Comparing AWS, Azure, and Google Cloud in 2025

Core Services Comparison

Global Infrastructure

Compute Services: Pricing and Performance

• AWS EC2:
  Offers a wide range of instance types (general, compute, memory, storage optimized). Spot and Savings Plans provide significant discounts.
  bash
# Launching an EC2 instance (CLI)
aws ec2 run-instances --image-id ami-123456 --instance-type t3.micro --key-name MyKey

• Azure VMs:
  Comparable instance types. Azure Hybrid Benefit allows use of on-premise Windows licenses, reducing costs for enterprise workloads.
  azurecli
# Create a VM
az vm create --resource-group MyGroup --name MyVM --image UbuntuLTS --admin-username azureuser

• GCP Compute Engine:
  Custom machine types offer granular control over vCPU and memory, minimizing over-provisioning. Sustained use discounts apply automatically.
  bash
# Create a VM instance
gcloud compute instances create my-vm --machine-type=e2-custom-4-8192 --image-family=debian-11

Storage Services: Features and Pricing

• Object Storage
• AWS S3: Industry standard, robust ecosystem. Intelligent-Tiering for automatic cost optimization.
• Azure Blob Storage: Strong integration with Microsoft tools. Tiers: Hot, Cool, Archive.
• GCP Cloud Storage: Uniform API, dual-region buckets, and Turbo Replication.

Managed Kubernetes: Feature Comparison

AI/ML Services: Practical Examples

• AWS SageMaker: End-to-end ML lifecycle with built-in algorithms and notebook instances.
  python
import sagemaker
session = sagemaker.Session()
estimator = sagemaker.estimator.Estimator(
image_uri='123456789012.dkr.ecr.us-west-2.amazonaws.com/my-image:latest',
role='arn:aws:iam::123456789012:role/SageMakerRole',
instance_count=1,
instance_type='ml.m5.large'
)
estimator.fit('s3://my-bucket/data/')

• Azure Machine Learning: Designer drag-and-drop and code-first experiences.
  python
from azureml.core import Workspace, Dataset
ws = Workspace.from_config()
dataset = Dataset.get_by_name(ws, name='my-dataset')

• Google Vertex AI: Unified platform for data labeling, training, and deployment. AutoML and custom models.
  python
from google.cloud import aiplatform
aiplatform.init(project='my-project', location='us-central1')
model = aiplatform.Model.upload(display_name='my-model', artifact_uri='gs://my-bucket/model/')

Hybrid and Multi-cloud Capabilities

Security, Compliance, and Identity

• AWS: IAM, Organizations, KMS, Macie. Compliance: FedRAMP, HIPAA, GDPR, C5.
• Azure: Active Directory, RBAC, Key Vault, Sentinel. Compliance: Government, Financial, Healthcare.
• GCP: IAM, Cloud Identity, Security Command Center, DLP API. Compliance: PCI DSS, ISO 27001, GDPR.

Serverless Offerings

Example: Deploying a basic function (GCP)

gcloud functions deploy hello_world --runtime python310 --trigger-http --allow-unauthenticated

Pricing and Cost Management

• AWS: AWS Cost Explorer, Budgets, Savings Plans.
• Azure: Cost Management + Billing, Azure Advisor, Reservations.
• GCP: Billing Reports, Cost Tables, Committed Use Discounts (CUDs).

DevOps and CI/CD Integrations

• AWS: CodePipeline, CodeBuild, CodeDeploy, native GitHub Actions integrations.
• Azure: Azure DevOps, Pipelines, GitHub Actions (deepest integration).
• GCP: Cloud Build, Cloud Deploy, Cloud Source Repositories.

Summary Table: Platform Strengths (2025)