Asynchronous Remote Snapshot Replication Consistency Window
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for remotely replicating snapshots between storage systems require significant manual effort and resource consumption, leading to performance degradation, especially with high write activity and frequent snapshot creation.
Innovation Solution
A method for asynchronously replicating snapshots from a first storage system to a second storage system, involving recording write operations, suspending new write requests during snapshot creation, and resuming processing after replication, with indications to replicate snapshots across systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If manual snapshot replication methods are used, then replication can be performed, but significant manual effort and resource consumption occur leading to performance degradation
Solution Approach 1:
The storage system automatically manages snapshot replication without requiring manual intervention. The system self-manages the replication process by detecting when snapshots need to be replicated, coordinating with remote storage systems, and handling the replication operations autonomously, thereby eliminating manual effort while maintaining performance
Solution Approach 2:
The system prepares for replication by pre-establishing replication relationships between storage systems and pre-allocating resources. Snapshot metadata is prepared in advance with replication instructions, allowing the actual replication to occur efficiently without manual coordination when needed
2Reliability
If frequent snapshot creation is performed, then data protection is improved, but resource consumption increases and performance degrades
Solution Approach 1:
The system changes the parameters of snapshot replication by using asynchronous replication mode, where snapshots are replicated without blocking write operations. The replication timing and methodology are adjusted to occur in the background, reducing resource consumption during snapshot creation while maintaining data protection through frequent snapshots
Solution Approach 2:
Write operations continue uninterrupted during snapshot replication. The system maintains continuous useful action by allowing write requests to proceed while snapshot replication occurs asynchronously in the background, preventing resource consumption spikes that would occur with synchronous replication
3Reliability
If synchronous replication is used, then data consistency is maintained, but write operations are blocked and performance decreases
Solution Approach 1:
The system dynamically adjusts replication timing based on operational needs. Instead of blocking writes for synchronous replication, the system uses asynchronous replication where write operations proceed dynamically without blocks, and replication occurs when appropriate without interfering with ongoing write operations
Solution Approach 2:
The system performs replication preparation in advance by recording write operations during replication cycles and pre-positioning snapshot metadata with replication instructions. This preliminary action allows replication to occur without blocking current write operations, maintaining both consistency and performance
Data Source
AI summary
Snapshots may be remotely replicated asynchronously from a first LSU (R1) on a first storage system (A) to a second replica LSU (R2) on a second storage system (A2). The storage system A1 may open a consistency window to suspend initiating processing of new write operations received on A1. While the consistency window is open, A1 may: take a first snapshot, SS11, of R1; record, in association with the first replication cycle, an indication to replicate SS11 on A2; and initiate a next replication cycle to record write operations of the next new write requests to be received from hosts. After initiating a next replication cycle, A1 may close the consistency and transmit the first replication cycle to A2. A2 may apply the write operations of the first replication cycle to R2, and then take a second snapshot SS12 of R2, which should be a replica of SS11.