AWS Cloud Adoption Framework (CAF)

The AWS CAF is a framework that brings AWS experience and best practices to companies preparing to migrate to the AWS Cloud.

The framework provides tools to help accelerate the migration journey, organize resources, and align management during the transition.

Benefits

AWS CAF provides benefits for migrations to reduce business risk and improve sustainability and corporate transparency.
Companies can grow revenue by creating new products and services in their cloud transformation.
They can also reduce operational costs, increase productivity, and improve customer experience in their new cloud environment.

You can use AWS CAF to migrate technology like legacy infrastructure and applications.
You can also use it to migrate and optimize business processes, operations, and even create new business models with the move to the cloud.

Perspective	Tagline	Key capabilities
Business	Cloud investment produce business outcome.	Strategy, portfolio, product, innovation, data monetization
People	Humans able to absorb change.	Culture evolution, cloud fluency, workforce transformation, org design
Governance	Orchestrate initiative, contain risk and cost.	Program mgmt, risk, benefits realization, cloud financial mgmt, data governance
Platform	Build the thing teams ship on.	Platform + data architecture, platform engineering, provisioning, CI/CD, modern app dev
Security	Confidentiality, integrity, availability.	IAM, threat detection, vulnerability mgmt, data protection, incident response
Operations	Run it at agreed service level.	Observability, event/incident/problem, change and release, patch, capacity, continuity

Top split: Business + People + Governance = business capabilities. Platform + Security + Operations = technical capabilities.

Platform + Security + Operations = technical capabilities.

Governance vs Security — Security has its own “security governance” capability; not a duplicate. Governance perspective = enterprise risk, spend, portfolio, benefits. Security governance = scoped to security control ownership and standards.
Platform vs Operations — Platform builds capability (CI/CD pipeline, landing zone). Operations runs what exists. Pipeline design = Platform. Pipeline on fire at 2am = Operations.
Governance owns FinOps — cloud financial management sits in Governance, not Business, not Operations. Most-misassigned capability.
Business vs Governance — Business decide which bet. Governance measure whether bet paid off, and cap risk while it runs.
People vs Governance — People = skill and culture capacity. Governance = decision rights and program control.

Strategy	Tagline	Detail
Retire	Turn it off.	No longer needed. Cheapest win, found during portfolio discovery
Retain	Leave it where it is, revisit later.	Compliance block, recent capex, dependency not ready, no business case
Relocate	Move hypervisor-level, change nothing above it.	VMware estate → VMware Cloud on AWS. No OS touch, no rewrite, same tooling
Rehost	Lift and shift.	Server-by-server into EC2, no code change. AWS MGN
Replatform	Lift, tinker, shift.	Swap components for managed services. MySQL on EC2 → RDS. Architecture unchanged
Repurchase	Drop and shop.	Abandon the code, buy SaaS. On-prem CRM → Salesforce
Refactor / Re-architect	Change the shape of the app.	Monolith → microservices/serverless. Most effort, most benefit

Rehost vs Relocate
- Rehost = instance level, per-server cutover, agent-based replication.
  - Memorize: Resource is switching to a new host.
- Relocate = hypervisor level, whole vSphere estate, no agent, near-zero downtime. Relocate keeps VMware operations intact; rehost lands you on native EC2.
  - Memorize: Resource container is switching to a new location.
  - Location has a host.
Replatform vs Refactor — Test: did the application’s shape change?
- No → replatform.
- Yes → refactor. Managed-service swap is not re-architecture.
Repurchase vs Refactor
- Refactor, you still own the code.
- Repurchase, you throw the code away and rent someone else’s product. License model changes too.
Retain vs Retire — Both “not migrating now,” opposite lifecycle.
- Retain keeps running.
- Retire kills it.
Retire is underused — discovery routinely surfaces dead workloads nobody owns. Fastest cost reduction in the portfolio.

Rough effort/benefit ladder: Retire → Retain → Relocate → Rehost → Replatform → Repurchase → Refactor

Assess
- Migration Evaluator — Make business case for AWS
Mobilize
- Application Discovery Service — Look at on-premises server to create detailed migration plan
- Migration Hub — Track all migration-related information and todos in one convenient hub
Migrate and Modernize
- Application Migration Service — Implement the actual migration

Primary Difference:

AWS DMS handles migrating a database without changing the underlying structure of the database (e.g., MySQL to MySQL).
AWS converts database schemas (e.g., MySQL to PostgreSQL).

These are not opposing tools, but work in tandem to bring a user’s database over to AWS and convert it to a new schema if desired.

AWS DataSync — very fast online data transfer from on-premises to AWS
AWS Transfer Family — transferring files securely
Direct Connect — dedicated private connection between network and virtual private cloud (VPC) in AWS