# Migration and Hybrid Architectures

> Chapter 14 of the Courseiva AWS-CLOUD-PRACTITIONER curriculum — https://courseiva.com/learn/aws-cloud-practitioner/migration-and-hybrid-architectures

**Official objective:** 1.3 — Describe migration and hybrid cloud strategies

## Introduction

Migration and Hybrid Architectures.  These two concepts form the practical bridge between the old world of IT – where you owned every server and cable – and the new world of the cloud.  For anyone studying for the CLF-C02 exam, understanding how businesses actually move to the cloud and why they often keep some things on their own servers is essential, because the exam tests not just what cloud services are, but how real organisations use them over time.

## The House Renovation Analogy

Remember the last time you helped a friend move house?  Or maybe you’ve helped a relative renovate their home.  What if you were asked to move your entire life into a new house – every piece of furniture, every photo album, every box of cables you swear you’ll organise – while still sleeping in the old house until the new one is ready?

That is cloud migration in a nutshell.  In IT, “migration” means moving your digital stuff (your applications, databases, and files) from where it currently lives – usually on physical computers in your own office (called “on-premises”) – into the cloud (like AWS).  But here’s the twist: you can’t just pack a lorry and drive away.  Your business must keep running during the move.  You might keep some things in the old house while you set up the new kitchen.  That state, where part of your house is still the old one and part is the shiny new one, is the “hybrid” part.  In cloud terms, a hybrid architecture means some of your computing resources stay on your own physical servers (the old house), while others run in the cloud (the new house), and they all talk to each other as if they were under one roof.  This lets you take your time moving, or keep certain things (like that antique grandfather clock that’s too fragile to ship) exactly where they are, while still enjoying the benefits of a modern cloud-based system for everything else.  The big question the whole world of migration answers is: “How do I move my entire digital life to the cloud without breaking anything, and how do I decide what stays behind?”

## Core explanation

Let’s start with the big picture.  For decades, if a company wanted to run a website, manage employee emails, or store customer data, they had to buy physical computers (servers), put them in a room (a data centre), pay for electricity and cooling, and hire people to fix them when they broke.  This is called on-premises infrastructure, or “on-prem” for short.  It worked, but it was expensive and inflexible.

Then came the cloud – companies like Amazon Web Services (AWS) who let you rent computing power and storage by the hour or by the minute.  You don’t own the servers; you just use them over the internet and pay for what you consume.  This is where the exam concept of migration comes in.  Migration is the process of moving an organisation’s digital assets – applications, data, and entire IT operations – from the on-premises environment into the cloud.  The exam uses the phrase “migration and hybrid cloud strategies” because very few businesses move everything in one go.  Most keep a foot in both worlds for a while, or even permanently.

Why would a business keep anything on-premises?  Several reasons.  Some older software (legacy applications) simply cannot run in the cloud without being completely rewritten.  Sensitive customer data might be subject to regulations that require it to stay in a particular country or on a specific physical machine.  Some companies have existing contracts for equipment they have already paid for and want to get full use out of before switching.  This mixed state, where some workloads run on your own hardware and others run on AWS, is called a hybrid architecture.  The key concept here is that the cloud and your on-prem systems are connected through a secure network (often a dedicated private connection like AWS Direct Connect, which is like a private fibre-optic cable just for you, or a secure VPN over the public internet).

Now, how do you actually migrate?  AWS defines six strategies commonly known as the “6 Rs” of migration.  You will see these on the exam:

- Rehosting (often called “lift and shift”): You take your application exactly as it is – the same operating system, the same code – and move it onto a virtual server in the cloud (specifically an Amazon EC2 instance, which is simply a virtual computer).  You change almost nothing.  This is the fastest migration method and the most common for the exam.

- Replatforming (“lift, tinker, and shift”): You make a few small optimisations during the move.  For example, you might move your database from a physical server to a managed cloud database service (like Amazon RDS) that handles backups and updates for you, without changing the application code.

- Refactoring/Re-architecting: You redesign the application to take full advantage of cloud-native features, such as breaking a large program into smaller, independent pieces (microservices) that can scale automatically.  This is the most expensive and time-consuming strategy but often delivers the best long-term results.

- Repurchasing: You stop using your current software and switch to a cloud-based version.  The classic example is replacing a custom-built customer management tool with a SaaS (Software as a Service) offering like Salesforce.  You are not moving your data – you are buying a new product.

- Retiring: You identify applications that are no longer useful and simply turn them off.  You do not migrate them at all.  This reduces cost and complexity.

- Retaining: You leave some applications exactly where they are (on-premises) because they are critical, require specialised hardware, or are too risky to move.  This is the “hybrid” part – you retain certain systems while migrating others.

A related key term is the AWS Cloud Adoption Framework (AWS CAF).  This is a set of guidelines (not something you need to memorise in detail, but you should recognise the name) that helps organisations plan their migration journey.  It breaks down the process into perspectives like business, people, governance (rules and policies), platform, security, and operations.

Another important concept is data migration.  Moving terabytes or petabytes of data (a petabyte is a million gigabytes) over the internet can take weeks or months.  For very large datasets, AWS offers the AWS Snow Family – physical devices like the Snowcone or Snowball, which are rugged boxes full of hard drives.  AWS ships them to your office, you load your data onto them, and then ship them back.  AWS then transfers the data directly into their cloud.  This is often faster and cheaper than using the internet.  This is known as offline data transfer.

Finally, the exam wants you to understand that hybrid architectures are not just a temporary phase.  Many organisations run permanent hybrid setups.  For example, a factory might have a control system that runs on a local server because it must respond instantly (low latency) and cannot afford any internet delay.  Meanwhile, their HR and payroll systems run in the cloud.  These two environments communicate securely.  This concept is also tied to the idea of edge computing, where some processing happens close to where data is generated (like in a factory or a store) rather than sending everything to the cloud.  AWS has services like AWS Outposts that literally place AWS hardware inside your own data centre, giving you a consistent hybrid experience.  On the exam, you will be asked to identify the best migration strategy for a given scenario, and to distinguish between hybrid, on-premises, and all-in cloud solutions.

## Real-world context

Imagine a mid-sized company called “GreenLeaf Books,” a chain of brick-and-mortar bookshops that also sells online.  For the past twenty years, GreenLeaf has run its entire business on servers in the basement of its headquarters.  Their systems include a customer database (which stores loyalty card information), an inventory system (tracking every book in every store), a payment processing system, and a staff scheduling tool.

GreenLeaf is growing fast and the basement server room is running out of space.  The cooling system keeps breaking, and the old server that handles the online store is so slow that customers are abandoning their shopping carts.  The IT director decides it is time to migrate to AWS.  Here is exactly what an IT professional at GreenLeaf would do, step by step:

First, they would perform a discovery and assessment.  This involves inventorying every single application and piece of data in the company.  They use tools like AWS Migration Hub and AWS Application Discovery Service to automatically map out which servers talk to which, how much CPU (computing power) and memory each uses, and what dependencies exist.  For example, they discover that the inventory system (running on a very old version of Windows) cannot be simply lifted because it requires a specialised piece of hardware called a “dongle” that plugs into the server.  This application must be retained on-premises or replaced entirely.

Next, they plan the migration strategy using the 6 Rs.  They decide:

- Rehost (lift and shift) the online store’s web server and the payment processing system.  These are well-behaved applications that will run fine on Amazon EC2 instances.
- Replatform the customer database by moving it from the on-premises Microsoft SQL Server to Amazon RDS for SQL Server.  This is a small change (they pick a managed service) but it removes the need for GreenLeaf to manage backups and patching.
- Retire three old report-generating applications that nobody uses anymore.
- Retain the inventory system because it relies on the physical dongle.  This becomes a permanent hybrid architecture – the inventory system stays on-premises, but everything else moves to the cloud.

Then comes the actual migration.  GreenLeaf sets up a secure VPN connection between their office network and their new AWS Virtual Private Cloud (VPC) – think of a VPC as your own private, isolated section of the AWS cloud.  For the lift-and-shift applications, they use AWS Server Migration Service (SMS) to replicate the live servers to the cloud.  Once the copies are tested and working, they switch all user traffic to the cloud copies in a cutover event (often done over a weekend to minimise disruption).  For the customer database, they use AWS Database Migration Service (DMS) which continuously copies data from the old database to the new one, keeping them in sync until the moment of the final switch.

After the migration, the IT team monitors everything using AWS CloudWatch (for performance alerts) and AWS CloudTrail (for auditing who did what).  They also set up AWS Backup to automatically back up the cloud databases.  The retained on-premises inventory system now talks to the cloud systems through the secure VPN – this is the hybrid architecture in action.  The inventory system sends stock updates to the cloud customer database, and the online store checks stock levels against the on-premises system in real time.

In this scenario, the IT professional spent months on planning, used several AWS migration tools, made strategic decisions about what to keep and what to move, and left the company with a permanent hybrid setup.  This is exactly the kind of real-world story the CLF-C02 exam expects you to understand at a conceptual level.

## Exam focus

The CLF-C02 exam tests “Migration and Hybrid Architectures” in a specific way.  You will not be asked to write a migration plan or configure a hybrid connection.  Instead, the exam presents scenario-based multiple-choice questions that check whether you can identify the correct concept, strategy, or tool for a given situation.  Here is exactly what you need to know.

First, the exam loves the 6 Rs of Migration.  You must be able to read a short description of a company’s situation and pick the right R.  The most common exam pattern is this: a company wants to move to the cloud quickly with minimal changes.  The correct answer is Rehost (lift and shift).  If the description says they want to make a minor improvement like using a managed database service, the answer is Replatform.  If they want to completely redesign the application to use serverless or microservices (small, independent functions), the answer is Refactor/Re-architect.  If they are going to ditch their old software and buy a cloud-based version, the answer is Repurchase.  If they discover that an application is no longer needed, the answer is Retire.  If they have a good reason to keep something on-premises (compliance, legacy hardware, latency), the answer is Retain.

Second, the exam tests the difference between online and offline data transfer.  Trap question: “A company needs to transfer 200 TB of data to AWS.  The internet connection is slow and the project has a tight deadline.”  A beginner might pick “use a faster internet connection” or just shrug.  The correct answer is to use an AWS Snowball device (a physical data transfer device).  The exam expects you to know that for very large datasets (typically over 10 TB or when the internet is slow), offline transfer via the Snow Family is the recommended approach.

Third, hybrid architectures appear in questions that contrast “all-in cloud” vs “on-premises” vs “hybrid.”  A common trap: they describe a scenario where a factory needs a control system with extremely low latency (less than 1 millisecond) and suggest using the cloud.  The correct answer is to retain that system on-premises or use AWS Outposts (which puts AWS hardware in your facility).  The exam wants you to understand that hybrid architectures solve specific problems where the cloud alone cannot meet requirements like latency, data residency (data must stay in a particular country), or compatibility with old hardware.

Fourth, you need to know what the AWS Cloud Adoption Framework (AWS CAF) is.  You do not need to memorise all six perspectives (Business, People, Governance, Platform, Security, Operations are the ones you will see on the exam).  However, you should recognise that AWS CAF is a guide that helps organisations plan their cloud migration and adoption journey.  The exam may ask, “Which AWS framework helps a company think about the people and process changes needed for a cloud migration?”  The answer is AWS CAF.

Fifth, the exam tests the concept of the AWS Snow Family.  Memorise these names:

- AWS Snowcone: The smallest device, for small amounts of data (up to 8 TB).
- AWS Snowball Edge: A larger device (up to 80 TB) that can also run compute (processing) jobs.
- AWS Snowmobile: A massive shipping container on a truck that can transfer up to 100 PB – yes, petabytes.  This is for truly enormous data centre moves.

Common exam traps:

- Trap: “The company has a slow internet connection and needs to move 1 GB of data.”  The answer is NOT a Snowball – that would be overkill.  For small amounts of data, just use the internet (AWS CLI or AWS DataSync).  Snow devices are for large volumes.
- Trap: Confusing “hybrid” with “multi-cloud.”  Hybrid means on-prem + cloud.  Multi-cloud means using two or more cloud providers (e.g. AWS and Azure).  The exam distinguishes them clearly.
- Trap: Thinking that migrated data is automatically secure.  The exam tests that you must still configure security (encryption, IAM roles) after migration.

Finally, the exam may include a question about the AWS Direct Connect vs VPN.  Direct Connect is a dedicated private physical connection between your office and AWS – it is more reliable and has consistent speed, but costs more.  A VPN runs over the public internet and is cheaper but less consistent.  If a question mentions “consistent network performance” or “dedicated private connection,” pick Direct Connect.

To summarise the exam focus: memorise the 6 Rs, the Snow Family names and use cases, the definition of hybrid, and the purpose of AWS CAF.  That covers 90% of the questions on this topic.  Practice scenario questions until you can spot the R instantly.

## Step by step

1. **Step 1: Discovery and Assessment** — Inventory all existing applications, data, and infrastructure.  Use tools like AWS Application Discovery Service to understand dependencies between systems, resource utilisation, and what workloads are candidates for migration.  This step identifies which applications will be migrated, retained, or retired.
2. **Step 2: Planning a Migration Strategy** — For each workload, decide which of the 6 Rs to apply.  Prioritise based on business goals: speed (Rehost), optimisation (Replatform), or modernisation (Refactor).  Also plan the network connectivity – either a VPN over the internet or a dedicated AWS Direct Connect link – and decide on a timeline for the cutover.
3. **Step 3: Setting Up the Target Environment** — Create the cloud infrastructure that will host the migrated workloads.  This includes defining a Virtual Private Cloud (VPC), subnets, security groups (firewalls), IAM roles for access control, and creating the target EC2 instances or database services (like Amazon RDS).  Test the network connection between on-premises and the cloud.
4. **Step 4: Executing the Migration** — Use AWS migration tools to move data and applications.  For servers, AWS Server Migration Service (SMS) can replicate live on-premises servers to the cloud.  For databases, AWS Database Migration Service (DMS) continuously syncs data until the cutover.  For large datasets, ship an AWS Snowball device to load data physically.  After replication, perform a cutover by redirecting user traffic to the new cloud environment.
5. **Step 5: Validation and Optimisation** — Test that all applications work correctly in the cloud environment.  Monitor performance, check security configurations, and verify data integrity.  After validation, decommission the original on-premises resources if they are no longer needed.  For retained systems, ensure the hybrid connection (VPN or Direct Connect) is stable and secure.  Finally, use cloud services like Auto Scaling and CloudWatch to optimise cost and performance.

## Comparisons

### Hybrid Cloud vs Multi-Cloud

**Hybrid Cloud:**
- Uses both on-premises infrastructure and one cloud provider (e.g., AWS)
- The on-premises part may be required for latency or compliance reasons
- Communication between on-prem and cloud uses VPN or Direct Connect

**Multi-Cloud:**
- Uses two or more public cloud providers (e.g., AWS and Azure)
- No on-premises infrastructure is involved
- Each cloud provider is used for different applications or redundancy

### AWS Snowball Edge vs AWS DataSync

**AWS Snowball Edge:**
- Physical device shipped to your location for offline data transfer
- Ideal for very large datasets (over 10 TB) or slow internet
- Can also run compute (processing) jobs on the device

**AWS DataSync:**
- Online data transfer tool that moves data over the internet or Direct Connect
- Ideal for smaller datasets or regular syncs
- Automatic encryption and validation of data in transit

### Rehost (Lift and Shift) vs Refactor (Re-architect)

**Rehost (Lift and Shift):**
- Application is moved with zero changes to code or configuration
- Fastest migration approach – minimal planning required
- Does not automatically bring cloud-native benefits like auto-scaling

**Refactor (Re-architect):**
- Application is redesigned to use cloud-native services (e.g., Lambda, DynamoDB)
- Most time-consuming and expensive migration approach
- Delivers long-term benefits like scalability, resilience, and lower operational overhead

### AWS Cloud Adoption Framework (AWS CAF) vs AWS Well-Architected Framework

**AWS Cloud Adoption Framework (AWS CAF):**
- Focuses on the process of moving to the cloud, including people and business changes
- Six perspectives: Business, People, Governance, Platform, Security, Operations
- Used to plan and manage cloud migration projects

**AWS Well-Architected Framework:**
- Focuses on designing and reviewing cloud architectures after migration
- Six pillars: Operational Excellence, Security, Reliability, Performance Efficiency, Cost Optimization, Sustainability
- Used to evaluate and improve existing cloud workloads

## Common misconceptions

- **Misconception:** The cloud is always cheaper than on-premises. **Reality:** Cloud can be cheaper, but not always.  Costs depend on usage, instance sizes, data transfer fees, and whether the workload is predictable.  Sometimes a long-running, predictable workload is cheaper on a dedicated on-premises server than on a cloud instance that runs 24/7. (Cloud providers market cost savings heavily, so beginners assume it is universally cheaper.  The exam tests that you understand migration decisions are case-by-case.)
- **Misconception:** If you move an application to the cloud using lift and shift (rehost), you automatically get scalability and high availability. **Reality:** Lift and shift moves the application exactly as-is.  If the on-premises application was designed to run on a single server with no redundancy, it will still be a single point of failure in the cloud unless you specifically redesign it or add additional services like Auto Scaling and Elastic Load Balancing. (People hear ‘cloud is scalable’ and assume it applies to everything automatically.  The exam sets traps where a lift-and-shift application still crashes under high load because the company did not configure scaling.)
- **Misconception:** Hybrid architecture means you are using two different cloud providers at the same time. **Reality:** Hybrid architecture means you have some resources on your own on-premises servers and some in the cloud (any cloud provider).  Using two cloud providers (e.g., AWS and Azure) without any on-premises is called multi-cloud, not hybrid. (The term ‘hybrid’ sounds like a mix of two things, so beginners assume it means mixing clouds.  The exam explicitly tests the difference between hybrid and multi-cloud.)
- **Misconception:** Once you migrate to the cloud, you can delete all your on-premises equipment immediately. **Reality:** Many organisations retain on-premises systems permanently for reasons like old software that cannot be migrated, compliance requirements, or latency-sensitive applications.  You should plan a phased migration and only decommission on-premises hardware after verifying that the cloud versions are stable and meet all requirements. (Beginners think migration is an all-or-nothing switch.  The exam tests that retention is a valid strategy (the 6 Rs include ‘Retain’).)
- **Misconception:** AWS Snowball devices are only used for data transfer. **Reality:** Snowball Edge devices can also run compute jobs (processing data on the device itself) using Amazon EC2 instances or AWS Lambda functions.  They are useful for processing data locally when internet connectivity is poor, and then shipping the results to the cloud. (The word ‘data transfer’ is in the product description, so beginners miss the compute capability.  The exam sometimes asks about edge computing use cases for Snowball Edge.)
- **Misconception:** You must move all applications to the cloud to benefit from AWS services. **Reality:** You can use many AWS services from your on-premises environment as long as you have a network connection.  Services like Amazon S3 (cloud storage) and AWS IAM (identity and access management) can be used in a hybrid setup.  AWS also offers AWS Storage Gateway, which makes cloud storage appear as a local drive to on-premises servers. (People assume the cloud is a separate world you must move into.  The exam tests that hybrid architectures allow you to use cloud services while keeping some data or applications on-premises.)

## Key takeaways

- The 6 Rs of Migration are Rehost, Replatform, Refactor, Repurchase, Retire, and Retain – each describes a different strategy for moving or handling applications during a cloud migration.
- Hybrid architecture means running some workloads on your own on-premises servers and others in the cloud, connected by a secure network.
- AWS Snow Family (Snowcone, Snowball Edge, Snowmobile) provides physical devices for offline data transfer when internet uploads are too slow or expensive.
- AWS Direct Connect establishes a dedicated private network connection from your data centre to AWS, offering consistent performance and higher security than a VPN over the internet.
- The AWS Cloud Adoption Framework (AWS CAF) provides guidance across six perspectives (Business, People, Governance, Platform, Security, Operations) to help organisations plan their cloud migration.
- Data migration can happen online via the internet (using AWS CLI, AWS DataSync, or AWS Database Migration Service) or offline using physical devices from the Snow Family.
- Retaining some applications on-premises is a valid strategy (the 6th R) for reasons of compliance, latency, or legacy system dependencies.
- A lift-and-shift (Rehost) migration does not automatically bring cloud benefits like scalability or cost savings – you must configure those features separately.

## FAQ

**What is the difference between rehosting and replatforming in AWS migration?**

Rehosting (lift and shift) moves an application to the cloud without any changes.  Replatforming makes minor adjustments during the move, like switching a database to a managed service (Amazon RDS) without changing the application code.

**When should I use AWS Snowball instead of just uploading data over the internet?**

Use AWS Snowball when you have more than 10 TB of data to transfer, or when your internet connection is slow, unreliable, or expensive.  Shipping a physical device is often faster and more cost-effective than uploading over the network.

**What does ‘hybrid cloud’ mean in the AWS context?**

Hybrid cloud (or hybrid architecture) means you run some of your IT resources on your own on-premises servers and some on AWS, connecting them securely.  It is not the same as using multiple cloud providers (that is multi-cloud).

**Is AWS Direct Connect the same as a VPN?**

No.  AWS Direct Connect is a dedicated physical private connection between your data centre and AWS, offering consistent network performance and higher security.  A VPN runs over the public internet and is cheaper but less reliable and has variable performance.

**Can I keep any of my on-premises servers after moving to AWS?**

Yes.  Retaining some servers on-premises is one of the 6 Rs of migration.  You might do this for compliance (data must stay in a specific country), latency (a factory control system that needs instant response), or because you have old software that cannot run in the cloud.

**What is the AWS Cloud Adoption Framework (AWS CAF)?**

AWS CAF is a set of guidelines that helps organisations plan and execute their cloud migration.  It covers six areas (perspectives): Business, People, Governance, Platform, Security, and Operations.  You do not need to memorise the details, just remember its purpose.

**What does ‘lift and shift’ mean in cloud migration?**

Lift and shift (also called rehosting) means taking an application exactly as it is – the same operating system, same configuration, same code – and moving it to a cloud server (EC2 instance) without any modifications.  It is the fastest way to migrate.

---

Interactive version with quiz and diagrams: https://courseiva.com/learn/aws-cloud-practitioner/migration-and-hybrid-architectures