Distributed Video Processing Scheduling for Local Bottlenecks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video processing systems face resource bottlenecks and increased costs due to limited local computing resources, necessitating deployment in cloud services, which incur network overhead and expenses.
Innovation Solution
A management service monitors resource usage trends across multiple video processing computers, generating schedules to distribute video processing operations among these computers, prioritizing local processing when possible and offloading to other computers or cloud services as needed, based on resource availability, software, hardware, and networking requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If video processing operations are deployed in cloud services to overcome limited local computing resources, then processing capacity is improved, but network overhead and costs increase
Solution Approach 1:
The system segments video processing operations into multiple independent tasks that can be distributed across different computing resources. The video processing computer divides incoming video data into separate processing streams, assigning different operations (e.g., object detection, tracking, analysis) to different available resources, thereby optimizing the balance between local processing and cloud offloading.
Solution Approach 2:
The system introduces a scheduling dimension by implementing a video processing schedule that assigns processing tasks across multiple time periods and computing resources. This temporal and spatial scheduling allows the system to dynamically allocate operations to local or cloud resources based on current resource availability, reducing the need for constant cloud dependency and thereby reducing network overhead.
2Loss of energy
If more video processing operations are performed locally to reduce network overhead, then network overhead is reduced, but resource availability becomes a bottleneck
Solution Approach 1:
The system dynamically adjusts the distribution of video processing operations based on real-time resource availability. The video processing schedule is generated and updated based on monitored resource usage trends, allowing the system to adaptively shift processing loads between local and cloud resources. This dynamic scheduling ensures that processing throughput is maintained by utilizing cloud resources when local resources are constrained, while minimizing network overhead when local resources are available.
3Device complexity
If video processing tasks are concentrated on a single computer to simplify system management, then system complexity is reduced, but resource utilization efficiency decreases
Solution Approach 1:
The system implements a universal scheduling mechanism that can manage multiple video processing computers and cloud resources through a single centralized video processing schedule. This schedule serves multiple functions: resource allocation, task distribution, performance monitoring, and dynamic reconfiguration. By using this multi-functional scheduling system, the patent achieves efficient multi-resource utilization without proportionally increasing system management complexity.
Data Source
AI summary
Systems, methods, and software to manage the distribution of video data for processing across multiple computers. In at least one implementation, a management service obtains resource usage associated with video processing computers, wherein each of the video processing computers manages video processing of video data associated with one or more cameras. The management service further identifies trends based on the resource usage and, for at least a first video processing computer, generates a processing schedule for first video processing computer to distribute the video processing to one or more other video processing computers. The management service also configures the first video processing computer with the video processing schedule.