AWS re:Invent 2025 - How Zydus implemented regulated DR on AWS (NTA320)

Zydus' Journey to Implementing Regulated Disaster Recovery on AWS

Overview of Zydus

• Zydus is a global pharmaceutical company headquartered in India
• Founded in 1952, it has grown to over 29,000 employees worldwide
• Operates in multiple countries including the US, Europe, Brazil, and Mexico
• Has 41 manufacturing plants, 16 of which are US FDA approved
• Focuses on developing new chemical entities (NCEs), biosimilars, vaccines, and niche technologies

Zydus' Disaster Recovery Challenges

• Previously had high RPO (24 hours) and RTO (24-48 hours) for their critical applications
• This resulted in long recovery times and potential regulatory compliance issues
• Significant cost and operational overhead to maintain on-premises disaster recovery sites
• Security and skills challenges in managing multiple disaster recovery sites

AWS Disaster Recovery Service (DRS) Solution

• Zydus sought a solution that could provide:
  • Minimal RPO (15 minutes) and RTO (< 3 hours)
  • Cost-effective and easy to manage
  • Compliance with regulatory requirements
• AWS DRS was selected to address these needs:
  • Provides account isolation, immutable snapshots, and point-in-time recovery
  • Supports a wide range of operating systems, servers, and databases
  • Highly automated and non-disruptive to production environments

Technical Architecture and Implementation

• Zydus set up replication agents on their on-premises servers to continuously replicate data to AWS
• Data is compressed, encrypted, and copied to EBS volumes in the staging environment
• Pre-configured EC2 launch templates are used to quickly spin up recovery instances during failover
• Secure and resilient networking was established using AWS Direct Connect, Direct Connect Gateway, and Transit Gateway
  • Enabled isolated DR environment and optimized replication traffic

Key Learnings and Benefits

• Established a secure and resilient cloud foundation as the foundation for the DR solution
• Addressed database replication issues by tuning firewall policies, dedicated replication servers, and increasing IOPS/throughput
• Implemented centralized monitoring and automated alerting to proactively manage the DR environment
• Achieved:
  • RPO of < 15 minutes and RTO of < 3 hours
  • No capital investment required for disaster recovery infrastructure
  • Simplified management and operations
  • Compliance with regulatory requirements

Next Steps

• Zydus is expanding their cloud journey beyond disaster recovery:
  • Implementing cloud-based file servers and secure, scalable storage at plant locations
  • Upskilling technical teams on AWS technologies like Amazon QuickSight
  • Evaluating migration of non-critical primary workloads to the cloud