Hyper-Converged Node Partitioning via Capability Discovery
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing hyper-converged infrastructure systems do not efficiently allocate tasks due to treating connected nodes as having homogeneous processing and storage capabilities, despite actual hardware differences, leading to suboptimal efficiency.
Innovation Solution
The system detects new node elements, assigns them an electronic address, boots using a pre-existing image, discovers capabilities through a federated control plane, and assigns them to a global storage pool based on determined resources, allowing for efficient task allocation by matching tasks with node partition groups optimized for specific capabilities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If tasks are allocated without regard to node hardware differences, then system simplicity is maintained, but task execution efficiency deteriorates
Solution Approach 1:
The system segments nodes into different capability groups based on hardware specifications (compute, storage, memory). This segmentation allows the task allocation mechanism to match tasks with appropriately capable nodes, resolving the contradiction by introducing structured complexity that directly improves execution efficiency without overwhelming system complexity
Solution Approach 2:
The system changes the parameter of node classification from homogeneous to heterogeneous based on actual hardware capabilities. By discovering and utilizing node-specific parameters (CPU, memory, storage), the system optimizes task allocation efficiency while maintaining manageable complexity through automated discovery and grouping
2Productivity
If node capabilities are discovered and tasks are allocated based on capabilities, then task execution efficiency is improved, but system complexity increases
Solution Approach 1:
Nodes perform self-discovery of their own capabilities and automatically register with the control plane. This self-service approach reduces the need for manual configuration and complex centralized management, allowing the system to achieve capability-based task allocation while minimizing the increase in system complexity
Solution Approach 2:
The control plane is designed to handle multiple functions (node discovery, capability assessment, task allocation, group management) through a unified mechanism. This multi-functionality reduces overall system complexity by consolidating management tasks rather than requiring separate systems for each function
3Ease of operation
If homogeneous node treatment is used, then ease of operation is maintained, but resource utilization efficiency deteriorates
Solution Approach 1:
The system performs preliminary discovery of node capabilities during node initialization and registration, before task allocation begins. This preliminary action automates the complexity of capability assessment, maintaining ease of operation while enabling efficient resource utilization through capability-matched task allocation
Data Source
AI summary
Technology for partitioning nodes based on capabilities in a hyper-converged infrastructure is disclosed. In an example computer system, the system detects connection of a new node element to the computer system. The system assigns the new node element an electronic address in the computer system. The computer system then boots the new node element using a pre-existing bootable image stored at the hyper-converged infrastructure system. The computer system uses a federated control plane to discover the new node element. The federated control plane determines a capability of the new node element. The federated control plane assigns the new node element to a global storage pool.