Pod Manager Node Recovery in Composable Cloud
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In composable and disaggregated cloud computing infrastructure, there is a critical need for efficient recovery solutions to prevent node failures, as failures in data centers can have significant impacts and existing technologies lack effective on-demand recovery mechanisms.
Innovation Solution
A pod manager system that reserves and allocates pooled hardware resources to build both primary and backup nodes, allowing for swift composition of backup nodes when critical nodes fail, utilizing reserved resources to ensure continuous operation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If pooled hardware resources are allocated to build primary nodes, then system productivity is improved, but system reliability deteriorates due to lack of backup capacity
Solution Approach 1:
The pod manager reserves a portion of pooled hardware resources in advance to build backup nodes before primary node failures occur. This preliminary action ensures that backup capacity is already available when needed, resolving the contradiction by maintaining both productivity (through primary nodes) and reliability (through pre-positioned backup nodes).
Solution Approach 2:
The system dynamically adjusts the allocation parameters of pooled hardware resources, partitioning them between primary and backup nodes based on reliability requirements. By changing the resource allocation parameter to include a backup portion, the system achieves both high productivity and reliability simultaneously.
2Reliability
If backup nodes are composed when primary nodes fail, then system reliability is improved, but recovery time increases due to resource allocation delays
Solution Approach 1:
The pod manager performs preliminary actions by reserving hardware resources and pre-building backup nodes before failures occur. When a primary node fails, the backup node is already constructed and ready for immediate activation, eliminating resource allocation delays and reducing recovery time while maintaining reliability.
Solution Approach 2:
The system creates backup nodes as copies of primary nodes using reserved resources. These copied backup nodes are prepared in advance, so when a primary node fails, the corresponding backup copy is immediately available, significantly reducing recovery time compared to composing backup nodes from scratch after failure.
3Reliability
If reserved pooled hardware resources are allocated for backup nodes, then system reliability is improved, but resource utilization efficiency deteriorates
Solution Approach 1:
The pod manager dynamically manages reserved pooled hardware resources, allowing them to be allocated to backup nodes when needed and released back to the pool when not required. This dynamic allocation ensures that reserved resources serve reliability purposes only when necessary, maintaining resource utilization efficiency while improving reliability through on-demand backup node composition.
Solution Approach 2:
The system recovers and releases hardware resources from backup nodes back to the pooled resources when primary nodes are restored or when backup nodes are no longer needed. This discarding and recovering mechanism ensures that reserved resources do not remain idle indefinitely, maintaining resource utilization efficiency while still providing reliability guarantees when required.
Data Source
AI summary
In an aspect of the disclosure, a method, a computer-readable medium, and an apparatus are provided. The apparatus may be a pod manager. The pod manager receives receive a request for composing a target composed-node. The pod manager employs a first set of pooled hardware resources of the computing pod to build the target composed-node. The pod manager determines to reserve a second set of pooled hardware resources of the computing pod for a backup node of the target composed-node. The pod manager determines that the target composed-node has failed. The pod manager employs the second set of pooled hardware resources to build the backup node.