Storage Performance Management via Predictive Data Migration
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional storage apparatus management systems do not systematically address future performance issues arising from flash drive devices reaching their limits, focusing only on ex-post facto measures based on current configurations of flash memory and hard disk drives.
Innovation Solution
A performance management device and method that creates a virtual volume using multiple storage mediums with different rewriting limits and response performances, migrating data to unused storage areas with higher or lower response performance based on IO load, and estimating storage capacity decrements and performance influences to display timely migration plans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is migrated based only on current usage status thresholds, then power consumption is reduced and performance is maintained, but future performance issues are not proactively addressed
Solution Approach 1:
The patent applies preliminary action by estimating future performance influences and degradation timings before they actually occur. The performance influence estimation unit calculates when flash drives will reach their rewriting limits and when performance degradation is likely to happen, enabling proactive data migration before actual performance issues arise. This transforms the reactive threshold-based migration into a predictive system that acts in advance.
Solution Approach 2:
The patent implements feedback by continuously monitoring usage status, estimating future performance degradation, and using this information to guide data migration decisions. The performance influence display unit provides feedback to administrators about estimated degradation timings, enabling informed decisions about when to migrate data to prevent future performance issues.
2Reliability
If the number of drives is increased in advance to prevent future limits, then future performance issues are prevented, but device complexity and cost increase
Solution Approach 1:
Instead of physically adding drives in advance, the patent uses preliminary estimation of performance degradation to plan capacity expansion timing. The performance influence estimation unit calculates when current drives will reach their limits, allowing administrators to plan capacity additions at optimal times rather than reacting to emergencies or over-provisioning in advance.
Solution Approach 2:
The patent makes the storage system more versatile by supporting both flash drives and hard disk drives in a unified managed environment. The system can dynamically allocate data between different storage medium types based on their characteristics and current/future status, making the storage infrastructure more adaptable without requiring separate management systems for different drive types.
3Speed
If data is migrated frequently between storage mediums, then performance is optimized, but system complexity and energy consumption increase
Solution Approach 1:
The patent reduces unnecessary migrations by using preliminary performance influence estimation. Instead of migrating data based solely on current threshold crossings, the system estimates future degradation and migrates data proactively before performance issues occur. This reduces the frequency of reactive migrations while maintaining performance optimization.
Solution Approach 2:
The patent changes the migration decision parameters from simple threshold-based current status to include estimated future performance influence and degradation timing. This allows the system to make more informed migration decisions that balance performance optimization with energy consumption, migrating data based on predicted needs rather than reactive threshold crossings.
Data Source
AI summary
Proposed are a method and a device for managing the performance of a storage apparatus which uses a combination of different types of drive devices which mutually have different limits as its storage medium so that measures can be systematically taken against problems that may arise in the future as a result of the drive devices reaching their limit. Based on a premise of being able to tier two or more types of storage mediums which have mutually different rewriting limits and response performances and migrate data between the respective tiers according to the response performance, a performance management device estimates a decrement of a storage capacity pursuant to each of the two or more types of storage mediums reaching a rewriting limit, estimates a performance influence on the virtual volume and a timing of occurrence, and displays information related to the estimated performance influence and the timing of occurrence.