Storage Object Migration via Parent Vdisk Ancestry
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for migrating storage objects between storage servers, such as vdisks, face inefficiencies due to reliance on unique fingerprints and versioning approaches, which can lead to redundant data transfer and incorrect block transmission, resulting in decreased network bandwidth and increased storage capacity requirements.
Innovation Solution
A novel migration system that determines a common parent vdisk available at the destination, transferring only the changed blocks between the child and parent vdisks, thereby avoiding redundant data transfer and ensuring accurate block transmission by leveraging the cloned relationship between vdisks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If versioning approaches are used to determine common data between storage objects, then storage capacity can be optimized, but incorrect block transmission may occur and reliability decreases
Solution Approach 1:
The patent uses cloning to create a parent indicator that points to the source storage object. This copying approach allows the destination to identify common data blocks through the parent-child relationship without relying on versioning, thereby maintaining reliability while optimizing storage capacity.
2Measurement precision
If unique fingerprints are used to identify storage blocks, then measurement precision improves, but fingerprint collisions can occur leading to incorrect block identification
Solution Approach 1:
The patent introduces a parent indicator as an intermediary element that establishes a hierarchical relationship between storage objects. This parent indicator serves as a mediator that allows the destination to identify common blocks through the ancestry relationship, eliminating the need for fingerprint comparisons and preventing fingerprint collision errors.
3Reliability
If all blocks are transferred during migration, then reliability is maintained, but network bandwidth is wasted due to redundant data transfer
Solution Approach 1:
The patent extracts only the changed blocks from the source storage object by comparing the child object's blocks with its parent's blocks. The destination uses the parent indicator to identify which blocks are already present and only requests the differences, thereby eliminating redundant data transfer while maintaining data integrity.
4Quantity of substance
If fingerprint-based deduplication is used, then storage efficiency improves, but incorrect blocks may be transmitted due to fingerprint collisions
Solution Approach 1:
The patent copies the parent indicator from the source storage object to enable the destination to identify common blocks through the hierarchical relationship. This copying mechanism provides a reliable method for deduplication that does not depend on fingerprint matching, thereby maintaining storage efficiency while eliminating the risk of fingerprint collision errors.
Data Source
AI summary
Embodiments of the present invention provide a method and system, in a network storage system, for efficiently migrating a storage object (vdisk) from a first storage server (source) to a second storage server (destination) using data of a related vdisk at the destination. A related vdisk includes an earlier created vdisk (parent) from which the vdisk to be migrated (child) is generated (cloned). By virtue of the cloned relationship, the parent and child have a portion of data blocks in common. When a parent is available at the destination, only the changed data blocks between the parent and child need to be migrated for efficiently constructing the child at the destination.