Edge Device Application Selection via Resource Headroom Profiling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current mechanisms for deciding which computing devices to install new applications on are limited in their configuration and operation, lacking an efficient method to optimize resource allocation and application selection based on available resource headroom.
Innovation Solution
A method and system that utilize a processor to receive resource headroom data and resource utilization data of applications, compute a fitness score, and generate rewards to refine the selection of applications for installation on edge devices, employing deep reinforcement learning and convolutional neural networks to optimize resource usage.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional application selection mechanisms are used, then device compatibility is maintained, but resource utilization efficiency deteriorates
Solution Approach 1:
The patent implements feedback loops where resource utilization data from edge devices is continuously collected, processed through machine learning models, and used to refine application selection decisions. The system learns from past deployment outcomes and resource consumption patterns to improve future selections, creating a closed-loop optimization system that enhances resource utilization while adapting to changing device states.
Solution Approach 2:
The system dynamically adjusts selection parameters based on real-time resource headroom data. Instead of using fixed selection criteria, the machine learning models evaluate multiple parameters including available compute power, memory, storage, and network bandwidth, adjusting the weight and threshold of each parameter based on current device conditions to optimize resource utilization efficiency.
2Adaptability or versatility
If more applications are installed on edge devices, then service coverage is improved, but resource constraints are violated
Solution Approach 1:
The patent employs dynamic application selection that adapts to changing resource conditions. The system continuously monitors resource headroom and adjusts which applications are deployed or removed from edge devices based on real-time availability. This dynamic approach allows the system to maintain service coverage by deploying applications when resources are available while ensuring resource constraints are never violated through continuous validation and automated remediation.
Solution Approach 2:
The system performs preliminary resource allocation analysis and application compatibility assessment before deployment. Machine learning models predict resource consumption patterns and evaluate potential impacts on existing applications, allowing the system to pre-validate deployment decisions to ensure resource constraints will not be violated, thereby maintaining reliability while expanding service coverage.
3Productivity
If manual application deployment is used, then control precision is maintained, but deployment efficiency deteriorates
Solution Approach 1:
The patent implements self-service automated deployment systems that independently evaluate edge devices, select appropriate applications, and execute deployment without manual intervention. The machine learning models autonomously analyze device characteristics, resource availability, and application requirements to make deployment decisions, dramatically improving deployment efficiency while the modular architecture manages system complexity through standardized interfaces and automated workflows.
Data Source
AI summary
A system and method include receiving a resource headroom data of an edge device, receiving resource utilization data of a plurality of applications, and selecting a group of applications from the plurality of applications for installation on the edge device based upon the resource headroom of the edge device. The system and method also include computing a fitness score based upon a suitability of the selected group of applications for the edge device, generating a reward based on the fitness score, and using the reward to refine the selection of the group of selections in subsequent iterations.