Why This Matters

Hosting determines performance, security, and cost. Choosing wrong leads to slow responses, security gaps, or runaway costs.

Concept Explained

Three hosting models: local (dev only), cloud (AWS/Azure/GCP), on-prem (behind firewall). Trade-offs: speed, security, cost, maintenance.

How It Works

Local for dev. Cloud for production (scalable, managed). On-prem for regulated industries (full control).

Step-by-Step Tutorial

1. Identify needs

2. Compare models

3. Choose

4. Plan scaling

5. Plan security

6. Document

Real-World Example

A regulated fintech chose on-prem hosting for their Extension. Full control over data, no external cloud. Slower to scale but met compliance. Trade-off accepted.

Example Prompts / Commands / Code

Model        Use case              Pros                       Cons
------------------------------------------------------------------------
Local       Dev only              Fast iteration             Not for prod
Cloud       Production, scaling  Scalable, managed          Data leaves premises
On-prem     Regulated industries Full control, compliance   Slower scaling, maintenance

# Serverless Extension
# - Auto-scales
# - Pay per request
# - Managed

# Lambda function:
def handler(event, context):
    # Parse Copilot request
    # Call internal service
    # Return response
    pass

# API Gateway: HTTPS endpoint
# Lambda: business logic
# Internal service: data source

Common Mistakes

• Using local for production — not scalable, not secure.
• Cloud without security review — data exposure.
• On-prem without capacity planning — slow under load.
• Not documenting hosting decision — tribal knowledge.

Best Practices

• Local for dev only.
• Cloud for scalable production.
• On-prem for regulated industries.
• Plan scaling and security.
• Document hosting decision.

Troubleshooting

Practical Exercise

Professional Challenge

Frequently Asked Questions

Further Reading

• Building Extensions

Official References

• Building Extensions

Related lessons: PR-21, PR-23