Cascading Engine for Instant Virtual Copy Management
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems for maintaining multiple instant virtual copies in storage environments face increased overhead due to write operations, especially as the number of images increases, which can lead to inefficiencies and capacity issues.
Innovation Solution
The cascading engine creates a dependency chain between storage volumes using intermediate volumes and maps, allowing for efficient management of instant virtual copies by minimizing writes and optimizing capacity usage, independent of the number of images.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If multiple instant virtual copies are maintained using traditional copy-on-write methods, then point-in-time copies are immediately available for read/write access, but write overhead increases significantly as the number of images increases
Solution Approach 1:
The patent segments the copy management system into two distinct components: instant virtual copies for immediate read access and a single cascaded image for write operations. This segmentation allows read-heavy workloads to access multiple point-in-time copies simultaneously while write operations are consolidated to a single target, eliminating the write overhead that traditionally increased with the number of copies.
Solution Approach 2:
The patent introduces an intermediary mechanism where a single cascaded image acts as a mediator between the source volume and multiple target volumes. Write operations are directed to this intermediate cascaded image, which then propagates changes to the appropriate target volumes. This intermediary approach decouples the write operation from the multiple targets, preventing write overhead from scaling with the number of images.
2Adaptability or versatility
If multiple instant virtual copies are maintained with separate write capabilities, then each target volume can be modified independently, but storage capacity requirements increase proportionally with the number of images
Solution Approach 1:
The patent merges the write-capable image into a single cascaded target that serves all copy relationships. Instead of maintaining separate write-capable copies, the system combines write operations into a single image that can be modified independently while still providing point-in-time access to multiple targets through the instant virtual copy mechanism.
Solution Approach 2:
The patent uses copy-on-write technology to create instant virtual copies that appear as independent, modifiable volumes but actually share underlying storage until writes occur. When writes are needed, they are directed to the single cascaded image, and the copy-on-write mechanism handles the duplication of data blocks only when necessary, dramatically reducing storage capacity requirements compared to maintaining full independent copies.
3Reliability
If traditional copy methods are used to create multiple images, then complete data redundancy is achieved, but the time and resources required to create and maintain copies increase
Solution Approach 1:
The patent performs preliminary action by creating instant virtual copies through metadata modifications before any physical data copying is needed. The system prepares the copy relationships in advance using copy-on-write semantics, so that when writes occur, the data is already in the correct location and format, eliminating the need for time-consuming physical copy operations.
Solution Approach 2:
The patent replaces the mechanical system of physical data copying with a metadata-based approach. Instead of physically moving data blocks to create copies, the system uses metadata modifications and copy-on-write semantics to create instant virtual copies. This substitution eliminates the time-consuming mechanical copying process while maintaining data redundancy through the virtual copy relationships.
Data Source
AI summary
Provided are techniques for maintaining instant virtual copies. A request to perform an instant virtual copy operation to create an instant virtual copy from a first volume to a new volume is received. It is determined that the first volume has not been modified since a last instant virtual copy operation has been performed. It is determined whether an intermediate volume and an intermediate map have already been created. In response to determining that the intermediate volume and the intermediate map have not already been created, the intermediate volume and the intermediate map are created, the intermediate volume is made dependent on the first volume in a dependency chain based on the intermediate map, and the new volume is made dependent on the intermediate volume in the dependency chain.