Data Migration Version Comparison to Filter Deleted Blocks

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

In clustered storage systems, data migration processes often result in resource wastage due to ineffective data deletion, where data remanence occurs because the source disk does not receive deletion instructions, leading to unnecessary I/O operations.

Innovation Solution

Implement a method where the first node compares version numbers of data from the second node and third nodes to filter out data for deletion, ensuring that only up-to-date data is migrated and deleted, thereby reducing resource wastage and improving migration performance and reliability.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If data is migrated from source disk to target disk without version number verification, then data migration can be completed, but data remanence occurs and I/O resources are wasted

Engineering Contradiction:
Improvedata migration efficiencyVSAvoidI/O resource consumption
Core Design Contradiction:
ProductivityVSLoss of energy

Solution Approach 1:

The patent applies preliminary action by comparing version numbers before data migration to identify and filter out data that has already been deleted or modified at the source. This pre-check prevents unnecessary migration of obsolete data, avoiding subsequent deletion operations and wasting I/O resources. The version number comparison is performed in advance to determine which data blocks need actual migration.

Inventive Principle:
Principle #10Preliminary action

2Loss of energy

If version number comparison is performed before data migration, then I/O resource consumption is reduced, but system complexity increases

Engineering Contradiction:
ImproveI/O resource consumptionVSAvoidmigration process complexity
Core Design Contradiction:
Loss of energyVSDevice complexity

Solution Approach 1:

The patent implements feedback by using version numbers as metadata that provide information about the current state of data at the source. This feedback mechanism allows the migration system to make informed decisions about whether to migrate specific data blocks. The version number comparison creates a feedback loop that guides the migration process, enabling intelligent resource allocation without significantly increasing system complexity.

Inventive Principle:
Principle #23Feedback

3Reliability

If all data is migrated regardless of deletion status, then data completeness is maintained, but resource wastage increases

Engineering Contradiction:
Improvedata completenessVSAvoidmigration resource efficiency
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent applies the extraction principle by separating data blocks into two categories: those that need migration (where source version differs from target version) and those that don't need migration (where versions match or source data is deleted). This extraction of only the necessary data blocks from the complete data set eliminates wasteful migration operations while maintaining data completeness for the data that actually needs to be transferred.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentEP3816783B1Method and device for data migration
Publication Date: 2024.04.17 HUAWEI TECH CO LTD
  • EP3816783B1 patent drawingFigure 1A~1B
  • EP3816783B1 patent drawingFigure 2A~2B
  • EP3816783B1 patent drawingFigure 3

AI summary

This application provides a data migration method and apparatus. The method includes: in a process in which a second node migrates data to a first node, reading, by the first node from the second node, to-be-migrated data and a version number of the to-be-migrated data, and reading, by the first node from at least one third node, a version number of data that belongs to a same first service as the to-be-migrated data, where the data of the first service is distributively stored in the second node and the at least one third node; and then when determining that the version number of the to-be-migrated data read from the second node is different from a version number of data read from any one of the third node, discarding, by the first node, the to-be-migrated data read from the second node. In this way, during data migration, the version number of the to-be-migrated data and a version number of data that is in another node and that belongs to a same service as the to-be-migrated data are compared, to filter out to-be-deleted data. This can reduce a waste of IO resources for ineffective migration and subsequent deletion.