Storage System Virtual Volume Grouping for Tier Relocation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing storage systems face performance issues due to files with high speed requirements being migrated to lower-tier drives with low access frequencies, leading to delayed application responses and suboptimal access performance.

Innovation Solution

Virtual volumes are grouped and relocation conditions are set based on required response performance, including maximum/minimum tier usage, I/O frequency weighting, and targeted average response time, to optimize tier allocation and reduce performance degradation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If dynamic tiering migrates pages based solely on I/O access frequency, then storage cost is optimized by placing low-access pages on lower-tier drives, but response performance deteriorates when files requiring high speed are incorrectly placed on low-speed drives

Engineering Contradiction:
Improvestorage cost optimizationVSAvoidresponse performance
Core Design Contradiction:
Quantity of substanceVSReliability

Solution Approach 1:

The patent segments the file into two parts: the file header storing metadata including required response performance level, and the file body containing actual data. This segmentation enables the system to distinguish between files with different performance requirements and make appropriate placement decisions for each part, resolving the contradiction between cost optimization and performance reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by assigning different storage tiers to different parts of the file based on their specific requirements. The file header with high I/O frequency is placed on high-speed drives to ensure quick access to metadata, while the file body with lower I/O frequency can be placed on lower-cost, lower-speed drives. This localized differentiation resolves the contradiction by ensuring performance where needed without sacrificing cost optimization elsewhere.

Inventive Principle:
Principle #3Local quality

2Reliability

If real data of files requiring high speed response is allocated to upper-tier drives, then response performance is satisfied, but the upper-tier drive capacity is continuously increased leading to higher storage costs

Engineering Contradiction:
Improveresponse performanceVSAvoidupper-tier drive capacity
Core Design Contradiction:
ReliabilityVSQuantity of substance

Solution Approach 1:

By segmenting the file into header and body, the system can place only the header (which requires high-speed access due to frequent I/O operations) on upper-tier drives, while placing the larger file body on lower-tier drives. This segmentation prevents unnecessary consumption of expensive upper-tier capacity while maintaining performance for critical operations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by differentiating storage requirements within the file itself. The header, requiring high-speed access for metadata operations, is placed on high-speed upper-tier drives, while the file body, accessible at lower speeds, is placed on cost-effective lower-tier drives. This resolves the contradiction by optimizing each part according to its specific performance needs rather than placing the entire file uniformly.

Inventive Principle:
Principle #3Local quality

3Reliability

If pages are migrated to upper-tier drives to satisfy response performance requirements, then access performance is improved, but the complexity of managing tier allocation increases

Engineering Contradiction:
Improveaccess performanceVSAvoidtier allocation management
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent incorporates required response performance level information directly into the file header during file creation or initialization. This preliminary action stores the performance requirement metadata before the file is written to storage, eliminating the need for complex real-time analysis and decision-making during I/O operations. The system simply reads the pre-stored requirement and makes automatic placement decisions, significantly reducing management complexity while maintaining performance.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP2864862B1Storage system and storage control method
Publication Date: 2020.01.01 HITACHI LTD
  • EP2864862B1 patent drawingFigure 1
  • EP2864862B1 patent drawingFigure 2(a)~2
  • EP2864862B1 patent drawingFigure 3

AI summary

When a plurality of applications operating in a host computer or a storage system uses a same pool, actual data constantly allocated to an upper tier drive satisfying the required response performance will continue to increase. Therefore, when data is migrated by IOPS in page units, the access performance such as the response speed may not be satisfied. Therefore, according to the present invention, multiple virtual volumes used in applications are subjected to grouping. Relocation conditions corresponding to performances required for each group are set in advance, and reallocation of tiers according to the relocation conditions is performed.