Modernize Infrastructure and Applications

Organizations migrating to the cloud follow different strategies based on their applications, timelines, and goals. Google Cloud and industry analysts identify several common migration patterns, often called the "6 Rs" of migration.

Strategy	Description	When to Use	Effort
Rehost (Lift & Shift)	Move as-is to cloud VMs	Quick wins, legacy apps with no changes	Low
Replatform (Lift & Optimize)	Minor optimizations during move	Use managed services (e.g., Cloud SQL instead of self-managed MySQL)	Low-Medium
Refactor (Re-architect)	Redesign for cloud-native	Take advantage of microservices, containers, serverless	High
Repurchase	Replace with SaaS	Move from on-prem email to Google Workspace	Medium
Retire	Decommission the application	App no longer needed	Low
Retain	Keep on-premises	Regulatory requirements, not ready to migrate	None

Migration Phases

1. Assess — Inventory workloads, assess dependencies, determine TCO, identify migration candidates.
2. Plan — Choose migration strategy per workload, design target architecture, plan network connectivity.
3. Deploy — Set up foundation (VPC, IAM, billing), migrate workloads in waves, validate functionality.
4. Optimize — Right-size resources, implement autoscaling, apply cost controls, modernize further.

Compute Engine provides virtual machines (VMs) running on Google's infrastructure. It is the IaaS compute service — you have full control over the operating system, networking, and installed software.

Machine Type Families

Family	Description	Use Cases
E2 (General)	Cost-optimized, shared-core options	Dev/test, small apps, microservices
N2/N2D (General)	Balanced compute, Intel/AMD	Web servers, databases, medium workloads
C2/C2D (Compute)	Highest per-core performance	Gaming, HPC, single-threaded apps
M2/M3 (Memory)	Ultra-high memory (up to 12 TB)	SAP HANA, in-memory databases
A2/A3 (Accelerator)	GPU-attached (NVIDIA A100/H100)	ML training, rendering, simulations

Key Compute Engine Features

# Create a VM instance
gcloud compute instances create web-server \
  --zone=us-central1-a \
  --machine-type=e2-medium \
  --image-family=debian-12 \
  --image-project=debian-cloud \
  --tags=http-server

# Create a managed instance group with autoscaling
gcloud compute instance-groups managed create my-mig \
  --template=my-template \
  --size=2 \
  --zone=us-central1-a

gcloud compute instance-groups managed set-autoscaling my-mig \
  --max-num-replicas=10 \
  --target-cpu-utilization=0.6 \
  --zone=us-central1-a

Containers package an application with all its dependencies into a portable, lightweight unit that runs consistently across environments. Docker is the most common container runtime. Kubernetes is the open-source system for orchestrating (managing) containers at scale.

Containers vs. VMs

Aspect	Virtual Machines	Containers
Isolation	Full OS per VM (strong isolation)	Shared OS kernel (process isolation)
Size	Gigabytes (includes full OS)	Megabytes (app + dependencies only)
Startup	Minutes	Seconds
Portability	Limited (OS-dependent)	High (runs anywhere Docker runs)
Density	10s per host	100s per host
Use case	Legacy apps, full OS control needed	Microservices, CI/CD, cloud-native apps

GKE — Google Kubernetes Engine

GKE is Google's managed Kubernetes service. Google manages the control plane (API server, scheduler, etcd) while you manage the worker nodes and workloads. GKE Autopilot goes further by also managing the nodes.

Feature	GKE Standard	GKE Autopilot
Node management	You manage nodes	Google manages nodes
Pricing	Pay for nodes (VMs)	Pay per pod (resources used)
Configuration	Full control over node pools	Best practices enforced
Security	You configure hardening	Hardened by default
Best for	Teams needing custom node config	Teams wanting fully managed K8s

# Create a GKE Autopilot cluster
gcloud container clusters create-auto my-cluster \
  --region=us-central1

# Deploy an application
kubectl create deployment hello-app \
  --image=us-docker.pkg.dev/google-samples/containers/gke/hello-app:1.0

# Expose it as a service
kubectl expose deployment hello-app \
  --type=LoadBalancer \
  --port=80 \
  --target-port=8080

Serverless means Google manages all infrastructure — servers, scaling, patching, and capacity planning. You focus only on your code. Google Cloud offers three main serverless compute options.

Feature	Cloud Run	Cloud Functions	App Engine
Unit of deployment	Container image	Single function	Application
Language support	Any (container-based)	Node, Python, Go, Java, .NET, Ruby, PHP	Standard: Python, Java, Go, PHP, Node, Ruby. Flexible: any
Scale to zero	Yes	Yes	Standard: Yes. Flexible: Min 1 instance
Max timeout	60 min	9 min (1st gen) / 60 min (2nd gen)	Varies
Concurrency	Up to 1000 requests per instance	1 per instance (1st gen) / configurable (2nd gen)	Automatic

Service	Model	Control	Scaling	Best For
Compute Engine	IaaS (VMs)	Full (OS, network, storage)	Manual or MIG autoscaler	Legacy apps, custom OS, GPU/TPU workloads
GKE	CaaS (Containers)	High (pods, nodes, networking)	HPA + cluster autoscaler	Microservices at scale, multi-container apps
Cloud Run	Serverless containers	Low (just container image)	Auto (0 to N instances)	Stateless HTTP services, APIs
Cloud Functions	FaaS	Lowest (just function code)	Auto per invocation	Event handlers, glue code, webhooks
App Engine	PaaS	Low-Medium	Auto	Web apps, RESTful backends

Apigee is Google Cloud's full-lifecycle API management platform. It helps organizations design, secure, deploy, monitor, and monetize APIs. APIs are the building blocks of modern application architectures and microservices.

Why APIs Matter

• Microservices architecture — Services communicate via APIs, enabling independent development and deployment.
• Partner integrations — Expose business capabilities to partners securely via APIs.
• Mobile/web backends — APIs serve as the backend for mobile apps and single-page web applications.
• Monetization — Treat APIs as products with rate limiting, usage tracking, and billing.

Apigee Key Features

Anthos is Google Cloud's platform for managing applications across multiple environments: on-premises data centers, Google Cloud, and other public clouds (AWS, Azure). It provides a consistent development and operations experience everywhere.

Anthos Components

Component	Purpose
Anthos on GKE	Managed Kubernetes on Google Cloud (the foundation)
Anthos on-premises	Run GKE on your own hardware in your data center
Anthos on AWS/Azure	Attach and manage Kubernetes clusters on other clouds
Anthos Service Mesh	Managed Istio for service-to-service communication, security, and observability
Anthos Config Management	GitOps-based policy and configuration management across all clusters
Migrate to Containers	Automatically containerize VMs for migration to GKE

Hybrid Connectivity

• Cloud VPN — Encrypted IPsec tunnel over the internet. Quick to set up, lower cost, but throughput limited by internet bandwidth.
• Cloud Interconnect (Dedicated) — Direct physical connection to Google's network. 10 or 100 Gbps. Lowest latency and highest throughput. Requires colocation facility.
• Cloud Interconnect (Partner) — Connection through a supported service provider. Lower bandwidth than dedicated but easier to set up.

Quick Reference

• "Move VMs to cloud without changes" → Rehost (Compute Engine)
• "Run containers without managing servers" → Cloud Run
• "Orchestrate 100s of microservices in containers" → GKE
• "Run same workloads on-prem and in cloud" → Anthos
• "Manage and secure APIs at enterprise scale" → Apigee
• "Connect on-prem to GCP with high bandwidth" → Cloud Interconnect
• "What is unique about Google Cloud VMs?" → Live migration (zero-downtime maintenance)