Isolated HPC Resource Partitioning for Secure Multi-Tenant Utilization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current high-performance computing devices are limited by high initial investment costs, inefficient resource utilization, and security concerns in on-premise and as-a-service models, particularly when used by multiple entities with heterogeneous needs.
Innovation Solution
A high-performance computing device with adaptable computing power is designed to distribute resources into physically isolated high-performance computers, each managed by a local resource manager and accessed through a global manager, ensuring secure and efficient utilization by multiple entities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high-performance computing resources are dedicated to a single entity, then security and performance are ensured, but resource utilization rate decreases and initial investment costs increase
Solution Approach 1:
The patent segments the HPC device into multiple physically isolated computers, each dedicated to a specific entity. This segmentation allows security to be maintained for each entity while enabling the overall system to serve multiple entities, thereby improving resource utilization. The physical isolation ensures that security requirements are met while the multi-tenant architecture increases utilization rates.
Solution Approach 2:
The HPC device is designed with multi-functionality to serve multiple entities simultaneously. Each computer within the device can be independently allocated to different entities based on their computing needs, making the system universal in its application while maintaining dedicated security boundaries for each user.
2Productivity
If HPC resources are shared among multiple entities, then resource utilization increases, but security concerns and performance degradation occur
Solution Approach 1:
The system segments the HPC device into multiple physically isolated computers, allowing resources to be shared at the device level while maintaining security through physical separation. Each entity receives dedicated computational resources within its isolated computer, ensuring security is not compromised while achieving high resource utilization across the entire device.
Solution Approach 2:
The patent introduces a resource allocation mechanism that acts as an intermediary between multiple entities and the physical HPC resources. This intermediary manages the allocation of computing resources fairly and securely, enabling high utilization while maintaining security boundaries through controlled access and isolation mechanisms.
3Ease of manufacture
If standard hardware is used for HPC as a service, then initial investment costs decrease, but execution time increases and energy consumption becomes disproportionate
Solution Approach 1:
The patent applies local quality by allocating specific high-performance computing resources to each entity based on their particular needs. Rather than using uniform standard hardware for all workloads, the system provides locally optimized computing power tailored to each entity's computational requirements, ensuring fast execution times while managing costs effectively.
4Power
If HPC device is sized large to meet peak demands, then sufficient computing power is available, but resource utilization rate decreases
Solution Approach 1:
The HPC device is designed with multi-functionality to serve multiple entities with different computing needs. By allocating resources across multiple tenants, the system can maintain large computing power capacity while achieving high utilization rates, as each entity uses the device during different time periods or for different workloads.
Solution Approach 2:
The system implements dynamic resource allocation that can adapt to varying demands of different entities. The resource allocation mechanism can dynamically adjust the distribution of computing resources based on current workload requirements, ensuring that sufficient computing power is available when needed while maximizing overall utilization across the device.
Data Source
AI summary
A high-performance computing device with adaptable computing power in which several high-performance computers exist that are isolated from each other to ensure data security. Each computer is run by a local resource manager that manages the computer at the node level. All or some of the local managers are managed via a global resource manager that the high-performance computing device manages at the level of the computers it contains.