Cloud Platforms: AWS, Azure, Google Cloud – Features, Services & Comparisons

Core Concepts: What Are Cloud Platforms?

Cloud platforms are on-demand, scalable computing environments provided by third-party vendors. They offer infrastructure, platforms, and software as services (IaaS, PaaS, SaaS) over the internet. The three giants—AWS, Azure, and Google Cloud—dominate ~90% of the market, each with unique strengths.

Why Do Companies Use Cloud Platforms?

• Cost Efficiency: Pay-as-you-go models eliminate upfront hardware costs.
• Scalability: Instantly scale resources (e.g., Daraz during sales, Ncell during data surges).
• Global Reach: Deploy services in multiple regions (e.g., AWS’s 105 Availability Zones).
• Reliability: Built-in redundancy (e.g., Google Cloud’s 99.99% uptime SLA).

1. Amazon Web Services (AWS): The Market Leader

AWS, launched in 2006, is the oldest and most mature cloud platform, used by 90% of Fortune 500 companies.

Key Services

How AWS Works: A Real Example

Scenario: eSewa needs to handle 10x traffic during Dashain.

1. Auto Scaling: AWS EC2 automatically spins up 50+ servers.
2. Load Balancing: Distributes traffic across servers to avoid crashes.
3. S3 for Static Content: User profiles and payment pages load faster.
4. RDS for Transactions: Database queries process in milliseconds.

sequenceDiagram
    participant User as eSewa User
    participant LB as AWS Load Balancer
    participant EC2 as EC2 Instances (50)
    participant RDS as RDS Database
    participant S3 as S3 Storage

    User->>LB: Request (e.g., "Pay Rs. 500")
    LB->>EC2: Route to least-loaded server
    EC2->>RDS: Query: "Check balance"
    RDS-->>EC2: Return balance
    EC2->>S3: Fetch user profile
    S3-->>EC2: Return profile
    EC2->>User: Confirm payment

2. Microsoft Azure: The Enterprise Giant

Azure, launched in 2010, is Microsoft’s cloud platform, tightly integrated with Windows, Office 365, and Active Directory. It’s the #2 choice for enterprises (e.g., banks, government systems).

Key Services

Why Azure?

• Seamless Windows Integration: Ideal for Nepali banks using Windows Server.
• Hybrid Cloud: Azure Arc lets companies mix cloud and on-premises (e.g., NTC’s gradual migration).
• Enterprise Support: 24/7 Microsoft support for critical systems.

Worked Example: Nepal Rastra Bank (NRB) migrates its core banking system to Azure.

1. Lift-and-Shift: Moves existing Windows Server apps to Azure VMs.
2. Azure SQL Database: Replaces on-prem SQL Server for better security.
3. Azure Active Directory: Manages employee access centrally.
4. Backup: Uses Azure Site Recovery for disaster recovery.

3. Google Cloud Platform (GCP): The AI/ML Powerhouse

GCP, launched in 2011, is Google’s cloud platform, known for AI/ML, data analytics, and Kubernetes.

Key Services

Why GCP?

• Superior AI/ML Tools: Vertex AI and TensorFlow integration.
• Global Network: Google’s private fiber backbone (lowest latency for global apps).
• Open-Source Friendly: Native support for Kubernetes, Docker, and open-source tools.

Worked Example: Daraz uses GCP for global order processing.

1. Multi-Region Deployment: Orders from Kathmandu and Dubai route to the nearest GCP region.
2. Firestore: Real-time inventory updates across all warehouses.
3. BigQuery: Analyzes customer purchase patterns for recommendations.
4. GKE: Runs microservices for payments, logistics, and customer support.

graph TD
    subgraph Kathmandu
        A["User Order"] --> B["GCP Load Balancer"]
    end
    subgraph Singapore
        B --> C["GKE Cluster"]
        C --> D["Firestore DB"]
        C --> E["BigQuery Analytics"]
    end
    subgraph Dubai
        B --> F["GKE Cluster"]
        F --> G["Firestore DB"]
    end

## In the Real World

1. eSewa (AWS)

   • Service Used: AWS Lambda + API Gateway
   • How: During festivals, eSewa uses AWS Lambda to auto-scale payment processing without manual server management. API Gateway routes requests to microservices handling UPI, credit card, and mobile top-ups.

2. Ncell (Azure)

   • Service Used: Azure IoT Hub + Machine Learning
   • How: Ncell uses Azure IoT Hub to monitor network towers in real-time. ML models predict outages (e.g., during monsoons) and auto-trigger maintenance alerts to field teams.

3. Daraz (GCP)

   • Service Used: Google Cloud Load Balancing + Firestore
   • How: During sales like 11.11, Daraz’s global traffic is distributed via GCP’s global load balancer. Firestore syncs inventory across 100+ warehouses in milliseconds, preventing overselling.

Comparison Table: AWS vs. Azure vs. GCP

## How to Choose Between AWS, Azure, and GCP?

Use this decision flowchart:

flowchart TD
    A["Need Cloud Platform?"] --> B{"Windows/Office 365?"}
    B -->|"Yes"| C["Azure"]
    B -->|"No"| D{"AI/ML Focus?"}
    D -->|"Yes"| E["Google Cloud"]
    D -->|"No"| F{"Global Scalability?"}
    F -->|"Yes"| G["AWS"]
    F -->|"No"| H{"Enterprise Legacy?"}
    H -->|"Yes"| C
    H -->|"No"| E

Real-World Tie-In:

• Nepal Rastra Bank → Azure (Windows Server + hybrid cloud).
• Daraz → GCP (AI recommendations + global traffic).
• eSewa → AWS (pay-as-you-go scalability).

## Exam Tip

What to Expect in TU/PU/NEB Exams

1. Short Questions (2-5 marks)

   • Define AWS EC2, Azure Arc, or GCP BigQuery.
   • Compare IaaS vs. PaaS in AWS vs. Azure.
   • Example:

     "Differentiate between AWS Lambda and Azure Functions."

2. Long Questions (10-15 marks)

   • Scenario-Based: "How would you deploy a Nepali bank’s core banking system on Azure?"
     • Steps: VMs → SQL Database → Active Directory → Backup.
   • Trace Diagrams: Draw a sequence diagram for eSewa’s payment flow using AWS services.
   • Comparison Tables: "Compare AWS S3 and Google Cloud Storage for a Daraz-like e-commerce site."

3. Case Studies (15-20 marks)

   • Given: "NTC wants to migrate its legacy billing system to the cloud. Recommend AWS/Azure/GCP with justification."
   • Your Answer Must Include:
     • Chosen platform (e.g., Azure for hybrid + Windows Server).
     • Services (e.g., Azure Virtual Machines for legacy apps, Azure SQL for databases).
     • Security (e.g., Azure Active Directory for access control).
     • Cost estimation (e.g., Azure Hybrid Benefit for Windows licenses).

4. Common Pitfalls

   • ❌ Saying "AWS is only for startups" (it’s used by Netflix, NASA).
   • ❌ Ignoring regional availability (e.g., AWS has no region in Nepal; use Singapore/Mumbai).
   • ❌ Mixing up Azure Arc (hybrid) with AWS Outposts (on-prem AWS).

How to Score Full Marks

• Use Real Examples: Always tie answers to eSewa, Daraz, Ncell, or banks.
• Draw Diagrams: For every flow (e.g., sequence diagram for a payment system).
• Compare Features: Use the comparison table in your answer.
• Justify Choices: "We choose GCP for Daraz because of its global load balancing and Firestore for real-time inventory."

## Quick Revision Checklist

Before the exam, ensure you can: ✅ List 5 AWS services and their use cases. ✅ Explain Azure Arc and Google Kubernetes Engine (GKE). ✅ Draw a sequence diagram for a cloud-based payment system. ✅ Compare AWS, Azure, and GCP in a table. ✅ Recommend a cloud platform for 3 Nepali scenarios (bank, e-commerce, telecom).