Portability Strategy

The stack uses standard technologies with no vendor lock-in, enabling migration between deployment platforms without code changes. Start with rapid iteration on Vercel + Supabase, then migrate to Google Cloud or AWS for production security when needed.

Philosophy

Start fast, scale secure:

1. Development - Rapid iteration on Vercel + Supabase
2. Production - Migrate to Google Cloud / AWS for enterprise security
3. Zero code changes - Only configuration changes required

Why Portability Matters

• ✅ No vendor lock-in - Standard technologies, no proprietary APIs
• ✅ Security flexibility - Migrate to environments with KMS, VPC, HSM
• ✅ Cost optimization - Choose best provider for each stage
• ✅ Compliance - Meet enterprise security requirements when needed

Migration Paths

flowchart LR
  Dev[Development<br/>Vercel + Supabase] --> Prod1[Production Option 1<br/>Google Cloud Run + Cloud SQL]
  Dev --> Prod2[Production Option 2<br/>AWS ECS + RDS]
  Dev --> Prod3[Production Option 3<br/>Self-hosted / On-premises]
  
  Prod1 --> Enterprise[Enterprise Security<br/>KMS, VPC, HSM]
  Prod2 --> Enterprise
  Prod3 --> Enterprise

Technology Choices for Portability

Runtime: Node.js LTS

• ✅ Standard Node.js - Runs anywhere (Vercel, GCP, AWS, on-premises)
• ✅ No platform-specific code - Pure Node.js application
• ✅ Universal support - All cloud providers support Node.js

Framework: Fastify

• ✅ Standard HTTP - No framework lock-in
• ✅ Portable - Runs as standard Node.js process
• ✅ Migration - Change deployment target only, no code changes

Database: PostgreSQL + Drizzle

• ✅ Standard SQL - Works with any PostgreSQL host
• ✅ Plain SQL queries - Drizzle generates standard PostgreSQL
• ✅ No proprietary APIs - Standard DATABASE_URL connection
• ✅ Migration - Only requires changing DATABASE_URL

Deployment Options

Development: Vercel + Supabase

Best for: Rapid iteration, preview environments, fast setup

• Vercel - Serverless functions, edge network, preview deployments
• Supabase - Managed Postgres, branching, local development
• Fast setup - Get started in minutes

Production: Google Cloud

Best for: Enterprise security, compliance, Google Cloud ecosystem

• Cloud Run - Serverless containers, auto-scaling
• Cloud SQL - Managed PostgreSQL with backups
• Security - KMS encryption, VPC isolation, Cloud IAM
• Migration - Change DATABASE_URL, deploy to Cloud Run

Production: AWS

Best for: AWS ecosystem, enterprise compliance

• ECS / EC2 - Container orchestration or VMs
• RDS - Managed PostgreSQL with backups
• Security - KMS encryption, VPC isolation, IAM
• Migration - Change DATABASE_URL, deploy to ECS/EC2

Enterprise: Self-Hosted / On-Premises

Best for: Maximum control, air-gapped environments, compliance

• Self-hosted Postgres - Full control over database
• On-premises deployment - Complete infrastructure control
• Security - Hardware Security Modules (HSM), air-gapped networks
• Migration - Deploy to on-premises infrastructure

Migration Process

Step 1: Prepare Database

1. Export schema and data from Supabase
2. Create new database (Cloud SQL, RDS, or self-hosted)
3. Import schema and data
4. Update DATABASE_URL environment variable

Step 2: Deploy Application

1. Build Docker image (if using containers)
2. Deploy to target platform (Cloud Run, ECS, etc.)
3. Configure environment variables
4. Update DNS/routing

Step 3: Verify

1. Test API endpoints
2. Verify database connectivity
3. Check logs and monitoring
4. Update documentation

Security Controls

Encryption

• At rest - Database encryption via KMS (Cloud KMS, AWS KMS)
• In transit - TLS/SSL for all connections
• Secrets - Environment variables, secret managers

Network Isolation

• VPC - Private networking for database and services
• VPC-SC - Google Cloud VPC Service Controls
• Private endpoints - Private Service Connect, VPC endpoints

Access Control

• IAM - Cloud IAM, AWS IAM for service accounts
• Database users - Role-based access control
• API keys - Secure key management

Compliance

• SOC 2 - Available on enterprise cloud providers
• HIPAA - Can be configured on GCP/AWS
• GDPR - Data residency controls

Code Changes Required

Answer: None!

The entire stack uses standard technologies:

• ✅ Fastify - Standard Node.js, no platform-specific code
• ✅ Drizzle - Generates plain SQL, no proprietary runtime
• ✅ PostgreSQL - Standard SQL, works with any Postgres host
• ✅ Environment variables - Only DATABASE_URL needs to change

Related Documentation

• Backend Stack - Technology choices