Hybrid Cloud Recovery Replication with Parallel Shadow Volumes
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data mirroring technologies in hybrid cloud environments face latency bottlenecks and inefficiencies in replicating I/O groups to remote sites, particularly in asynchronous mirroring, which can slow down disaster recovery processes and hinder cache space utilization.
Innovation Solution
The system splits outstanding I/O groups into multiple sets, mirrors these sets in parallel to shadow and emulated physical volumes, and merges them in the original order using snapshot and clone technology to maintain data consistency and accelerate replication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If data is mirrored sequentially to remote site, then data consistency is maintained, but replication speed is slow and latency is high
Solution Approach 1:
The patent segments outstanding I/O groups into multiple sets and creates corresponding shadow physical volumes for parallel processing. By dividing the data replication task into independent segments that can be processed simultaneously, the system achieves faster replication speeds while maintaining data consistency through ordered merging of results.
Solution Approach 2:
The patent creates shadow physical volumes in advance before the actual data replication occurs. These pre-prepared shadow volumes serve as ready-made containers for the segmented I/O groups, enabling immediate parallel writing operations without waiting for volume creation during the replication process, thus reducing overall latency.
2Productivity
If parallel mirroring is implemented, then replication speed increases, but data consistency and ordering become difficult to maintain
Solution Approach 1:
The patent introduces shadow physical volumes as intermediary structures between the source and destination storage systems. These shadow volumes act as temporary holding areas where parallel replicated data can be organized and buffered before being merged back in the correct order, thus enabling both parallel processing and data consistency.
Solution Approach 2:
The patent creates shadow copies of physical volumes that can be written to in parallel. These shadow copies are exact replicas that serve as intermediate storage during the acceleration process, allowing multiple I/O groups to be replicated simultaneously without affecting the original data consistency requirements.
3Loss of time
If traditional mirroring is used, then system complexity is low, but cache space utilization is inefficient and recovery time is long
Solution Approach 1:
The patent implements a dynamic mirroring system that can switch between normal and accelerated modes based on conditions. The system dynamically creates and manages shadow physical volumes only when acceleration is needed, and dynamically merges them back when recovery is complete. This dynamic approach reduces average recovery time while maintaining acceptable system complexity through on-demand activation.
Data Source
AI summary
A computer-implemented method may include receiving an alert indicating a need for accelerated I/O group mirroring; creating a shadow physical volume of an emulated physical volume present on an RPV server; communicating a capability of the RPV server to mirror I/O groups in parallel; mirroring the I/O groups to the shadow physical volume and the emulated physical volume present on the RPV server; receiving a merge command; and merging the shadow physical volume and the emulated physical volume.