Introduction

Disaster recovery (DR) is a critical concern for any organization, but for financial services companies, the stakes are even higher. When a regional outage occurs, clients need uninterrupted access to data, and recovery time objectives (RTO) and recovery point objectives (RPO) are often measured in minutes, not hours. S&P Global Market Intelligence faced this exact challenge with their Capital IQ platform, which supports global clients who rely on timely and accurate financial data.

In this article, we'll explore how they leveraged Amazon FSx for NetApp ONTAP to build an innovative DR solution that enables immediate failover to read-only mode in a secondary region within 15 minutes, followed by full read-write recovery when needed. This approach not only meets strict RTO/RPO requirements but also reduces costs and extends the solution's value beyond disaster scenarios.

AWS architecture diagram showing cross-region disaster recovery with FSx for NetApp ONTAP Technical Structure Concept

The Solution: SnapMirror + FlexClone

The core of S&P Global's strategy is a two-pronged approach:

  1. Immediate failover to DR in read-only mode – using ONTAP snapshots and FlexClone technology for sub-15-minute recovery.
  2. Conversion from read-only to read-write mode – following a geo-cluster design with SnapMirror replication.

How FlexClone Works

FlexClone volumes are created from existing SnapMirror snapshots in the DR region. The process is automated daily, identifying the latest snapshot and creating a clone that can be presented to a read-only SQL Server instance. Here's a typical CLI command:

# Create a FlexClone volume from the latest SnapMirror snapshot
volume clone create \
  -vserver dr-svm \
  -flexclone ciq_data_readonly \
  -parent-volume ciq_data_mirror \
  -parent-snapshot snapmirror.latest \
  -type RW

This command creates a read-write clone (though used as read-only initially) that shares data blocks with the parent volume, minimizing storage overhead.

Cloud infrastructure with data replication between two AWS regions Developer Related Image

Key Advantages and Considerations

Advantages

  • Sub-15-minute recovery: FlexClone creation completes in under 2 minutes, making failover a simple application cutover.
  • Storage efficiency: Clones share data blocks, reducing DR storage costs.
  • Operational isolation: Clones don't interfere with ongoing SnapMirror replication.
  • Beyond disasters: Read-only instances also support availability during production code releases.

Limitations and Caveats

While this solution is powerful, it's important to consider:

  • RPO variability: The actual RPO depends on production activity. During high transaction volumes, data loss could approach 15 minutes.
  • Complexity: Implementing a geo-distributed WSFC cluster across regions requires significant expertise.
  • Cost: While storage costs are reduced, the compute and networking costs for a multi-region setup can be substantial.

If you're new to AWS disaster recovery, you might find it helpful to start with AWS's official documentation to understand the basics.

Network diagram illustrating SQL Server failover cluster spanning regions Programming Illustration

Conclusion

S&P Global's implementation demonstrates that rapid DR is achievable without sacrificing cost efficiency. By combining SnapMirror replication with FlexClone technology, they built a solution that is faster, leaner, and more flexible than traditional on-premises approaches. For financial services organizations navigating cloud migration, this serves as a proven blueprint: replicate what works, modernize how it runs, and maintain the same level of data protection.

Next Steps for Learning

  • Dive deeper into Amazon FSx for NetApp ONTAP security and replication features.
  • Explore how to automate FlexClone creation using AWS Lambda or other orchestration tools.
  • Study the geo-cluster design patterns for SQL Server on AWS.

For more insights on cloud architecture and data management, check out our guide on GPU programming or this tutorial on Pandas indexing.

This content was drafted using AI tools based on reliable sources, and has been reviewed by our editorial team before publication. It is not intended to replace professional advice.