Compute Options — ⚙️ VMs, Containers, and Serverless

Cloud providers offer multiple ways to run your application. Each model has different tradeoffs between control, complexity, and cost.

🎯 Virtual Machines (VMs)

You rent a virtual computer that runs your operating system and applications.

What you do: Install OS, deploy applications, manage everything

What provider does: Manages physical hardware

Best for: Full control, legacy applications, custom configurations

Example: AWS EC2, Google Compute Engine, Azure Virtual Machines

Your app
────────────
Operating System
────────────
Hypervisor (Provider manages)
────────────
Physical Hardware (Provider manages)

Pros:

• Maximum control and flexibility

• Run any operating system

• Compatible with legacy applications

Cons:

• You manage OS updates and security patches

• You manage scaling (add/remove VMs manually)

• More expensive per application

💡 Containers

Applications run in isolated containers (like lightweight virtual machines) that include your application and its dependencies.

What you do: Build container image, deploy it

What provider does: Manages container orchestration and infrastructure

Best for: Consistent deployments, microservices, rapid scaling

Example: Docker, Kubernetes, AWS ECS, Google Cloud Run, Azure Container Instances

Your app 1
────────────
Container Runtime (Docker)
────────────
Kubernetes/Orchestration (Provider manages)
────────────
Physical Hardware (Provider manages)

Pros:

• Package everything your app needs (consistent deployments)

• Lightweight and fast

• Scales automatically

• Easy to move between clouds

Cons:

• Learning curve (Docker, Kubernetes concepts)

• More complex setup than VMs

⚡ Serverless (Functions as a Service)

Upload your code (function), specify when it runs, cloud runs it automatically.

What you do: Write code, deploy function

What provider does: Manages infrastructure, scaling, everything

Best for: Event-driven workloads, APIs, background jobs

Example: AWS Lambda, Google Cloud Functions, Azure Functions

Your function
────────────
Serverless Runtime (Provider manages)
────────────
Auto-scaling (Provider manages)
────────────
Physical Hardware (Provider manages)

Pros:

• No infrastructure management

• Pay only when function runs

• Automatic scaling

• Fastest deployment (seconds)

Cons:

• Limited execution time (typically 5-15 minutes)

• No persistent storage on function

• Scaling latency (cold starts)

• Vendor lock-in

🎨 Real-World Examples

VMs: Netflix Backend

Complex, stateful services. Need maximum control. Runs on VMs for flexibility.

Containers: Spotify

Microservices architecture. Each service in container. Easy to scale independently.

Serverless: Mobile App Backend

Mobile app sends data → Lambda function processes → Stores in database. Pay $0 when app not used.

📊 Comparison Table

💰 Cost Comparison (Example)

Running same application:

VMs: $200/month (you pay even when idle)

Containers: $100/month (scale with demand)

Serverless: $10/month (pay per execution, idle = $0)

*Actual costs vary by workload and utilization*

🔑 Key Takeaways

• ✅ VMs give maximum control but require management

• ✅ Containers provide balance of control and automation

• ✅ Serverless offers maximum simplicity and cost efficiency

• ✅ Choose based on: control needed, complexity tolerance, cost requirements

• ✅ Modern applications often use combination of all three

Want architectural patterns? ⚙️ Compute Options (Experienced)