Letters travel through Ghana every day. From Accra HQ, to district post offices, all the way to your village kiosk. The system isn't random โ it's carefully designed. Same letter, multiple paths. Same building, multiple branches. That's exactly how AWS designs the global cloud.
Follow the letterGhana Post has its headquarters in Accra. It's a massive operation โ sorting machines, mail trucks, hundreds of staff. Everything that happens with mail in this country gets coordinated from this one geographic area.
But the HQ isn't ONE building. It's actually a complex of three separate post offices spread across Accra โ one in Osu, one in Tema, one in Kasoa. They're close enough to share trucks and routing, but far enough apart that if one floods, the others keep running.
That's an AWS RegionAn AWS Region is a geographic area like 'us-east-1' (Virginia) or 'eu-west-1' (Ireland). Each region is fully isolated and contains multiple data centers.. A geographic area where AWS has built a cluster of data centers. Cape Town. London. Tokyo. Each region is independent โ what happens in London stays in London.
You pick a region close to your users so your app feels fast. Build in Cape Town for African users. Build in London if your customers are mostly in the UK.
Look closer at one of those post offices. Osu has its own building, its own generators, its own water supply, its own backup power, its own staff. If a fire breaks out in Osu, the Tema branch doesn't even notice โ they have their own everything.
But they're connected by fast, dedicated roads. A letter sorted in Osu can be in Tema within minutes. They're separate enough to fail independently, close enough to work together.
That's an Availability ZoneAn Availability Zone (AZ) is one or more discrete data centers within a Region, with redundant power, networking, and connectivity. AZs in a region are physically separated by meaningful distance. in AWS. Each AZ has its own buildings, power, cooling, and networking. They're typically tens of kilometers apart in different parts of a city.
You always launch your servers in at least two AZs. That way, if one data center has a power cut, your app keeps running from the other one.
Storm hits Osu. The roof leaks, the generator dies, the sorting machine shorts out. The Osu post office is down for the day.
But the letters don't stop. Tema and Kasoa pick up the slack. People drop their mail at Tema instead. Letters that were headed FOR Osu get rerouted through Kasoa. Service slows down a bit but nothing stops.
This is why AWS designs apps to span multiple AZs. If your servers run in just one AZ and that AZ has an issue (yes, it happens โ it's rare but real), your app dies. Spread across two or three AZs, and one failure becomes a hiccup, not a disaster.
The pattern is called Multi-AZ deploymentMulti-AZ means deploying your application across two or more Availability Zones, so that if one AZ fails, the others can continue serving traffic. This is the foundation of high availability in AWS.. It's the foundation of high availability.
Exam tip: "High availability" almost always means Multi-AZ. "Disaster recovery" usually means Multi-Region.
Now imagine you live in a small village in the Volta Region. The HQ is in Accra โ that's hundreds of kilometers away. Every time you want to send a letter, do you travel to Accra? Of course not.
Ghana Post has a small drop-off kiosk right at the village junction. You drop your letter there. A truck swings by twice a day, picks up everything, and runs it to the regional hub. Same for incoming mail โ it gets pre-sorted at the regional hub and trucked TO the kiosk for you to pick up.
That kiosk doesn't have sorting machines, generators, or a hundred staff. It's just a small outpost CLOSE TO YOU. It exists for one reason: speed of access.
That's an Edge LocationEdge locations are AWS sites that cache content close to end users for low-latency delivery. There are over 600 globally โ far more than there are AWS Regions. in AWS. Hundreds of small outposts scattered around the world โ one in Lagos, one in Nairobi, one in Mumbai. They cache popular content (videos, images, web pages) close to users so the experience feels fast.
This is what powers CloudFront โ AWS's content delivery network. Your video doesn't load from London every time. It loads from the edge location nearest you.
Ghana isn't the only place with mail. Nigeria has its own postal HQ. South Africa has its own. The UK has theirs. Each country runs its own postal system, with branches and kiosks.
Sometimes a letter has to cross between them โ a Ghana โ UK letter goes through ports, planes, customs. But internally, each country's postal service operates on its own.
AWS works exactly the same way. There are AWS Regions all over the world โ Ireland, Tokyo, Sydney, Cape Town, Sรฃo Paulo, Mumbai, Bahrain, and more being added every year. Each is independent. Each has multiple AZs. Each has Edge Locations near it.
When you build an app, you choose:
โข Which region to host it in (close to your main users)
โข Which AZs to spread across (for failure protection)
โข Whether to use edge locations (CloudFront) for global users
When you read "deploy your app to us-east-1 across two AZs and serve via CloudFront" โ picture Ghana Post. The HQ. The branches. The village kiosks. It all clicks.
A geographic area with multiple data centers. You pick one close to your users.
One physical building. Own power, own staff, own everything. Spread your app across 2+ for safety.
If one branch fails, others keep running. The foundation of high availability in AWS.
A small outpost close to users for fast access. CloudFront uses these to deliver content globally.
"AWS isn't a single building. It's a postal network. Knowing the network is half the cert."
Read the story again