Virtual Storage Volume Snapshot Data Transfer
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for increasing the replication factor of a virtual storage volume in containerized storage systems cause significant disruption by requiring freezing access and lengthy data copying processes, which is unacceptable in high-bandwidth environments.
Innovation Solution
The solution involves a two-phase process: a transfer phase where data is iteratively copied from the virtual storage volume to a newly added replica using snapshots, and a resynchronization phase where recently changed data is synchronized, allowing for increased replication factor without blocking access.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional methods are used to increase replication factor, then data redundancy is improved, but system availability deteriorates due to access freezing and lengthy copying processes
Solution Approach 1:
The patent applies preliminary action by creating snapshots of the source virtual storage volume before initiating data transfer. These snapshots capture data blocks at specific points in time, allowing the replication process to proceed without freezing access to the source volume. The snapshots are created in advance of the actual data copying operation, enabling continuous access while replication occurs.
Solution Approach 2:
The patent segments the data transfer process into discrete data blocks that are identified and transferred individually or in batches. Instead of copying the entire volume at once, the system divides the data into manageable blocks, transfers them progressively, and allows continued access to the source volume throughout the process. This segmentation enables parallel operations without mutual exclusion.
2Reliability
If data copying is performed to increase replication factor, then data redundancy is improved, but processing time deteriorates due to lengthy copying processes
Solution Approach 1:
The patent implements periodic action by creating snapshots at regular intervals or at specific checkpoints during the replication process. Instead of performing a single lengthy copy operation, the system periodically captures data blocks and transfers them in multiple shorter operations. This periodic approach reduces the time any single operation needs to complete and allows the system to remain responsive throughout the replication process.
Solution Approach 2:
The patent applies preliminary action by pre-identifying and pre-transferring data blocks that will be needed for replication. By using snapshots to capture data blocks before they are potentially modified or deleted, the system prepares the replication data in advance, allowing the actual copy operation to proceed more quickly without requiring access freezing or lengthy blocking operations.
3Stability of the object's composition
If access freezing is implemented during replication, then data consistency is improved, but system performance deteriorates due to blocked operations
Solution Approach 1:
The patent applies copying by creating snapshots that capture copies of data blocks at specific points in time. Instead of requiring exclusive access during replication, the system creates point-in-time copies of the necessary data blocks through snapshots. These copies can then be transferred and processed without interfering with ongoing operations on the source volume, maintaining both consistency and performance.
Solution Approach 2:
The patent introduces snapshots as an intermediary mechanism between the source volume and the replication process. The snapshots act as a mediator that captures the state of data blocks without requiring the source volume to be locked or frozen. This intermediary layer allows the replication process to proceed independently while the source volume remains accessible, eliminating the need for access freezing while maintaining data consistency.
Data Source
AI summary
An example method includes transferring, for each of a plurality of snapshots of a source virtual storage volume mounted at a first compute node, at least a portion of a plurality of data blocks for each of the snapshots to a target virtual storage volume at a second compute node; and after the data blocks are transferred, resynchronizing the target virtual storage volume with the source virtual storage volume.