FaaS - Function as a Service

Understanding Function as a Service (FaaS) serverless computing model, covering AWS Lambda, Azure Functions, and Google Cloud Functions with practical examples.

Function as a Service (FaaS) is a serverless computing model where developers write individual functions that execute in response to events. The cloud provider manages all infrastructure — servers, scaling, patching, and availability — letting developers focus purely on business logic.

FaaS Execution Model

┌─────────────────────────────────────────────────────────────┐ │ FaaS EXECUTION MODEL │ │ │ │ EVENT SOURCE FUNCTION RESULT │ │ ┌──────────┐ ┌────────┐ ┌──────────┐ │ │ │ HTTP Req │────────►│ │────────►│ Response │ │ │ └──────────┘ │ │ └──────────┘ │ │ ┌──────────┐ │ Your │ ┌──────────┐ │ │ │ S3 Event │────────►│ Code │────────►│ DB Write │ │ │ └──────────┘ │ │ └──────────┘ │ │ ┌──────────┐ │(single │ ┌──────────┐ │ │ │ Schedule │────────►│function│────────►│ Email │ │ │ └──────────┘ │ │ └──────────┘ │ │ ┌──────────┐ │ │ ┌──────────┐ │ │ │ Queue Msg│────────►│ │────────►│ Log │ │ │ └──────────┘ └────────┘ └──────────┘ │ │ │ │ Provider handles: Scaling, Availability, Patching │ │ You pay for: Execution time (milliseconds) │ └─────────────────────────────────────────────────────────────┘

FaaS Platforms

Platform	Provider	Languages	Max Timeout
AWS Lambda	Amazon	Python, Node, Java, Go, .NET, Ruby	15 min
Azure Functions	Microsoft	C#, JavaScript, Python, Java, PowerShell	Unlimited*
Cloud Functions	Google	Node.js, Python, Go, Java, .NET, Ruby	60 min
IBM Functions	IBM	Node.js, Python, Swift, PHP	10 min
Cloudflare Workers	Cloudflare	JavaScript, Rust, WASM	30 sec (free)

Writing FaaS Functions

AWS Lambda (Python)

python example

import json
import boto3

def lambda_handler(event, context):
    """Process an API Gateway request"""
    
    # Extract request data
    body = json.loads(event.get('body', '{}'))
    name = body.get('name', 'World')
    
    # Business logic
    greeting = f"Hello, {name}! Welcome to serverless."
    
    # Return response
    return {
        'statusCode': 200,
        'headers': {
            'Content-Type': 'application/json',
            'Access-Control-Allow-Origin': '*'
        },
        'body': json.dumps({
            'message': greeting,
            'timestamp': context.get_remaining_time_in_millis()
        })
    }

Azure Functions (JavaScript)

Example

module.exports = async function (context, req) { const name = req.query.name || req.body?.name || 'World'; context.res = { status: 200, body: { message: `Hello, ${name}! From Azure Functions.`, timestamp: new Date().toISOString() } }; };

Google Cloud Functions (Python)

python example

import functions_framework

@functions_framework.http
def hello_http(request):
    request_json = request.get_json(silent=True)
    name = request_json.get('name', 'World') if request_json else 'World'
    
    return {
        'message': f'Hello, {name}! From Google Cloud Functions.',
        'status': 'success'
    }

Deploying FaaS Functions

bash example

# AWS Lambda deployment
zip function.zip lambda_function.py
aws lambda create-function \
  --function-name my-api-handler \
  --runtime python3.11 \
  --handler lambda_function.lambda_handler \
  --zip-file fileb://function.zip \
  --role arn:aws:iam::123456789:role/lambda-role \
  --timeout 30 \
  --memory-size 256

# Add API Gateway trigger
aws apigatewayv2 create-api \
  --name my-api \
  --protocol-type HTTP \
  --target arn:aws:lambda:us-east-1:123456789:function:my-api-handler

# Google Cloud Functions deployment
gcloud functions deploy hello_http \
  --runtime python311 \
  --trigger-http \
  --allow-unauthenticated \
  --memory 256MB \
  --region us-central1

# Azure Functions deployment
func azure functionapp publish my-function-app

FaaS Pricing Model

│ Free Tier	1M requests + 400,000 GB-seconds/mo │
│ After free tier	│
│ Example	10M requests/mo, 128MB, 200ms avg │
│ Compute	10M × 0.2s × 0.125GB = 250,000 GB-sec │
│ Cost	(250,000 × $0.0000166667) + (10 × $0.20) │
│ Total	$4.17 + $2.00 = $6.17/month │
│ Equivalent server	~$50-100/month (always on) │

FaaS vs Traditional Servers

Aspect	FaaS	Traditional Server
Billing	Per-execution (ms)	Per-hour (running)
Scaling	Automatic (0 to thousands)	Manual or auto-scaling config
Cold Start	Yes (100ms-5s)	No (always running)
State	Stateless	Stateful
Max Duration	15 min (Lambda)	Unlimited
Maintenance	Zero	OS patches, updates

Interview Questions

What is FaaS and how does it differ from traditional server-based computing?

FaaS executes individual functions in response to events, with zero server management. Unlike traditional servers that run continuously, FaaS functions are ephemeral, stateless, scale automatically, and charge only for actual execution time.

What is a cold start in FaaS and how can you mitigate it?

Cold start is the delay when a function is invoked after being idle — the platform must provision a container and load the runtime. Mitigations: provisioned concurrency (Lambda), smaller deployment packages, lighter runtimes, and keeping functions warm with scheduled pings.

When is FaaS NOT appropriate?

For long-running processes (>15 min), stateful applications, real-time systems requiring consistent low latency, high-throughput constant workloads (cheaper on reserved servers), and applications needing persistent connections.

Explain event-driven architecture in the context of FaaS.

Functions are triggered by events (HTTP requests, file uploads, database changes, messages, schedules). This creates loosely coupled systems where components communicate through events rather than direct calls, improving scalability and maintainability.

Compare AWS Lambda, Azure Functions, and Google Cloud Functions.

Lambda: most mature, largest ecosystem, 15-min timeout. Azure Functions: best for .NET, unlimited duration on premium plan, Visual Studio integration. Cloud Functions: simpler API, 60-min timeout, tight GCP integration, good for data pipelines.

FaaS Execution Model

FaaS Platforms

Platform	Provider	Languages	Max Timeout
AWS Lambda	Amazon	Python, Node, Java, Go, .NET, Ruby	15 min
Azure Functions	Microsoft	C#, JavaScript, Python, Java, PowerShell	Unlimited*
Cloud Functions	Google	Node.js, Python, Go, Java, .NET, Ruby	60 min
IBM Functions	IBM	Node.js, Python, Swift, PHP	10 min
Cloudflare Workers	Cloudflare	JavaScript, Rust, WASM	30 sec (free)

Writing FaaS Functions

AWS Lambda (Python)

python example

import json
import boto3

def lambda_handler(event, context):
    """Process an API Gateway request"""
    
    # Extract request data
    body = json.loads(event.get('body', '{}'))
    name = body.get('name', 'World')
    
    # Business logic
    greeting = f"Hello, {name}! Welcome to serverless."
    
    # Return response
    return {
        'statusCode': 200,
        'headers': {
            'Content-Type': 'application/json',
            'Access-Control-Allow-Origin': '*'
        },
        'body': json.dumps({
            'message': greeting,
            'timestamp': context.get_remaining_time_in_millis()
        })
    }

Azure Functions (JavaScript)

Example

Google Cloud Functions (Python)

python example

import functions_framework

@functions_framework.http
def hello_http(request):
    request_json = request.get_json(silent=True)
    name = request_json.get('name', 'World') if request_json else 'World'
    
    return {
        'message': f'Hello, {name}! From Google Cloud Functions.',
        'status': 'success'
    }

Deploying FaaS Functions

bash example

# AWS Lambda deployment
zip function.zip lambda_function.py
aws lambda create-function \
  --function-name my-api-handler \
  --runtime python3.11 \
  --handler lambda_function.lambda_handler \
  --zip-file fileb://function.zip \
  --role arn:aws:iam::123456789:role/lambda-role \
  --timeout 30 \
  --memory-size 256

# Add API Gateway trigger
aws apigatewayv2 create-api \
  --name my-api \
  --protocol-type HTTP \
  --target arn:aws:lambda:us-east-1:123456789:function:my-api-handler

# Google Cloud Functions deployment
gcloud functions deploy hello_http \
  --runtime python311 \
  --trigger-http \
  --allow-unauthenticated \
  --memory 256MB \
  --region us-central1

# Azure Functions deployment
func azure functionapp publish my-function-app

FaaS Pricing Model

│ Free Tier	1M requests + 400,000 GB-seconds/mo │
│ After free tier	│
│ Example	10M requests/mo, 128MB, 200ms avg │
│ Compute	10M × 0.2s × 0.125GB = 250,000 GB-sec │
│ Cost	(250,000 × $0.0000166667) + (10 × $0.20) │
│ Total	$4.17 + $2.00 = $6.17/month │
│ Equivalent server	~$50-100/month (always on) │

FaaS vs Traditional Servers

Aspect	FaaS	Traditional Server
Billing	Per-execution (ms)	Per-hour (running)
Scaling	Automatic (0 to thousands)	Manual or auto-scaling config
Cold Start	Yes (100ms-5s)	No (always running)
State	Stateless	Stateful
Max Duration	15 min (Lambda)	Unlimited
Maintenance	Zero	OS patches, updates

Interview Questions

What is FaaS and how does it differ from traditional server-based computing?

What is a cold start in FaaS and how can you mitigate it?

When is FaaS NOT appropriate?

Explain event-driven architecture in the context of FaaS.

Compare AWS Lambda, Azure Functions, and Google Cloud Functions.

FaaS - Function as a Service

FaaS Execution Model

FaaS Platforms

Writing FaaS Functions

AWS Lambda (Python)

Azure Functions (JavaScript)

Google Cloud Functions (Python)

Deploying FaaS Functions

FaaS Pricing Model

FaaS vs Traditional Servers

Interview Questions

Continue learning this concept

FaaS - Function as a Service

FaaS Execution Model

FaaS Platforms

Writing FaaS Functions

AWS Lambda (Python)

Azure Functions (JavaScript)

Google Cloud Functions (Python)

Deploying FaaS Functions

FaaS Pricing Model

FaaS vs Traditional Servers

Interview Questions

Continue learning this concept