Dynamic Path Selection for Storage System Load Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing multipath technologies in storage systems face challenges in effectively managing bandwidth saturation and load imbalance, particularly with the introduction of solid-state disk (SSD) enclosures, which can lead to performance bottlenecks and resource inefficiencies due to lack of dynamic congestion control and system-wide performance monitoring.
Innovation Solution
A method and apparatus for dynamic path selection in storage systems that monitor system-wide performance metrics, detect performance-related events, and reroute workloads across multiple HBA ports based on real-time statistics to balance load and avoid bottlenecks, using a combination of performance monitoring, event detection, and path selection algorithms.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If SSD enclosures are introduced into a chain to increase throughput, then system performance is improved, but bandwidth saturation occurs more quickly and HDD enclosures may experience starvation
Solution Approach 1:
The patent implements a feedback mechanism that continuously monitors bandwidth consumption and performance metrics of different storage devices in the chain. Based on this feedback, the system dynamically adjusts congestion control parameters and path selection to prevent SSDs from monopolizing bandwidth and to ensure fair resource allocation to HDDs, thereby resolving the contradiction between high throughput and resource allocation fairness
Solution Approach 2:
The system dynamically changes operational parameters such as congestion control thresholds, weight factors for different device types, and path selection criteria based on real-time performance conditions. This allows the system to optimize throughput when resources are abundant while preventing starvation when bandwidth is constrained, effectively managing the trade-off between productivity and reliability
2Quantity of substance
If multiple HBA ports are used to increase storage capacity, then system capacity is improved, but load imbalance and performance bottlenecks occur without dynamic path selection
Solution Approach 1:
The patent implements dynamic path selection that continuously monitors performance metrics of multiple HBA ports and storage devices, then adaptively routes I/O operations through optimal paths. This dynamic adjustment ensures balanced load distribution across all available resources, preventing bottlenecks while utilizing the full storage capacity provided by multiple HBA ports, thereby resolving the contradiction between quantity and productivity
3Ease of operation
If static path selection is used to simplify system operation, then ease of operation is improved, but bandwidth saturation and load imbalance cannot be effectively managed
Solution Approach 1:
The system implements self-service through automated performance monitoring, event detection, and dynamic path selection algorithms that operate without manual intervention. The multipath module automatically adjusts routing decisions based on real-time conditions, maintaining ease of operation while achieving optimal resource utilization efficiency through intelligent, adaptive resource management
Data Source
AI summary
Embodiments of the present disclosure provide methods and apparatus for path selection in a storage system. The storage system includes a plurality of storage devices, each storage device being connected to a plurality of different HBA ports of a server via a plurality of paths; the method comprises: monitoring a system-wide performance metric of the storage system to obtain dynamic performance statistics; detecting a performance-related event based on the dynamic performance statistics; and selecting, from the plurality of paths, an active path for a storage device of the plurality of storage devices based on the dynamic performance statistics and a result of the detecting. With the methods or apparatus according to the embodiments of the present disclosure, performance bottleneck can be avoided or timely eliminated, load balance can be achieved, and system resources can be utilized more effectively.