ML-Based Application Deployment Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Deploying applications across different computing environments, such as on-premises and off-premises systems, is prone to errors and resource inefficiencies due to the complexity of determining deployability and suitable environments, often requiring extensive expertise and multiple attempts.
Innovation Solution
The use of machine learning models to classify application modules based on metadata, enabling a one-step deployment process across various computing environments by distinguishing between deployability and performance-related criteria, and selecting appropriate environments using trained models for accurate and efficient deployment.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual assessment and decision-making is used for each business application deployment, then deployment accuracy can be maintained through expert judgment, but the process becomes intensive resource consuming and time-consuming
Solution Approach 1:
The patent replaces manual expert assessment (mechanical human judgment system) with an automated machine learning classification system. The ML model analyzes application metadata and characteristics to automatically determine deployability and select target environments, eliminating the need for intensive manual resource consumption while maintaining deployment accuracy through systematic classification criteria.
Solution Approach 2:
The system enables applications to be self-classified and self-deployed by automatically analyzing their own metadata and characteristics through the ML model. The application deployment process becomes self-service oriented, where the system autonomously determines deployability and selects appropriate target environments without requiring continuous human intervention for each deployment decision.
2Reliability
If multiple attempts and expertise are required to determine deployability, then deployment reliability can be improved through thorough assessment, but the complexity of the deployment process increases
Solution Approach 1:
The patent segments the deployment assessment process into distinct classification categories (e.g., cloud-native, containerized, virtualized, traditional applications). Each segment has specific metadata requirements and deployment criteria, allowing the system to handle different application types through standardized classification pathways rather than requiring comprehensive manual assessment of all possible scenarios.
Solution Approach 2:
The system changes the assessment parameters from qualitative expert judgment to quantitative metadata analysis. By transforming deployment criteria into measurable parameters (application characteristics, metadata attributes, technical specifications), the ML model can systematically evaluate deployability without requiring complex multi-step manual assessment procedures.
3Measurement precision
If traditional deployment assessment methods are used, then thorough evaluation of application characteristics can be achieved, but the time required for deployment decision-making increases
Solution Approach 1:
The patent implements preliminary classification of applications based on their metadata and characteristics before actual deployment. The ML model pre-assesses application deployability and identifies suitable target environments in advance, so that when deployment is initiated, the decision-making process is already complete or can be rapidly finalized, significantly reducing actual deployment time while maintaining thorough evaluation.
Solution Approach 2:
The system enables continuous automated assessment of application deployability through the ML model, eliminating interruptions and repeated manual evaluations. The classification process runs continuously as applications are submitted for deployment, maintaining constant readiness to deploy without the time losses associated with starting fresh assessments for each deployment attempt.
Data Source
AI summary
Aspects of the present invention disclose a method for deploying an application in a computing environment receiving an application module, determining values of a first set of metadata for the received application module, determining a classification of the received application module based at least in part on the values of the first set of metadata, and determining whether the received application is deployable in at least an off-premise system based at least in part on the classification.