Flexible Distributed Storage Deprovisioning via Orchestrator Demand Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Distributed storage systems in shared computing environments face challenges in scaling down storage capacity efficiently, as interlinking between nodes prevents straightforward deprovisioning, leading to performance bottlenecks and reduced deployment flexibility.
Innovation Solution
An orchestrator system measures current and projected storage demand factors, identifies underutilized storage nodes lacking active communication sessions, and selectively terminates or reconfigures them based on associations with other nodes to maintain performance and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If storage nodes are deprovisioned in distributed storage systems, then storage efficiency and resource utilization are improved, but system reliability and data availability deteriorate due to interlinking dependencies between nodes
Solution Approach 1:
The system performs preliminary actions by identifying storage nodes that lack active communication sessions before deprovisioning. The orchestrator proactively detects underutilized nodes and schedules their termination during low-impact time windows, preventing reliability issues before they occur by ensuring data is safely stored or transferred elsewhere first.
Solution Approach 2:
The system implements dynamic deprovisioning by continuously monitoring storage node utilization and adjusting the deprovisioning schedule based on real-time conditions. Nodes are deprovisioned dynamically when they show sustained low activity, while maintaining flexibility to pause or resume operations based on changing system states and performance requirements.
2Loss of energy
If storage capacity is scaled down to match demand, then resource utilization and energy efficiency are improved, but deployment flexibility and adaptability worsen due to rigid internode dependencies
Solution Approach 1:
The orchestrator implements dynamic scaling by continuously monitoring storage demand and adjusting the number of active storage nodes accordingly. When demand decreases, nodes are gradually deprovisioned to improve energy efficiency. When demand increases, the system can rapidly provision new nodes, maintaining deployment flexibility despite previous rigid dependencies.
Solution Approach 2:
The system segments the storage node population into active and deprovisioning groups, allowing independent management of each segment. This segmentation enables gradual scaling operations where individual nodes can be taken offline or brought online without affecting the entire storage system, thereby maintaining adaptability during capacity adjustments.
3Productivity
If underutilized storage nodes are terminated to improve efficiency, then resource utilization is improved, but system performance and stability worsen due to unexpected associations with active nodes
Solution Approach 1:
The orchestrator implements feedback mechanisms by continuously monitoring communication sessions and performance metrics of storage nodes. Before deprovisioning any node, the system verifies that no active communication sessions exist and that performance thresholds are met. This feedback loop prevents premature termination of nodes that are actually critical for system performance.
Solution Approach 2:
The system performs preliminary validation checks before deprovisioning storage nodes. It proactively identifies and resolves any associations with active nodes, ensures data integrity, and confirms that deprovisioning will not impact system performance. Only after these preliminary actions confirm safety does the orchestrator proceed with node termination.
Data Source
AI summary
Flexible deprovisioning of distributed storage is disclosed. For example, an orchestrator executes on a processor to measure a current storage demand factor based on a current count of service guests and a storage capacity of a plurality of storage nodes. A projected storage demand factor is calculated by (i) adjusting the current count of service guests with a timing factor resulting in a projected count, and (ii) combining the projected count with a storage class associated with the service guests. The orchestrator determines that the projected storage demand factor is lower than the current storage demand factor, and in response requests termination of a first storage node of the plurality of storage nodes based on the first storage node lacking an active communication session with the service guests. Cancel termination of the first storage node based on an association between the first storage node and a second storage node.