AWS Architecture Blog·July 7, 2026Multi-Region SQL Server Disaster Recovery with AWS FSx for NetApp ONTAP
S&P Global implemented an innovative multi-region disaster recovery solution for their Capital IQ platform using Amazon FSx for NetApp ONTAP. This strategy leverages SnapMirror replication for cross-region data consistency and FlexClone technology for rapid, sub-15-minute read-only failover, followed by a full read-write recovery process. The solution addresses critical RTO/RPO requirements for financial data while optimizing costs.
Read original on AWS Architecture BlogThis article details S&P Global Market Intelligence's approach to building a robust disaster recovery (DR) strategy for their critical SQL Server infrastructure, leveraging AWS services, specifically Amazon FSx for NetApp ONTAP. The primary goal was to achieve a rapid Recovery Time Objective (RTO) of under 15 minutes for read-only access to financial data during regional outages, alongside maintaining strict Recovery Point Objectives (RPO).
Two-Pronged DR Strategy
The solution employs a two-pronged strategy to balance immediate availability with full recovery capabilities:
- Immediate Read-Only Failover: Utilizes ONTAP snapshots and FlexClone technology for sub-15-minute recovery of a read-only instance.
- Full Read-Write Recovery: Transitions the DR system from read-only to read-write mode using established geo-cluster design with SnapMirror replication, allowing for complete operational resumption.
Architecture Components
The architecture comprises several key layers across two AWS Regions (US-East-1 as primary, US-West-2 as DR):
- Compute Layer: A four-node geo-distributed Windows Server Failover Cluster (WSFC) spans both regions, hosting Microsoft SQL Server instances on Amazon EC2.
- Storage Layer: Two Amazon FSx for NetApp ONTAP file systems, one in each region, serve as the primary and DR storage.
- Data Replication: SnapMirror replication is configured from the primary to the DR FSx for ONTAP file system, with 15-minute intervals to maintain data consistency.
- Rapid Recovery: FlexClone volumes are created from the most recent SnapMirror snapshots in the DR region, providing a quickly accessible, read-only data instance.
FlexClone for Rapid Recovery
FlexClone technology is crucial for achieving the sub-15-minute RTO. By creating a thin, writable clone from a point-in-time snapshot, a read-only instance of the database can be made available almost instantly. This minimizes storage overhead as FlexClones share data blocks with the parent volume, only consuming additional space for changed blocks. This enables pre-provisioning of DR instances, transforming failover into a simple application cutover rather than a lengthy data restoration process.