Cloud Database Storage Advisor for In-Memory Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In-memory database management systems face a trade-off between Total Cost of Operation (TCO) and performance, with manual workload and access pattern analysis being time-consuming and challenging to achieve an optimal storage configuration, especially in cloud environments.
Innovation Solution
A configuration advisor system that collects statistics on data access in a cloud environment and automatically generates recommendations for page-loadable or column-loadable units and persistence memory layers, using rule-based heuristics and UIs to optimize storage configuration.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If information is stored in-memory, then processing speed is improved, but cost increases
Solution Approach 1:
The patent applies local quality by differentiating storage strategies for different data access patterns. Frequently accessed data is stored in-memory for fast processing, while rarely accessed data is stored on disk to reduce cost. The system automatically identifies and separates data based on access frequency statistics, applying appropriate storage quality locally rather than uniformly across all data.
Solution Approach 2:
The configuration advisor system performs self-service by automatically analyzing access patterns and generating storage configuration recommendations without manual intervention. The system collects statistics on data access, processes this information through rule-based heuristics, and produces actionable recommendations for optimizing the balance between in-memory storage performance and cost efficiency.
2Manufacturing precision
If manual analysis of workloads and access patterns is performed, then optimal storage configuration is achieved, but time consumption increases
Solution Approach 1:
The configuration advisor system enables self-service by automatically collecting access statistics, processing them through predefined heuristics, and generating optimized storage configurations without requiring manual analysis. This automation eliminates the time-consuming manual intervention while maintaining optimization quality through systematic rule-based processing.
Solution Approach 2:
The system implements feedback mechanisms by continuously collecting access pattern statistics, comparing them against optimization rules, and adjusting storage recommendations accordingly. This closed-loop feedback process enables automatic optimization that adapts to changing workloads without requiring manual reanalysis.
3Productivity
If automated storage configuration recommendations are generated, then time consumption is reduced, but system complexity increases
Solution Approach 1:
The system manages complexity by changing parameters such as access frequency thresholds, memory allocation ratios, and optimization rules that can be adjusted based on workload characteristics. These parameter changes allow the automated system to adapt to different scenarios without requiring complex custom logic for each case, maintaining productivity while controlling complexity through configurable parameters.
Data Source
AI summary
Embodiments may include a collected statistics data store that contains statistics about information accessed via an in-memory database management system that is executing in a cloud computing environment. A configuration advisor executing in the cloud computing environment may retrieve the statistics about information accessed via the in-memory database management system. Based on the retrieved statistics, the configuration advisor may automatically generate a data storage recommendation for the in-memory database management system. The data storage recommendation may be, for example, associated with page-loadable or column-loadable units and a persistence memory layer. According to some embodiments, the data storage recommendation may be further based on a storage cost, a performance metric, an access frequency, an object threshold (e.g., a threshold for a data column, a data partition, or a data table), recommendation rule-based heuristics, etc.