Application Boundary Detection Using Two-Stage Unit Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Modern software applications face challenges in scaling to handle large user bases, managing vast data, and responding quickly, necessitating software modernization, but existing methods lack effective tools for understanding and optimizing software ecosystems.
Innovation Solution
An application boundary detection service using a two-stage machine learning approach clusters application units into distinct groups, leveraging both supervised and unsupervised techniques to identify predicted application boundaries, incorporating customer feedback for improved accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If software applications are modernized to scale quickly and handle large data, then productivity and performance are improved, but the complexity of understanding and optimizing software ecosystems increases
Solution Approach 1:
The patent segments the software ecosystem into distinct application units and identifies boundaries between them using clustering algorithms. This segmentation transforms the complex monolithic software ecosystem into manageable, identifiable components with clear delimitations, enabling organizations to understand and optimize each segment independently while maintaining overall system productivity and scaling capability.
2Ease of operation
If existing software tools are used without boundary detection, then ease of operation is maintained, but measurement precision of application boundaries is insufficient
Solution Approach 1:
The patent implements self-service through automated boundary detection that operates without requiring manual intervention or complex user input. The system automatically analyzes software ecosystems, identifies application units, and determines boundaries using machine learning models, thereby maintaining ease of operation while significantly improving measurement precision of application boundaries.
Solution Approach 2:
The patent replaces manual analysis methods with automated machine learning-based boundary detection systems. This substitution eliminates the need for manual software ecosystem analysis while providing precise, data-driven identification of application boundaries, thereby improving measurement precision without sacrificing ease of operation.
3Measurement precision
If manual analysis of software ecosystems is performed, then measurement precision can be maintained, but loss of time increases
Solution Approach 1:
The patent applies preliminary action by pre-processing software metadata, collecting ecosystem data, and preparing training models before actual boundary detection is needed. This preparation enables rapid, automated boundary identification without requiring time-consuming manual analysis during the actual detection process, thereby maintaining measurement precision while significantly reducing analysis time.
Data Source
AI summary
Techniques for application boundary detection are described. A request to identify application boundaries within customer software executed in a cloud environment is received. Metadata associated with the customer software is obtained, the metadata including an identification of application units, features associated with each identified application unit, and connectivity between application units. The application units are clustered into initial application unit clusters based at least in part on the features associated with each application unit and the connectivity between application units. The initial application unit clusters are iteratively combined into at least some increasingly larger application unit clusters to generate final application unit clusters. A result including an identification of the application units in the final application unit clusters is output.