Storage System I/O Statistics Comparison for Dynamic Configuration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data storage systems often suffer performance issues due to misalignment between actual application behavior and selected storage services, leading to suboptimal performance and customer dissatisfaction, as users manually configure storage resources without accurately matching expected behaviors.
Innovation Solution
The data storage system automatically adjusts operating settings, such as tiering policies and RAID levels, based on comparisons between observed and expected I/O statistics, to align with actual application behaviors, including adjusting prefetching, compression, and deduplication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If human users manually configure storage services to match expected application behaviors, then the configuration can be customized for specific use cases, but the actual performance may suffer when actual application behaviors differ from expectations
Solution Approach 1:
The storage system automatically monitors its own I/O statistics and compares them with expected profiles to detect misalignments. The system self-diagnoses performance issues and triggers remedial actions without human intervention, allowing it to adapt to actual application behaviors dynamically rather than relying on static manual configuration
Solution Approach 2:
The system continuously collects I/O statistics feedback from actual application operations and compares this feedback against expected behavior profiles. This closed-loop feedback mechanism enables the system to detect when actual usage diverges from expectations and automatically adjust configurations to maintain optimal performance
2Productivity
If the storage system automatically adjusts operating settings based on observed I/O statistics, then performance can be optimized for actual application behaviors, but the system complexity increases
Solution Approach 1:
The system optimizes performance by dynamically changing operating parameters such as RAID level, tiering policy, prefetching settings, compression, and deduplication based on observed I/O statistics. Rather than redesigning the entire system architecture, the invention focuses on adjusting specific controllable parameters to adapt to different application behaviors
Solution Approach 2:
The patent introduces an intermediary layer consisting of expected I/O statistics profiles that mediate between the complex storage subsystem and the application layer. These profiles serve as a simplified interface, allowing the system to compare actual behavior against predefined patterns and trigger appropriate remedial actions without requiring complex real-time analysis of every I/O operation
3Reliability
If RAID 6 is configured for fault tolerance, then data protection is improved, but write performance deteriorates due to compute intensity
Solution Approach 1:
The system dynamically selects between different RAID levels based on observed application behavior patterns. For applications exhibiting sequential write patterns, the system can switch from RAID 6 to RAID 1, adapting the fault tolerance mechanism to match actual usage characteristics rather than applying a static configuration
4Quantity of substance
If data is placed on low performance storage tiers to reduce cost, then storage capacity utilization is improved, but access time increases when applications require fast response
Solution Approach 1:
The system dynamically adjusts tiering policies based on observed I/O statistics. When applications exhibit patterns requiring fast response times, the system can automatically adjust to prefer placing data on higher performance tiers, reversing the static cost-optimization approach to match actual performance requirements
Data Source
AI summary
A technique is directed to performing a tuning operation in data storage equipment. The technique involves generating, while the data storage equipment performs input/output (I/O) transactions, an observed I/O statistics profile based on performance of at least some of the I/O transactions. The technique further involves performing a comparison operation that compares the observed I/O statistics profile to an expected I/O statistics profile which is defined by a set of operating settings that controls operation of the data storage equipment. The technique further involves operating the data storage equipment in a normal state when a result of the comparison operation indicates that the observed I/O statistics profile matches the expected I/O statistics profile and in a remedial state which is different from the normal state when the result of the comparison operation indicates that the observed I/O statistics profile does not match the expected I/O statistics profile.