Storage Device Dynamic Load Balancing for Logical Volume Replacement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In storage devices, the existing technologies face challenges in balancing the write load across physical volumes during logical volume replacement, leading to increased load on specific physical volumes, which can result in delayed data writing and cache page release, causing errors and inefficiencies.

Innovation Solution

A storage device and method that dynamically select pairs of physical volumes for mirroring logical volumes to minimize the number of logical volumes mirrored to the same pair, allowing data from a high-load physical volume to be written to alternative volumes, thereby reducing the load on other mirrored volumes during replacement.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Speed

If data is written in parallel to multiple logical volumes through multiple channels, then writing speed is improved, but load concentration on specific physical volumes increases

Engineering Contradiction:
Improvewriting speedVSAvoidload balance
Core Design Contradiction:
SpeedVSEase of operation

Solution Approach 1:

The patent implements dynamic load balancing by continuously monitoring the load status of physical volumes and dynamically adjusting the mapping between logical volumes and physical volumes. The controller selects target physical volumes based on real-time load conditions, ensuring that data is distributed to volumes with lower current load, thereby preventing load concentration while maintaining parallel writing speed.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The system changes the parameter of volume selection from static mapping to dynamic parameter-based selection. The controller adjusts the mapping parameters based on load status, cache hit rates, and other runtime parameters, allowing the system to adapt to changing conditions and maintain balanced load distribution across physical volumes during parallel write operations.

Inventive Principle:
Principle #35Parameter changes

2Adaptability or versatility

If logical volume replacement is performed on high-load physical volumes, then storage flexibility is improved, but data writing delay increases

Engineering Contradiction:
Improvelogical volume replacementVSAvoiddata writing delay
Core Design Contradiction:
Adaptability or versatilityVSLoss of time

Solution Approach 1:

The patent performs preliminary actions by pre-selecting candidate physical volumes for logical volume replacement based on predicted future load conditions rather than current load. The controller anticipates load changes and prepares replacement targets in advance, ensuring that replacement operations are directed to volumes that will have lower load when the actual data writing occurs, thereby avoiding delays.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system implements feedback mechanisms by continuously monitoring the actual performance and load status of physical volumes after replacement operations. This feedback information is used to refine future replacement decisions and adjust the replacement strategy, ensuring that the system learns from past experiences and optimizes the timing and target selection for logical volume replacements to minimize writing delays.

Inventive Principle:
Principle #23Feedback

3Productivity

If cache pages are released based on LRU policy, then cache management efficiency is improved, but cache page release delay occurs under high load

Engineering Contradiction:
Improvecache management efficiencyVSAvoidcache page release delay
Core Design Contradiction:
ProductivityVSLoss of time

Solution Approach 1:

The patent introduces additional parameters beyond the traditional LRU (least recently used) criterion for cache page release decisions. The system considers the load status of the target physical volume, the importance of the data, and predicted access patterns when determining which cache pages to release. This multi-parameter approach allows the cache management to prioritize releases in a way that minimizes delays under high load conditions while maintaining overall cache management efficiency.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS10698820B2Storage device and storage method
Publication Date: 2020.06.30 FUJITSU LTD
  • US10698820B2 patent drawing
  • US10698820B2 patent drawing
  • US10698820B2 patent drawing

AI summary

A storage device includes a cache memory and a processor configured to perform, in accordance with an amount of dirty data stored in the cache memory, a determination of a first physical volume and a first logical volume placed in the first physical volume from the plurality of physical volumes and the plurality of logical volumes placed in the plurality of physical volumes, perform selection, from a first plurality of physical volumes in which the determined first logical volume is placed by mirroring, of a second physical volume other than the determined first physical volume, and perform selection, from the plurality of physical volumes, of a third physical volume other than the second plurality of physical volumes, and write data of the first logical volume read from the second physical volume, to the third physical volume, and delete the first logical volume from the determined first physical volume.