Logical Dedicated Processing Resources for Dynamic Allocation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional methods for allocating dedicated processing resources, such as GPUs, are inefficient as they typically allocate physical resources exclusively to a single application, failing to fully utilize the resources' capabilities and leading to wastage, especially in cloud-based computing systems where multiple applications require processing power.
Innovation Solution
The method involves dividing physical dedicated processing resources into logical resources, allowing for finer-grained scheduling and allocation, enabling multiple applications to share the same physical resources by creating virtual GPUs (vGPUs) based on predefined specifications and allocation strategies, which are managed by a controller that maps and allocates these resources dynamically.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If physical dedicated processing resources are allocated exclusively to a single application, then the application can utilize the full computing power of the resource, but the resource utilization rate decreases when multiple applications need processing power
Solution Approach 1:
The patent segments physical dedicated processing resources into multiple logical resources, allowing each logical resource to be independently allocated to different applications. This segmentation enables fine-grained resource allocation where a single physical resource can serve multiple applications simultaneously through time-multiplexed or space-multiplexed logical divisions, thereby improving overall resource utilization while maintaining adequate computing power for each application.
Solution Approach 2:
The patent implements dynamic allocation of logical resources to applications based on real-time demands. The system can dynamically adjust which logical resources are allocated to which applications, allowing the resource allocation to adapt to changing workload requirements. This dynamic approach enables the system to optimize resource utilization rates while ensuring each application receives appropriate computing power when needed.
2Loss of energy
If physical resources are divided into logical resources for finer-grained scheduling, then resource utilization improves, but system complexity increases
Solution Approach 1:
The patent introduces a resource management system that acts as an intermediary between physical resources and applications. This intermediary layer handles the complexity of dividing and managing logical resources, maintaining mapping relationships between physical and logical resources, and making allocation decisions. By centralizing this management functionality, the system achieves fine-grained resource control without requiring complex modifications to individual applications or the underlying physical hardware.
Solution Approach 2:
The patent creates logical copies of physical resources that can be independently managed and allocated. These logical resources are virtual representations that mirror the functionality of physical resources but can be divided, copied, and distributed more flexibly. The logical resource layer provides a simplified interface for allocation while the underlying physical resources remain unchanged, reducing the complexity of physical resource management.
Data Source
AI summary
A resource management method comprises: in response to receiving, from an application operating on a client, a resource allocation request indicating an amount of dedicated processing resources required by the application, acquiring a mapping between a group of physical dedicated processing resources provided by a group of servers and a group of logical dedicated processing resources, the group of physical dedicated processing resources being divided into the group of logical dedicated processing resources; determining allocation statuses of the group of logical dedicated processing resources; determining, based at least on the mapping and the allocation statuses, a first amount of logical dedicated processing resources to be allocated to the application from the group of logical dedicated processing resources; and indicating the first amount of logical dedicated processing resources to the application, to allow the application to utilize physical dedicated processing resources provided by at least one of the group of servers.