Database Query Optimization via Predictive Modeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing database query systems face limitations such as call limits and cost constraints, making it inefficient to submit queries, especially in scenarios where query responses follow learnable behavioral patterns, leading to increased query submissions and costs.
Innovation Solution
A computerized system that learns a physical process by sending queries, deriving a model from responses, and generating test queries to optimize database usage by predicting query responses, thereby reducing the number of queries needed to be sent to the database.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If queries are submitted to database to obtain accurate responses, then response accuracy is improved, but query submission cost and time consumption increase due to call limits
Solution Approach 1:
The system performs preliminary actions by sending a small initial set of queries to the database and using the responses to build a predictive model. This model is then used to forecast responses to future queries without actually submitting them to the database, thereby avoiding the time consumption of repeated queries while maintaining response accuracy through the learned patterns.
Solution Approach 2:
The system creates a virtual copy of the database behavior through the predictive model. Instead of repeatedly querying the actual database, the model replicates the database's response patterns based on the initial queries and learned behavioral patterns, providing accurate responses without the time and cost penalties of repeated actual queries.
2Measurement precision
If queries are submitted to database to obtain accurate responses, then response accuracy is improved, but query submission cost increases due to call limits and rate restrictions
Solution Approach 1:
The system performs preliminary actions by sending a small initial set of queries to the database and using the responses to build a predictive model. This model is then used to forecast responses to future queries without actually submitting them to the database, thereby avoiding the time consumption of repeated queries while maintaining response accuracy through the learned patterns.
Solution Approach 2:
The system creates a virtual copy of the database behavior through the predictive model. Instead of repeatedly querying the actual database, the model replicates the database's response patterns based on the initial queries and learned behavioral patterns, providing accurate responses without the time and cost penalties of repeated actual queries.
3Measurement precision
If more queries are submitted to database, then response accuracy is improved, but the number of queries needed increases due to call limits
Solution Approach 1:
The system performs preliminary actions by sending a small initial set of queries to the database and using the responses to build a predictive model. This model is then used to forecast responses to future queries without actually submitting them to the database, thereby avoiding the time consumption of repeated queries while maintaining response accuracy through the learned patterns.
Solution Approach 2:
The system creates a virtual copy of the database behavior through the predictive model. Instead of repeatedly querying the actual database, the model replicates the database's response patterns based on the initial queries and learned behavioral patterns, providing accurate responses without the time and cost penalties of repeated actual queries.
Data Source
AI summary
Minimizing queries by sending first query/ies, including query parameter/s having first query parameter value/s respectively, to a database, and receiving therefrom, response/s to the first query/ies respectively; and using at least the query/ies and response/s to compute a model which approximates the database's behavior, the behavior including generating responses pursuant to queries received. The model serves as current model, unless/until replaced by a later computed model. The method may iterate as follows at least once: selecting query/ies, more distant from the first query's query parameter/s than query/ies not selected; sending the selected query/ies to the database and using a received database's response, and the selected query/ies, and query-response pair/s already used, to compute another model of the database's behavior which serves as current model, unless and until replaced by a later computed model.