Multi-site Point-in-time Copy for Data Consistency
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current disaster recovery systems face challenges in maintaining data consistency and availability across multiple remote sites, particularly in gradual disasters where data updates may be lost, and existing mirroring technologies do not ensure instantaneous recovery from failures at multiple locations.
Innovation Solution
A system and method for maintaining mirror and point-in-time copies of volumes across multiple remote sites, where a first server initiates a point-in-time copy operation and transmits commands to secondary servers to create consistent copies, allowing for instantaneous creation of data structures without initial data transfer, and subsequent deferred data replication with minimal system impact.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If data is mirrored to multiple remote sites using traditional copy technologies, then data availability and disaster recovery capability are improved, but data consistency across sites deteriorates due to gradual disasters and lost updates
Solution Approach 1:
The system performs preliminary actions by creating point-in-time copies at multiple remote sites before a disaster occurs. The first server initiates copy operations to multiple sites simultaneously, establishing consistent baseline copies in advance. This preliminary action ensures that when a disaster strikes, pre-synchronized copies are already available at all sites, preventing data consistency loss without requiring real-time synchronization during the disaster event.
2Reliability
If traditional mirror copy operations are performed across multiple sites, then data redundancy is improved, but system performance and productivity deteriorate due to extensive data transfer requirements
Solution Approach 1:
The system performs preliminary copy operations to establish point-in-time copies at multiple remote sites before production use. By completing the data transfer and synchronization in advance during non-critical periods, the system eliminates the need for extensive real-time data transfers that would degrade performance. The preliminary copies are then used for disaster recovery, maintaining high productivity during normal operations.
Solution Approach 2:
The system creates point-in-time copies of data at multiple remote sites rather than maintaining continuous real-time synchronization. This copying approach allows discrete data transfers at controlled times, avoiding the performance degradation associated with continuous data mirroring. The copies are sufficient for disaster recovery purposes without requiring constant updates, thus preserving system productivity.
3Speed
If point-in-time copy operations are initiated from a first site to multiple remote sites, then recovery speed is improved, but device complexity increases due to coordination requirements across multiple servers
Solution Approach 1:
The system segments the disaster recovery function by designating a first server as the primary initiator while allowing second servers at remote sites to independently process copy operations. This segmentation enables parallel execution of copy operations to multiple sites simultaneously, improving recovery speed. The complexity is distributed across sites rather than centralized, with each server handling its own copy operations independently after receiving the initial command.
Data Source
AI summary
Provided is a method for maintaining mirror and storage system copies of volumes at multiple remote sites. A first server maintains a mirror copy relationship between a first storage system at a first site and a second storage system at a second site. The first server performs a first point-in-time copy operation from the first storage system to a first storage system copy, wherein the data for the first storage system copy is consistent as of the determined point-in-time. The first server transmits a command to a second server to create a point-in-time copy of the second storage system. The second server processes mirror data transferred from the first server as part of the mirror copy relationship to determine when to create a second point-in-time copy. The second server performs the second point-in-time copy operation.