Resource Allocation Mechanism for Clustered Storage Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In clustered network storage systems, resource over-commitment occurs due to simultaneous requests for storage efficiency operations exceeding the limited processing resources of data modules, leading to degraded system performance.
Innovation Solution
Implementing a resource allocation mechanism that allows conditional access to storage server node resources based on credits and prioritization, using a job submission module, job scheduler, and job queue to manage and schedule jobs effectively, preventing over-commitment and ensuring efficient resource utilization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If remote management is enabled to allow administrators to manage storage efficiency resources from any node, then ease of operation is improved, but resource over-commitment occurs when multiple simultaneous requests exceed processing capacity
Solution Approach 1:
The system performs preliminary actions by checking credit availability before allowing remote management operations. The job submission module verifies whether credits are available on the target node before submitting storage efficiency jobs, preventing over-commitment of processing resources while maintaining remote management capability.
Solution Approach 2:
The patent introduces an intermediary mechanism using credits as a mediator between remote management requests and actual resource allocation. Credits act as a buffer that mediates between the desire for remote management access and the limitation of processing capacity, allowing controlled resource allocation without over-commitment.
2Quantity of substance
If storage efficiency operations are performed to improve storage space efficiency, then storage space utilization is improved, but processing resources on data modules become overloaded
Solution Approach 1:
The system performs preliminary verification of credit availability before allowing storage efficiency operations to be submitted. This preliminary action ensures that processing resources are not overloaded by accepting more operations than the data module can handle simultaneously.
Solution Approach 2:
The patent changes the parameter of resource allocation control by introducing credit-based limits. Instead of allowing unlimited storage efficiency operations, the system uses credit parameters to control and limit the number of simultaneous operations, balancing storage efficiency improvements with processing capacity constraints.
3Adaptability or versatility
If multiple administrators simultaneously request storage efficiency operations, then management flexibility is improved, but system performance is degraded due to resource over-commitment
Solution Approach 1:
The system performs preliminary credit availability checks before allowing multiple administrators to submit storage efficiency operations. This preliminary action prevents resource over-commitment while maintaining the flexibility for multiple administrators to manage storage efficiently.
Solution Approach 2:
The patent implements feedback mechanisms where the job submission module receives feedback about credit availability and adjusts its behavior accordingly. When credits are unavailable, the system provides feedback to administrators that operations cannot be submitted, preventing resource over-commitment while maintaining management flexibility.
Data Source
AI summary
A clustered network storage system includes a management module and a resource allocation mechanism configured to allow remote management of node resources from the management module while preventing resource over-commitment due to the remote management. Preventing resource over-commitment includes conditionally granting access to the resources of a node according to a credit based system. Preventing resource over-commitment further includes scheduling jobs that access the resources of the node according to a prioritization scheme.