Edge Visual Resource Deployment for Vision Application Latency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing solutions for deploying visual resources in network systems, particularly for VR/AR applications, face challenges in ensuring timely delivery due to high data requirements, leading to latency issues, as they fail to accurately predict and pre-deploy resources based on processing capabilities and time requirements.
Innovation Solution
A method that utilizes a prediction engine to anticipate the resource needs of vision applications, identifies suitable edge devices near terminal devices based on processing capabilities, and deploys visual resources in advance to minimize latency, optimizing network efficiency by aggregating and sharing resources without increasing hardware infrastructure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of time
If visual resources are deployed on-demand without prediction, then network infrastructure requirements are reduced, but latency increases due to delayed resource delivery
Solution Approach 1:
The system performs preliminary actions by predicting future resource requirements of vision applications and pre-deploying visual resources to edge devices before they are actually needed. This advance preparation eliminates latency during runtime resource delivery while managing complexity through automated prediction algorithms and intelligent resource placement strategies.
2Speed
If visual resources are pre-deployed to edge devices, then latency is reduced, but network system complexity increases
Solution Approach 1:
The system implements feedback mechanisms that continuously monitor resource usage patterns, application performance metrics, and network conditions. This feedback informs dynamic adjustments to prediction accuracy and resource placement decisions, enabling the system to maintain high delivery speed while adapting to changing conditions and managing complexity through data-driven optimization.
Solution Approach 2:
The system changes parameters such as prediction time horizons, resource caching strategies, and edge device selection criteria based on observed performance and network conditions. By dynamically adjusting these parameters, the system optimizes resource delivery speed while managing the complexity of the deployment mechanism through adaptive parameter tuning rather than fixed complex rules.
3Power
If more edge devices are deployed to handle vision applications, then processing capability increases, but infrastructure cost and complexity increase
Solution Approach 1:
The system enables existing edge devices to serve multiple functions by deploying visual resources that can support various vision applications across different terminal devices. Rather than requiring dedicated infrastructure for each application, the same edge devices handle diverse workloads through resource virtualization and dynamic allocation, increasing processing capability without proportionally increasing infrastructure complexity.
Solution Approach 2:
The system creates virtual copies of visual resources across multiple edge devices rather than physically duplicating hardware infrastructure. This allows the network to handle increased processing demands by distributing resource instances across existing devices, effectively scaling processing capability while avoiding the complexity and cost of deploying additional physical infrastructure.
Data Source
AI summary
The present disclosure relates to a method, a device, and a program product for deploying a visual resource. In one method, a resource requirement of a vision application for the visual resource in a network system is acquired. Based on the resource requirement, the visual resource which will be called by the vision application is predicted. Based on processing capabilities of various edge devices and the visual resource in the network system, an edge device located near a terminal device in the network system is identified, wherein the terminal device is configured to run the vision application. Based on a time requirement in the resource requirement, the visual resource is deployed to the edge device. Further, a corresponding device and a corresponding program product are provided.