Cascade Replication via Consistency Groups and Snapshots
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data replication methods face challenges in maintaining consistency across multiple storage systems, particularly when involving more than two systems, due to issues related to data synchronization and replication modes.
Innovation Solution
The implementation of cascade replication techniques, where a first replication session is initiated to replicate data from a first consistency group to a second, creating an additional consistency group linked to the second, and a second replication session is initiated to replicate from this additional group to a third consistency group, ensuring consistency through periodic updates using snap-and-reassign operations and active snapshot sets.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional data replication methods are used across multiple storage systems, then data can be replicated between two systems, but maintaining consistency across more than two systems becomes difficult
Solution Approach 1:
The patent segments the multi-system replication problem into multiple bilateral replication relationships. Each storage system participates in separate replication sessions with its partners, breaking down the complex N-system consistency problem into manageable pairwise interactions. This segmentation allows each replication session to maintain consistency independently while collectively achieving cascade replication across multiple systems.
Solution Approach 2:
The patent introduces consistency groups as intermediary structures that mediate data replication between storage systems. These consistency groups act as logical containers that track and manage data state across systems, enabling indirect consistency maintenance through intermediate coordination layers rather than direct peer-to-peer synchronization.
2Reliability
If synchronous replication is used to ensure data consistency, then data is mirrored in real-time, but replication performance and speed are reduced
Solution Approach 1:
The patent implements dynamic replication mode selection that adapts between synchronous and asynchronous operations based on operational context. The system can switch between real-time mirroring and periodic updates, allowing optimization of replication speed when absolute consistency is not immediately required, while maintaining reliability when needed.
Solution Approach 2:
The patent employs periodic updates through snapshot-based replication where data is replicated at scheduled intervals rather than continuously. This periodic action reduces the overhead of constant synchronization while maintaining acceptable consistency levels, improving overall replication performance by eliminating unnecessary real-time synchronization operations.
3Productivity
If asynchronous replication is used to improve replication speed, then data is transferred periodically, but data consistency between systems deteriorates
Solution Approach 1:
The patent uses preliminary snapshot creation before data transfer to establish a known consistent state. By capturing snapshots of data before replication, the system ensures that even asynchronous transfers start from a verified consistent baseline, preventing consistency deterioration while maintaining the performance benefits of periodic updates.
Solution Approach 2:
The patent implements feedback mechanisms through consistency group tracking and snapshot validation that monitor replication status and detect inconsistencies. This feedback allows the system to identify and correct consistency issues that arise during asynchronous replication, maintaining reliability while benefiting from improved replication speed.
4Ease of repair
If storage volumes are refreshed or restored during ongoing replication, then data recovery is enabled, but the replication process must be terminated and restarted
Solution Approach 1:
The patent enables continuous replication operation during volume restoration activities by using snapshot-based updates. The replication process continues uninterrupted by applying snapshots before and after restoration operations, eliminating the need to terminate and restart replication while still allowing volume refresh and restore activities to proceed.
Solution Approach 2:
The patent creates protective snapshots before restoration operations to cushion against potential data loss or inconsistency. These pre-restoration snapshots serve as safety buffers that allow the system to recover from restoration operations without interrupting the overall replication process, maintaining continuity while enabling repairs.
Data Source
AI summary
An apparatus is configured to initiate a first replication session to replicate data of a first consistency group in a first storage system to a second consistency group in a second storage system, to create an additional consistency group linked to the second consistency group in the second storage system, and to initiate a second replication session to replicate data of the additional consistency group to another consistency group in a third storage system. The additional consistency group linked to the second consistency group in some embodiments is periodically updated against the second consistency group. For example, in one or more embodiments the second consistency group is updated based at least in part on an active snapshot set of the first replication session, and the additional consistency group is updated based at least in part on the second consistency group.