Storage Controller Data Relocation Evaluation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

In tiered storage systems, data relocation based solely on access-frequency statistical values may not sufficiently enhance system performance, as the expected changes in access frequency after relocating data between high-speed and low-speed storage devices do not always occur as anticipated, leading to inefficient use of storage capacity and access speed differences.

Innovation Solution

A storage controller with a logical volume management unit, access statistics collection controller, evaluation controller, and relocation controller is implemented to monitor access frequencies and evaluate the performance impact of data relocation, determining optimal relocation of logical chunks between high-speed and low-speed storage devices based on calculated evaluation values, rather than relying solely on access frequency counts.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If data relocation is performed chunk by chunk based on access-frequency statistical values, then storage capacity and access speed differences are utilized, but system performance enhancement is insufficient

Engineering Contradiction:

Improvesystem performanceVSAvoidaccess frequency prediction accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent applies preliminary action by performing data relocation based on predicted access frequency changes rather than waiting for actual access frequency statistics to materialize. The evaluation controller predicts whether access frequency will increase or decrease after relocation, and the relocation controller executes relocation proactively based on these predictions, thereby resolving the delay between relocation and performance manifestation.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent implements feedback through the evaluation controller that continuously monitors actual access frequency changes after relocation and uses this information to refine future relocation decisions. The system compares predicted access frequency changes with actual changes, adjusting the evaluation model accordingly to improve prediction accuracy and optimize subsequent relocation operations.

Inventive Principle:

Principle #23Feedback

2Ease of operation

If access-frequency statistical values are used as relocation index, then relocation decisions are made, but expected access frequency changes do not always occur

Engineering Contradiction:

Improverelocation decision makingVSAvoidaccess frequency improvement

Core Design Contradiction:

Ease of operationVSProductivity

Solution Approach 1:

The patent replaces the mechanical system of direct access-frequency-based relocation with an evaluation-based system. Instead of mechanically relocating based on raw access frequency counts, the evaluation controller analyzes multiple factors including access frequency changes, access patterns, and storage device characteristics to generate evaluation values that better predict actual performance impact, thereby achieving more reliable relocation outcomes.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

3Productivity

If high-speed storage device capacity is fully utilized, then access performance is optimized, but storage capacity constraints are exceeded

Engineering Contradiction:

Improveaccess performanceVSAvoidstorage capacity

Core Design Contradiction:

ProductivityVSQuantity of substance

Solution Approach 1:

The patent applies parameter changes by dynamically adjusting the evaluation criteria for data relocation based on real-time storage device status. The evaluation controller monitors available capacity on high-speed and low-speed devices, and modifies relocation decisions accordingly. When high-speed device capacity is approaching limits, the system shifts its evaluation parameters to prioritize different relocation strategies, thereby optimizing access performance within capacity constraints.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10095439B2Tiered storage system, storage controller and data location estimation method

Publication Date: 2018.10.09 KK TOSHIBA

AI summary

According to one embodiment, a tiered storage system includes first and second storage devices having access speeds different from each other, and a storage controller. The storage controller manages first and second access frequency statistical values. The first access frequency statistical value corresponds to a state in which data of a respective one of logical chunks is located on the first storage device, and the second access frequency statistical value corresponds to a state in which data of the respective one is located on the second storage device. The storage controller determines whether it is necessary to change the location state of the data of the respective one, based on the first and second access frequency statistical values of the respective logical chunks.