Storage Extent Relocation via Backend Connection Properties
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing storage systems fail to optimally relocate extents within the same tier during load balancing due to the lack of consideration for backend connection properties, such as interconnect bandwidth and protocol overhead, leading to suboptimal performance and potential performance degradation.
Innovation Solution
A storage system evaluates backend connection properties like interconnect bandwidth, protocol I/O stack overhead, and additional interconnect capabilities to perform optimal load balancing across physical storage devices, using rebalancing and tiering daemons to adjust threshold values and relocate extents based on these properties.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If load balancing relocates extents between physical storage devices within the same tier without considering backend connection properties, then the storage system can simplify the load balancing mechanism and reduce complexity, but the performance optimization is insufficient and I/O latency increases
Solution Approach 1:
The patent changes the parameters considered in load balancing from simple utilization metrics to include backend connection properties such as interconnect bandwidth, protocol overhead, and connection speed. This allows the system to optimize extent relocation decisions based on actual connection characteristics, reducing I/O latency without significantly increasing mechanism complexity.
Solution Approach 2:
The patent implements a feedback mechanism where the storage system continuously monitors backend connection properties and uses this information to dynamically adjust load balancing decisions. The system evaluates current connection states and relocates extents to optimize performance based on real-time feedback about interconnect conditions.
2Productivity
If the storage system evaluates backend connection properties for each extent relocation decision, then the performance optimization improves, but the computational overhead and processing time increase
Solution Approach 1:
The patent performs preliminary evaluation of backend connection properties and pre-identifies optimal destination storage devices before actual extent relocation is needed. By pre-assessing connection characteristics and maintaining updated information about backend capabilities, the system reduces the computational overhead during actual relocation decisions, improving both optimization quality and decision speed.
3Measurement precision
If the storage system considers multiple backend connection properties including interconnect bandwidth and protocol overhead, then the extent relocation accuracy improves, but the measurement and evaluation complexity increases
Solution Approach 1:
The patent creates a universal evaluation framework that handles multiple different backend connection properties through a single standardized assessment mechanism. The system uses a unified approach to evaluate diverse connection characteristics (interconnect bandwidth, protocol overhead, connection speed) using consistent methods, simplifying the measurement process while maintaining high relocation decision accuracy.
Data Source
AI summary
An approach is provided in which a storage system detects that an extent residing on a first one of a set of physical storage devices requires relocation. The storage system identifies a set of backend connection properties of each of a set of backend connections between the storage system and the set of physical storage devices. The set of backend connection properties includes at least a connection bandwidth between the storage system and at least one of the physical storage devices. In turn, the storage system relocates the extent from the first physical storage device to a second one of the set of physical storage devices based at least in part, on the set of backend connection properties.