Which Cloud Platform Wins? A Hands‑On Comparison of AWS, Azure, Google Cloud, and Oracle

Following a high‑profile AWS outage that disrupted a large portion of the internet, the author decided to evaluate the four major cloud providers—AWS, Azure, Google Cloud, and Oracle Cloud—by using each service hands‑on for several weeks.

01. Overview: Who Offers What?

All providers cover the same core domains (compute, storage, databases, networking, and specialized services), but the differences lie in the details of each offering.

02. Compute Services

AWS leads with EC2, EKS for Kubernetes, and Lambda for serverless workloads. Azure offers Virtual Machines, AKS, and Azure Functions, which excel for Windows‑based workloads. Google Cloud provides Compute Engine, GKE, and Cloud Functions, with GKE often regarded as the best managed Kubernetes service. Oracle supplies Virtual Machines, OCI Container Engine, and OCI Functions, entering the market later but with competitive pricing.

# AWS Lambda (Python)
def lambda_handler(event, context):
    return {
        'statusCode': 200,
        'body': 'Hello from Lambda'
    }
# Azure Functions (Python)
import azure.functions as func
def main(req: func.HttpRequest) -> func.HttpResponse:
    return func.HttpResponse("Hello from Azure Functions")

The syntax of these serverless functions is almost identical, making migration between platforms relatively easy.

03. Storage: Where Your Data Lives

AWS provides S3, EBS, and EFS. Azure offers Blob Storage, Managed Disks, and File Storage. Google Cloud supplies Cloud Storage and Persistent Disk. Oracle delivers Object Storage and Block Volumes.

Performance tests showed S3 and Google Cloud Storage excel at large‑file throughput, while Azure Blob Storage handles millions of small objects more efficiently.

# Typical cloud architecture diagram
[Load Balancer]
   |
   ---------
   |       |
[Web]   [Web]
   |       |
   ---------
   |
[App Layer]
   |
   ---------
   |       |
[Cache] [Database]
   |
[Storage]

04. Databases

AWS: RDS, DynamoDB, Redshift. Azure: SQL Database, Cosmos DB, Synapse Analytics. Google Cloud: Cloud SQL, Firestore, BigQuery. Oracle: Oracle Database, NoSQL Database, Autonomous Data Warehouse.

BigQuery completed a 10 TB query in 23 seconds, beating Redshift (1.2 min), Synapse (1.5 min), and Oracle ADW (1.8 min). For transactional workloads, Oracle and Azure SQL Database provide the strongest consistency guarantees.

05. Networking

All providers offer VPC capabilities, DNS services, and load balancers. AWS has VPC and Route 53; Azure provides Virtual Network and DNS; Google Cloud supplies VPC and Cloud DNS; Oracle offers Virtual Cloud Network and DNS. Performance is comparable, with differences mainly in configuration complexity.

06. Machine Learning

AWS SageMaker, Azure Machine Learning, Google Vertex AI, and Oracle Data Science all deliver end‑to‑end ML platforms. Google leads in pre‑trained models and AutoML capabilities.

# Google Cloud AutoML example
from google.cloud import automl
client = automl.AutoMlClient()
# Train a model with minimal code
model = client.create_model(
    parent="projects/my-project/locations/us-central1",
    model={
        "display_name": "my_model",
        "dataset_id": dataset_id,
        "image_classification_model_metadata": {}
    }
)

07. Pricing

Oracle Cloud is 30‑50 % cheaper than competitors for compute and storage. Google Cloud offers sustained‑use discounts automatically. AWS pricing is the most complex but also the most flexible. Azure provides hybrid benefits for existing Microsoft licenses.

For a medium workload (8 vCPU, 32 GB RAM, 1 TB storage), monthly costs were:

Oracle Cloud: $245

Google Cloud: $312

AWS: $387

Azure: $395

08. Ecosystem Factors

AWS has the largest market and third‑party integrations. Azure dominates enterprise environments, especially with Active Directory integration. Google Cloud shines in data analytics and ML workflows. Oracle is the best choice for traditional database migrations.

09. Cross‑Platform Microservice Architecture

Internet
   |
[API Gateway / Application Gateway]
   |
--- [Auth Service] --- [Key Vault / Secrets Manager]
   |
--- [User Service] --- [SQL Database]
   |
--- [Order Service] --- [NoSQL Database]
   |
--- [Notification Service] --- [Message Queue]
                                   |
                               [Email Service]

10. Final Recommendations

If you need the broadest service catalog and can handle complexity, choose AWS. If you are deeply tied to the Microsoft ecosystem or need strong hybrid capabilities, pick Azure. For data‑intensive, ML, or Kubernetes‑centric workloads, Google Cloud offers the best developer experience. Consider Oracle when you run Oracle databases or need the best price‑to‑performance ratio for compute‑heavy tasks.

The key takeaway is that no single cloud provider wins for every scenario; a multi‑cloud strategy often provides the best balance of features, cost, and risk mitigation.