CDL
Deep Dive
The 6 Rs of migration, Google Cloud tools for migrating workloads, and the application modernization path.
When a company moves workloads to the cloud, it chooses a migration strategy. The "6 Rs" are the industry-standard framework for categorizing these strategies.
Moves the app to the cloud as-is, without modifications. On-premises VM → Compute Engine. No optimization, maximum migration speed.
When to apply
Migrate quickly to exit a datacenter. First step before modernizing.
Small changes to take advantage of cloud capabilities without redesigning. E.g.: migrating from MySQL on VMs to Cloud SQL.
When to apply
You want basic cloud benefits (managed service, automatic backups) with minimal effort.
Switch to a different SaaS product. E.g.: moving from an in-house CRM to Salesforce, or from Exchange to Google Workspace.
When to apply
The app isn't a business differentiator and a SaaS exists that covers the use case.
Redesign the app to use cloud-native capabilities: microservices, containers, serverless. More effort, greater long-term benefit.
When to apply
The app has scale or agility issues that can only be solved by redesign.
Eliminate apps that are no longer needed. Typically 10-20% of the portfolio in large migrations can be retired.
When to apply
The app is duplicated, has no active users, or the business no longer needs it.
Keep the app on-premises for now for technical, contractual, or business reasons. Not everything needs to migrate on day 1.
When to apply
App with a long active contract, specific regulations, or migration not yet cost-effective.
| Tool | What it migrates | Key details |
|---|---|---|
| Migrate to Virtual Machines | VMs from VMware, AWS, Azure, or bare metal → Compute Engine | Continuous replication + minimal cutover. Test the VM on GCP before the final cutover. |
| Database Migration Service | Relational databases to Cloud SQL or AlloyDB | PostgreSQL, MySQL, SQL Server. CDC (Change Data Capture) for migration with minimal downtime. |
| Datastream | Real-time database changes (continuous replication) | Serverless CDC. Replicates from Oracle, MySQL, PostgreSQL to BigQuery, Cloud Storage, Spanner. |
| Storage Transfer Service | Data from AWS S3, Azure Blob, HDFS, on-premises → Cloud Storage | Scheduled or continuous transfers. For petabytes of data. |
| Transfer Appliance | Massive on-premises data (>20TB) → Cloud Storage | A physical device that Google ships. Copy data offline and return it to Google. |
| BigQuery Data Transfer Service | SaaS data (Google Ads, YouTube, Salesforce) → BigQuery | Native connectors for dozens of external data sources. |
Inventory and evaluate current apps. What to migrate, modernize, retire? Google Cloud Rapid Assessment & Migration Program (RAMP) helps with this analysis.
Lift-and-shift: move VMs as-is to Compute Engine. Fast, no code changes. Foundation for later modernization.
Containerize apps with Docker, deploy to GKE or Cloud Run. Migrate databases to managed services (Cloud SQL, Spanner). Adopt managed services.
Redesign as microservices, use Cloud Run functions for event-driven logic, Pub/Sub for messaging, and managed APIs. Maximum use of GCP.
What used to be Anthos (later GKE Enterprise) is now part of GKE: fleets manage clusters on GCP, on-premises (Google Distributed Cloud), and on AWS or Azure (GKE Multi-Cloud) with a unified control plane for Kubernetes anywhere.
GKE clusters
Managed Kubernetes on-premises or in other clouds. Same GKE API across all environments.
Config Management
Policy-as-code: applies configurations and security policies across all clusters from a Git repository.
Cloud Service Mesh
Managed Istio for service mesh: traffic, observability, and security (mTLS) between microservices.
Use case
Companies with regulations requiring on-premises, or a multi-cloud strategy without vendor lock-in.
Hybrid Cloud
A combination of on-premises infrastructure with one or more public clouds, working together. Companies keep some workloads on-premises and others in the cloud.
Multi-cloud
Deliberately using services from multiple cloud providers (GCP + AWS + Azure). Not to be confused with having a presence in two clouds by accident.
GKE multi-cloud: one control panel for everything
GKE with fleets, GKE Multi-Cloud, and Google Distributed Cloud (formerly Anthos / GKE Enterprise) enable managing apps across multiple environments (GCP, on-premises, AWS, Azure) from a single Kubernetes-based control panel. Config Management applies consistent security policies across all clusters. Cloud Service Mesh (Istio) manages communication between microservices in any environment.
Did you get this topic?
Test what you just learned
A company has 200 VMs in its on-premises datacenter running critical legacy applications. They want to migrate to Google Cloud as quickly as possible to exit the datacenter in 6 months. The apps cannot be modified. What strategy and tool should they use?
Put what you learned to the test with original CDL exam questions.
Start CDL mock examUnlimited mock exams with the Pro plan
Log in to track your progress.