Model Data Store Access Pattern Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In modeling environments, accessing data stores can lead to issues such as reading uninitialized data or losing information due to improper synchronization between model components, resulting in undesirable read and write patterns.
Innovation Solution
A system and method for analyzing models to identify patterns associated with data store access, using execution control graphs to detect and highlight anomalous access patterns, allowing for the identification and correction of issues like reading before writing or writing multiple times before reading.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If multiple portions of the model access the data store simultaneously, then information sharing between model portions is improved, but data integrity deteriorates due to read-before-write and write-before-read conflicts
Solution Approach 1:
The patent performs preliminary analysis of data store access patterns during model compilation or before execution. It identifies potential read-before-write and write-before-read conflicts in advance and generates warnings or errors to prevent these integrity issues before they occur during model execution.
Solution Approach 2:
The patent implements a feedback mechanism that analyzes the model's execution control graph to detect problematic access patterns. The system provides feedback to the user about identified conflicts, allowing the user to modify the model to eliminate read-before-write and write-before-read scenarios that would compromise data integrity.
2Productivity
If data store access is unrestricted, then model execution speed is improved, but loss of information increases due to uninitialized reads and premature writes
Solution Approach 1:
The patent performs preliminary analysis during model compilation to identify all data store access operations. It determines the initialization status and access timing of each operation, flagging potential uninitialized reads and premature writes before execution begins, allowing users to correct these issues without impacting execution speed.
Data Source
AI summary
In an embodiment, a technique that may be used to identify a pattern with respect to accessing a data store in a model. The pattern may be a desirable, undesirable, anomalous or some other type of pattern with respect to accessing the data store. The technique may include generating an execution control graph that represents an execution of the model. The execution control graph may be analyzed to identify the pattern. Analysis may include generating an expression based on the execution control graph and a condition to test for and determining, based on the expression, if the condition is met. If the condition is met, the pattern may be said to exist in the model. A result may generated based on the analysis and the result may be output.