Lessons Learned from Our Largest Migration (1.5 Years in the Making)

We recently completed our biggest cloud migration project ever, moving from AWS to GCP. This is the second post in a series sharing key lessons learned.

🌐 Regions, Zones, and Networking

• Regions: Both AWS and GCP use geographic regions to organize infrastructure.
• Availability Zones (AZs): While both clouds call them Availability Zones, the naming and locations don’t always match 1:1.

🛡️ VPC Differences

• AWS:
  • A VPC (Virtual Private Cloud) is a regional resource.
  • Inside a VPC, you create subnets, which are zonal.
  • Resources can be zonal (e.g., EC2 instances) or multi-AZ (e.g., Load Balancers, RDS).
• GCP:
  • A VPC is a global resource — it spans multiple regions.
  • Subnets inside a VPC are regional.
  • Resources like VMs are deployed in specific zones within those regional subnets.
  • Load Balancers in GCP are global, which makes them inherently suitable for worldwide failover scenarios.

This architecture makes GCP networking much better suited for global failover, offering a significant advantage when designing robust disaster recovery environments.