How does one company run websites, apps, and databases for customers all over the world without crashing every time a new Taylor Swift concert ticket sale starts? For the CLF-C02 exam, understanding this is vital because AWS's global infrastructure is the physical backbone that makes everything possible. It explains why your website loads quickly in Sydney and London, and how your data stays safe even if an entire data centre floods. Without this foundation, the cloud would just be someone else's computer in a single room.
Jump to a section
Have you ever ordered food from a takeaway chain and wondered how your pizza arrives hot, even though the restaurant is miles away?
That takeaway chain is a lot like the AWS global infrastructure. The chain doesn't have just one kitchen. It has them all over the city, and even in other towns. When you order, the system automatically sends your order to the kitchen nearest to you. Your pizza is cooked fresh, stays hot during the short journey, and arrives quickly. If that kitchen is closed or too busy, another one just down the road takes over without you noticing.
Now, think about the chain's menu. Every kitchen uses the same recipes (services) and ingredients (data), but they are independent. If one kitchen runs out of cheese, the others keep cooking. The chain also has a central head office that decides what new dishes to add and ensures every kitchen meets health standards (compliance).
In AWS terms, each kitchen is a Region. The delivery drivers are the edge locations (points of presence) that bring data closer to you. The head office is the AWS Management Console and shared control plane. Just like you don't care which kitchen cooked your pizza as long as it arrives hot, in the cloud, you don't care which server runs your app as long as it works fast and reliably. The global infrastructure hides all that complexity behind a simple interface.
The AWS global infrastructure is the physical network of data centres, servers, and networking equipment that Amazon Web Services uses to deliver its cloud services worldwide. It is not one big computer in a warehouse. It is a carefully designed system spread across the planet to ensure speed, reliability, and security.
Let's break it down from the largest piece to the smallest.
Regions are the biggest building blocks. A Region is a separate geographic area, like 'US East (N. Virginia)' or 'EU (Ireland)'. Each Region is completely independent from every other Region. They do not share power, cooling, or network connections. This is by design. If a natural disaster hits one Region (say an earthquake in Tokyo), the services in another Region (like Singapore) keep working without interruption. When you create an AWS resource, like a virtual server (called an Amazon EC2 instance), you must choose which Region to put it in. This affects latency (how fast data travels) and cost.
Inside each Region, there are multiple Availability Zones (AZs) . An AZ is one or more discrete data centres, each with redundant power, networking, and connectivity, housed in separate facilities. They are located close enough to provide low latency (under 10 milliseconds) between them, but far enough apart to be isolated from common failures like a power outage or a fire. Think of an AZ as a single, very reliable data centre building. A Region typically has a minimum of three AZs. This is crucial for high availability. You can run your application in two AZs so that if one data centre fails, the other instantly takes over. AWS services like Amazon RDS (relational database) and Elastic Load Balancing are designed to spread traffic across multiple AZs automatically.
Then we have Edge Locations (Points of Presence) . These are smaller sites, located in hundreds of cities around the world. They are not full Regions. Their main job is to cache (store copies of) content, like website images, videos, and APIs, so that when a user in Mumbai requests your website, the data comes from an Edge Location in Mumbai rather than travelling all the way from your server in Virginia. This service is called Amazon CloudFront, a Content Delivery Network (CDN). Edge Locations also help speed up DNS lookups (how your browser finds the IP address of a website) through Amazon Route 53, AWS's Domain Name System (DNS) service.
Finally, there are Regional Edge Caches. These sit between Edge Locations and Regions. They hold more content than a regular Edge Location, reducing the load on the origin Region.
Why does AWS build all this? Three reasons.
Performance (Latency): Users get data from the closest server, so web pages load faster.
High Availability and Fault Tolerance: If one data centre or Region fails, your application still runs from another location.
Compliance and Data Sovereignty: Many countries have laws requiring that customer data stays within that country's borders. AWS Regions allow you to keep data in Germany, for example, without it ever leaving the EU.
Before the cloud, a company would have to build its own data centre in one location. If that data centre had a power cut, the whole business went offline. AWS's global infrastructure replaces that single point of failure with a distributed, resilient network that big companies like Netflix and small startups rely on equally.
Choose your Region
When you sign into the AWS Management Console, you see a Region selector in the top right corner. You pick the Region physically closest to your end users to minimise latency. For example, if your users are in Australia, choose Asia Pacific (Sydney). This decision affects where your data will be stored and where your compute resources run.
Deploy across multiple Availability Zones
Inside your chosen Region, you launch your application (e.g., EC2 instances) in at least two different AZs. An Application Load Balancer distributes traffic between them. If one AZ experiences an outage, the load balancer automatically sends traffic only to the healthy AZ. This is the foundation of high availability in AWS.
Set up a Content Delivery Network
You create a CloudFront distribution and point it to your origin server (your EC2 instance or S3 bucket). CloudFront automatically replicates your static and dynamic content to Edge Locations around the world. When a user in London requests your website, CloudFront serves it from the London Edge Location, reducing load time dramatically.
Configure DNS routing
You use Amazon Route 53 to manage your domain name. You create a latency-based routing policy. This means when a user types your domain, Route 53 checks which Region gives the lowest latency for that user and directs them to that Region's IP address. This ensures each user reaches the fastest available server.
Plan for disaster recovery across Regions
To protect against a full Region outage, you optionally replicate your data (using services like DynamoDB Global Tables or S3 Cross-Region Replication) to a second Region. You configure Route 53 with a failover routing policy. If the primary Region health check fails, traffic automatically shifts to the secondary Region, keeping your application online.
You work for a startup that built a mobile game called 'Silly Squirrel Dash'. When a player in Japan opens the game, it takes 8 seconds to load the squirrel's acorn collection screen. Players are quitting. Your boss asks you to fix it.
Here is what you actually do using the AWS global infrastructure.
First, you check where your application servers are running. You find they are in a single EC2 instance in the US East (N. Virginia) Region. That is why Japanese players experience high latency. You decide to deploy the game's backend servers to the Asia Pacific (Tokyo) Region, closer to your Japanese users. - You create a new EC2 instance in the Tokyo Region. - You set up an Application Load Balancer that distributes incoming traffic between the Virginia and Tokyo instances. - You configure Amazon Route 53 with a latency-based routing policy. Now, when a player in Japan makes a DNS query, Route 53 automatically directs them to the Tokyo server, which is faster.
Next, you tackle the static content — the acorn images and sound files. You store these in an Amazon S3 bucket in the Virginia Region. This is still slow for Japan. - You enable Amazon CloudFront, a CDN, and point it to your S3 bucket. - CloudFront distributes copies of your game assets to Edge Locations worldwide, including Tokyo, Osaka, and Seoul. - Now, when a player in Japan downloads the acorn image, it comes from the local Edge Location, not Virginia. Loading time drops from 8 seconds to under 1 second.
Finally, you ensure reliability. Your game is now running in two Regions. But you want to be sure that if the entire Tokyo Region experiences a disaster, the game still works. - You set up an active-passive failover architecture. Under normal conditions, both Regions run. If CloudWatch (AWS monitoring service) detects that the Tokyo instance is unhealthy, Route 53 automatically reroutes all traffic to Virginia. - You replicate the game's database (Amazon DynamoDB) across Regions using global tables, so player scores are not lost.
The result: players everywhere get a snappy experience, and your boss is happy because the game did not go offline during a recent earthquake in Japan.
The CLF-C02 exam loves testing your understanding of the hierarchy and purpose of each component of the global infrastructure. You will not be asked to configure anything, but you must know the definitions and why they matter.
Here are the specific concepts the exam tests repeatedly.
Region vs. Availability Zone vs. Edge Location: You must be able to identify which is which from a description. A common trick: they describe an Edge Location but ask you to name a Region. Memorise: Regions are for running services and storing data; AZs are for high availability within a Region; Edge Locations are for caching content to reduce latency.
How many AZs does a Region have? The correct answer is 'at least two' or 'a minimum of two' (though most have three or more). AWS never says exactly how many, only that it is at least two.
Which services are global vs. regional? Services like IAM (Identity and Access Management), Route 53 (DNS), and CloudFront are global — they work across all Regions without you choosing one. Services like EC2, S3, and RDS are regional — you pick a Region when you create them. The exam will give you a list of services and ask which are global.
Data residency and compliance: You will get a scenario question: 'A company in France must keep all customer data within the European Union. Which AWS feature should they use?' The answer is 'choose the EU (Frankfurt) or EU (Ireland) Region.' Do not pick 'Edge Locations' — Edge Locations cache data but do not store the primary copy.
Disaster recovery and high availability: The exam asks how to make an application fault-tolerant. The correct answer nearly always involves using multiple Availability Zones within a single Region, or multiple Regions for cross-region disaster recovery. A single EC2 instance in one AZ is the wrong answer for any question about high availability.
Latency reduction: If the question is about speeding up content delivery to global users, the answer is CloudFront with Edge Locations. Do not pick 'adding more EC2 instances in one Region' — that does not help users far away.
Trap patterns to watch for: the exam might describe a scenario where a company has customers in Brazil and wants to reduce latency. A wrong answer might say 'deploy to the US East Region because it is the largest'. The correct answer is 'deploy to the South America (Sao Paulo) Region' or 'use CloudFront'. Another trap: confusing a Regional Edge Cache with an Edge Location. Remember, Edge Locations are smaller and more numerous; Regional Edge Caches are larger and sit between the Edge and the Region.
Key definitions to memorise for the exam: - Region: A physical location in the world where AWS has multiple AZs. - Availability Zone (AZ): One or more discrete data centres with redundant power and networking within a Region. - Edge Location: A site that CloudFront uses to cache copies of content for faster delivery. - Regional Edge Cache: A larger cache that sits between an Edge Location and the origin server.
A Region is a separate geographic area with at least two Availability Zones for fault tolerance.
An Availability Zone is one or more discrete data centres with independent power and networking within a Region.
Edge Locations cache content to reduce latency for users worldwide and do not run general-purpose compute.
Latency-based routing with Route 53 directs users to the closest Region for the fastest experience.
Data transfer between Regions always incurs cost, even within the same AWS account.
Global services like IAM and Route 53 do not require a Region selection; regional services like EC2 and S3 do.
These come up on the exam all the time. Here's how to tell them apart.
Region
A large geographic area (e.g., US East, EU Ireland).
Contains at least two Availability Zones.
Used for data residency and latency decisions.
Availability Zone
One or more data centres within a Region.
Isolated from other AZs for fault tolerance.
Used for high availability by deploying across multiple AZs.
Edge Location
Smaller site in hundreds of cities globally.
Caches content for CloudFront only.
Reduces latency for static and dynamic content.
Regional Edge Cache
Larger cache between Edge Locations and Regions.
Holds more data to reduce load on origin servers.
Used when content is not found in an Edge Location.
Global Service (e.g., IAM)
No Region selection required; works everywhere.
Data is stored in a single global region (US East).
Accessible from any Region without replication.
Regional Service (e.g., EC2)
Must choose a Region when creating the resource.
Data stays within the chosen Region unless explicitly moved.
Can be replicated across Regions for disaster recovery.
Mistake
An Availability Zone is the same as a single data centre building.
Correct
An Availability Zone is one or more discrete data centres. It can be a single building or a cluster of buildings close together, but it is isolated from other AZs.
The word 'zone' sounds singular, and AWS documentation sometimes says 'data centre' loosely.
Mistake
All AWS Regions are connected to each other with a direct private network link, so data travels between them for free.
Correct
AWS Regions are connected via the AWS global network backbone, but data transfer between Regions costs money. It is not free.
The fact that AWS has a private backbone makes people assume it is included, but inter-Region traffic is billed per gigabyte.
Mistake
If I launch an EC2 instance in one Region, I can access it from any other Region with the same low latency.
Correct
Latency increases significantly when accessing a resource from a distant Region. To get low latency, you must deploy resources in a Region close to your users.
New learners often think 'the cloud' is a single omnipresent entity, not a collection of geographically separate data centres.
Mistake
Edge Locations are mini data centres where you can run compute or store databases.
Correct
Edge Locations mainly cache content for CloudFront and accelerate DNS lookups. You cannot run general-purpose compute (like EC2) there. For compute at the edge, you use AWS Wavelength or AWS Outposts, not standard Edge Locations.
The word 'location' sounds like a general-purpose facility, and Amazon uses 'edge' in many product names confusingly.
Mistake
You must choose a Region when you sign up for AWS, and you can only use services in that Region.
Correct
Your AWS account is global. You can use any Region at any time. The Region is chosen per service resource, not per account.
Many cloud providers lock you to a specific region at signup. AWS does not, but this confuses beginners.
Reveal each answer, then mark whether you got it right. Score 60%+ to unlock the next chapter.
As of 2025, there are over 30 AWS Regions launched. The exam does not require you to memorise the exact number, but you should know they are spread across North America, South America, Europe, Asia Pacific, and the Middle East.
You cannot move a running EC2 instance across Regions. You must create an Amazon Machine Image (AMI) of the instance, copy that AMI to the target Region, and then launch a new instance from the copied AMI there.
Using one Region saves on data transfer costs between Regions, but may increase costs for your users due to higher latency, leading to slower applications and potentially lost business. The cost-benefit depends on where your users are located.
A Region is a large geographic area that contains at least two Availability Zones. An Availability Zone is a smaller facility (one or more data centres) inside that Region. You run your services in one or more AZs within a Region.
No. Standard Edge Locations only cache content (CloudFront) and handle DNS (Route 53). For compute at the edge, AWS offers AWS Wavelength (for 5G networks) and AWS Outposts (for on-premises), which are different services.
Services in other Regions continue to run normally. If your application is only deployed in that one Region, it will be unavailable until AWS restores the Region. To protect against this, you design your architecture to failover to a second Region.
You've just covered AWS Global Infrastructure — now see how well it sticks with free CLF-C02 practice questions. Full explanations included, no account needed.
Done with this chapter?