switching software/cloud infrastructure

AWS to Google Cloud: Migration Guide for Data‑Heavy Workloads

Network design, storage mapping, and IAM parity when moving from AWS to GCP.

Executive Summary

This guide outlines the architectural transition from AWS to Google Cloud Platform (GCP). The migration focuses on mapping AWS primitives (VPC, EC2, RDS, IAM) to GCP equivalents (VPC, Compute Engine, Cloud SQL, IAM/Resource Hierarchy). Unlike a lift-and-shift, this approach prioritizes GCP’s unique strengths in data analytics (BigQuery) and container orchestration (GKE).

Why Businesses Migrate

Drivers:

  • Kubernetes Maturity: Leveraging GKE’s superior management plane compared to EKS.
  • Data Analytics: Accessing BigQuery’s serverless architecture for real-time insights.
  • Global Network: Utilizing Google’s private global fiber backbone for lower inter-region latency.

Anti-patterns:

  • "Cloud-agnosticism" (over-engineering to avoid proprietary services).
  • Migrating monolithic legacy apps without evaluating containerization potential.

The 6 Rs of Cloud Migration

  • Rehost (Lift & Shift): Moving EC2 workloads to Compute Engine via Migrate to Virtual Machines (formerly Velostrata).
  • Replatform: Moving RDS to Cloud SQL or self-managed DBs on GKE; minimal code changes.
  • Refactor: Moving from EC2 to Cloud Run or GKE; rewriting for microservices.
  • Repurchase: Moving from AWS Marketplace software to Google Cloud Marketplace SaaS.
  • Retain: Keeping specialized AWS-only workloads (e.g., specific Lambda-heavy stacks).
  • Retire: Decommissioning shadow IT or redundant legacy systems.

Pre-Migration Assessment

  • Inventory: Use Google Cloud Migration Center to perform an automated discovery of your AWS environment.
  • Network Mapping: Map AWS Transit Gateway topologies to GCP Network Connectivity Center.
  • IAM: AWS uses Identity-based policies (JSON); GCP uses Role-Based Access Control (RBAC) via IAM hierarchy (Organization > Folder > Project).
  • Data Transfer Calc: Evaluate gsutil for cold storage vs. Storage Transfer Service for high-throughput AWS S3 to GCS synchronization.

Step-by-Step Execution Plan

  1. Pilot: Establish a "Landing Zone" (GCP Organization, Billing, Folders). Deploy a non-critical microservice.
  2. Foundation: Implement Shared VPCs for network isolation and Cloud Interconnect for hybrid AWS-GCP connectivity.
  3. Data Migration: Execute continuous block-level sync via Migrate to Virtual Machines. Use Database Migration Service (DMS) for RDS to Cloud SQL.
  4. Cutover: Update DNS (Cloud DNS) to point traffic to GCP Global Load Balancers. Perform a "Blue-Green" transition.

Troubleshooting Common Issues

  • IAM Latency: GCP IAM policies are eventually consistent; allow 30-60 seconds for propagation.
  • Network MTU: AWS defaults to 9001 (Jumbo) in some cases; GCP VPCs typically default to 1460. Ensure fragmentation isn't occurring.
  • Security Groups vs. Firewall Rules: AWS Security Groups are stateful; GCP Firewall Rules are also stateful but applied at the VPC level, not the individual NIC.

UK-Specific Compliance (GDPR, Data Residency)

  • Data Residency: Utilize GCP’s europe-west2 (London) region for PII storage to maintain GDPR compliance.
  • Org Policy Service: Enforce "Resource Location Restriction" policies to prevent developers from provisioning resources in non-EU regions.
  • Data Sovereignty: Use VPC Service Controls to create a security perimeter around sensitive data, preventing unauthorized egress.

Cost Modeling

  • Compute: Utilize Sustained Use Discounts (SUDs) and Committed Use Contracts (CUDs) in GCP.
  • Network: AWS charges for inter-AZ data transfer; GCP’s internal global network pricing model differs—factor in egress costs between regions.
  • Tools: Use the Google Cloud Pricing Calculator to compare AWS monthly spend against projected GCP consumption.

Conclusion

Migrating from AWS to GCP is not merely a change in provider; it is an evolution toward a data-driven, container-native architecture. By focusing on the Foundation phase and leveraging Google's automated migration suite, organizations can minimize downtime and capitalize on GCP's superior network and analytics capabilities.