Software Dependency Analysis via Static and Dynamic Agent Synthesis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Identifying and managing software dependencies within an enterprise is complex and time-consuming, especially when bugs are detected or updates are necessary, due to the intricate web of software assets and interdependencies, which can only become evident at runtime.
Innovation Solution
A computer-implemented method using technology agents to perform both static and dynamic analysis of software assets, assigning node IDs based on hash content, and communicating hierarchies and dependencies to synthesis agents to create a dependency graph, facilitating efficient identification and tracking of software relationships across the enterprise.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If static analysis is performed on software assets to identify dependencies, then dependency identification efficiency is improved, but dependencies that only become evident at runtime cannot be detected
Solution Approach 1:
The patent divides dependency analysis into two separate analysis modes: static analysis (performed on software assets without execution) and dynamic analysis (performed during runtime execution). Each analysis type is handled by separate agent components that collect dependency information through different mechanisms, allowing both to coexist and complement each other without interference.
Solution Approach 2:
The patent merges the results from static analysis and dynamic analysis into a unified dependency graph. The synthesis agent combines dependency information from both analysis types, creating a comprehensive view of software dependencies that includes both compile-time relationships and runtime-discovered relationships.
2Measurement precision
If comprehensive dependency analysis is performed across all software assets, then accuracy of dependency identification is improved, but time and complexity of analysis increases significantly
Solution Approach 1:
The patent performs static analysis as a preliminary step before dynamic analysis. The static analysis agent identifies and records dependencies that can be detected without execution, creating an initial dependency graph. This preliminary work reduces the scope and time required for subsequent dynamic analysis, as the system can focus runtime analysis on dependencies not already identified.
Solution Approach 2:
The system performs analysis in two phases rather than attempting complete analysis simultaneously. Static analysis covers all assets comprehensively at build time, while dynamic analysis focuses on runtime-specific dependencies. This partial sequencing of analysis actions manages overall time and complexity while maintaining comprehensive accuracy.
3Productivity
If multiple technology agents analyze different software types in parallel, then analysis throughput is improved, but system complexity and coordination overhead increases
Solution Approach 1:
The synthesis agent is designed as a universal component that handles integration of results from multiple specialized technology agents. Each technology agent focuses on analyzing a specific software type or asset, while the synthesis agent provides a common interface for collecting, normalizing, and merging dependency information from all sources into a unified dependency graph.
Solution Approach 2:
The synthesis agent acts as an intermediary between the multiple technology agents and the final dependency graph output. It receives dependency information from various agents through standardized communication protocols, reconciles different data formats and representations, and produces a unified dependency model, thereby simplifying coordination between parallel analysis processes.
Data Source
AI summary
An analysis of software stored and executing on one or more assets in an enterprise includes determining a hierarchy of nodes in the software and among the assets and analyzing dependencies among those nodes and assets through a static analysis of the software as it is stored as well as through a dynamic analysis of the software as it executes.