such as row store and column store database management systems, and machine learning predictive models to route queries based on latency data for efficient execution.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If a primary database system optimized for OLTP is used, then transaction processing performance is improved, but analytical processing capability deteriorates
Solution Approach 1:
The patent implements a hybrid database system where a single primary database system can handle both OLTP and OLAP workloads by dynamically routing queries to appropriate processing modes or storage configurations, making the system universal for both transactional and analytical processing
Solution Approach 2:
The patent segments the database system into different processing pathways or storage configurations within the primary system, allowing transactional data to be organized in row-store format for OLTP while enabling column-store views for OLAP operations without requiring separate physical systems
2Adaptability or versatility
If separate OLAP systems are added to handle analytical workloads, then analytical processing capability is improved, but infrastructure complexity increases
Solution Approach 1:
The patent merges OLTP and OLAP capabilities into a single hybrid database system, combining transactional and analytical processing functions in one infrastructure rather than maintaining separate systems, thereby reducing overall complexity while preserving analytical capability
Solution Approach 2:
The primary database system is designed to be universal, handling both transactional and analytical workloads through dynamic query routing and multi-mode processing, eliminating the need for separate OLAP systems and their associated infrastructure complexity
3Adaptability or versatility
If data is replicated across multiple accelerator database systems, then query execution flexibility is improved, but data synchronization latency increases
Solution Approach 1:
The patent implements a feedback mechanism where the primary database system continuously monitors data synchronization status and query performance across accelerator systems, using this feedback to dynamically adjust query routing decisions and prioritize critical data replication to minimize latency impact
Data Source
AI summary
Methods, computer program products, and systems are presented. The method computer program products, and systems can include, for instance: receiving by primary database system a query statement that references a first table, wherein the primary database system is associated to a first accelerator database system and a second accelerator database system, wherein modifications performed on tables of the primary database system are iteratively replicated into the first accelerator database system and the second accelerator database system; examining latency data that specifies a latency with which a modification to the first table has been replicated into the first accelerator database system and a latency with which a modification to the first table has been replicated into the second accelerator database system; selecting, in dependence on the examining, the first accelerator database system for execution of the query statement; routing the query statement to the first accelerator database system for execution in response to the selecting.