VF-Level Resource Allocation for Shared Server Resource Pools
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional rack-mounted servers face inefficiencies in resource management due to asynchronous lifecycles among different types of resources, leading to wasted CPU performance and power consumption when storage capabilities are inadequate, and existing resource pooling technologies lack effective methods for determining and managing resource configurations and allocations for multiple user hosts.
Innovation Solution
A device resource management system that separates physical and function configuration resources, allowing for finer granularity allocation of virtual functions (VF-level) to different user hosts, enabling flexible resource utilization and management through a management host that discovers, configures, and allocates resources at a VF-level granularity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If conventional rack-mounted servers are used, then server density is maintained at traditional levels, but resource utilization is low and CPU performance is wasted when storage capabilities are inadequate
Solution Approach 1:
The patent segments physical resources into multiple virtual functions (VF-level granularity), allowing independent allocation and management of different resource types. This enables storage resources to be upgraded or replaced without affecting CPU resources, resolving the asynchronous lifecycle problem and improving resource utilization while avoiding energy waste from premature decommissioning
Solution Approach 2:
The management host provides universal resource management capabilities that can handle multiple types of resources (CPU, storage, memory, I/O) through a unified virtualization framework. This multi-functional approach allows flexible allocation and optimization of mixed resource types, improving overall resource utilization across the server pool
2Productivity
If resource pooling technology is implemented, then server density and resource sharing are improved, but resource allocation flexibility and management granularity remain insufficient
Solution Approach 1:
The patent divides physical resources into virtual functions at VF-level granularity, enabling fine-grained allocation to multiple user hosts. This segmentation allows individual virtual functions to be allocated, monitored, and managed independently, providing the flexibility and adaptability needed for dynamic resource pooling while maintaining high resource sharing efficiency
Solution Approach 2:
The management host implements dynamic resource allocation where virtual functions can be allocated, reassigned, or reclaimed based on real-time user host requirements. This dynamic approach enhances adaptability by allowing the system to respond to changing workload demands while maintaining efficient resource sharing across the pooled environment
3Measurement precision
If VF-level resource allocation is implemented, then resource allocation granularity is improved, but system complexity increases
Solution Approach 1:
The management host serves as an intermediary between physical resources and user hosts, handling the complexity of VF-level resource allocation, configuration, and management. By centralizing these management functions in the intermediary management host, the system achieves fine-grained allocation precision while shielding user hosts from the underlying complexity of virtual function management
Data Source
AI summary
This application discloses a device resource management method, a related system, and a storage medium. The system may include at least one user host, at least one resource device, and a management host. Each resource device includes a first entity, at least one second entity, where a total of N second entities are included. The first entity includes a physical configuration resource, and the one or more second entities include a function configuration resource generated after the physical configuration resource is configured. The management host allocates, according to a resource allocation policy, the N second entities to the at least one user host for use, and the target user host obtains L allocated second entities. The target user host registers the L second entities and uses the L second entities. Thereby providing a finer resource allocation granularity, higher resource utilization, and greater flexibility resource allocation.