AWS Deep Dive & Global Infrastructure
AWS runs data centres all over the world. To design anything on AWS, you must understand how that world is organised: Regions, Availability Zones and edge locations. We also tour the main service groups and see how scaling saves money on a big sale day.
In this session
- What A-W-S means and the official definition
- Why AWS: availability, scale, pay-as-you-go
- Regions, Availability Zones and edge locations
- How to choose a Region
- The six service groups in the console
- A sale-day scaling example
01What does AWS stand for?
| Letter | Means | In simple words |
|---|---|---|
| A | Amazon | the company behind it (also Prime, Amazon Pay, Alexa) |
| W | Web | delivered over the internet |
| S | Services | computing, storage, networking, databases and more |
| Who uses AWS | Examples |
|---|---|
| Individuals | students, developers, researchers learning or running small projects |
| Businesses | startups, large companies, banks, governments (Netflix, Airbnb, many Indian unicorns) |
02Regions, Availability Zones and edge locations
AWS groups its data centres like this:
| Piece | What it is | India example |
|---|---|---|
| Region | a separate geographic area; your data stays there unless you move it | ap-south-1 Mumbai, ap-south-2 Hyderabad |
| Availability Zone | isolated location(s) inside a Region; spread your servers over 2+ AZs | ap-south-1a, 1b, 1c |
| Edge location | small sites in many cities that cache content close to users (CloudFront) | Mumbai, Delhi, Chennai, Bengaluru, Hyderabad, Kolkata and more |
Like a bank’s branches. A Region is a city; each AZ is a separate branch building in that city with its own power backup. If one branch floods, the others keep working. Edge locations are ATMs placed close to customers for quick service.
03How to choose a Region
| Ask | Why |
|---|---|
| Where are my users? | closer Region = lower latency (faster app). Indian users → Mumbai or Hyderabad |
| Do laws require data to stay in a country? | data residency and compliance (e.g. banking and government data) |
| Is the service available there? | new services reach some Regions first |
| What does it cost? | prices differ a little between Regions |
04A tour of the main service groups
| Group | Examples | Used for |
|---|---|---|
| Compute | EC2, Lambda | running your code on servers or serverless |
| Networking | VPC, Route 53, Elastic Load Balancing | private networks, DNS, spreading traffic |
| Identity & security | IAM | users, roles and permissions |
| Storage | S3, EBS, EFS | files, disks, backups |
| Database | RDS, DynamoDB | SQL and NoSQL databases |
| Monitoring | CloudWatch, CloudTrail | metrics, logs, alarms, audit trail |
Sale-day example. On normal days an e-commerce site needs 10 servers. On sale day traffic jumps 10×. With AWS it scales to 100 servers for those hours, then back to 10, and you pay only for the extra hours. Play with it:
- Region = geographic area; AZ = one or more isolated data centres in a Region; edge location = CloudFront cache near users.
- High availability = deploy across multiple AZs. Disaster recovery = across multiple Regions.
- Choose a Region by: compliance/data residency, latency to users, service availability, price.
05Summary
- AWS = Amazon Web Services: 200+ on-demand services, pay-as-you-go.
- Region (e.g. Mumbai) contains 3+ Availability Zones; each AZ is one or more isolated data centres.
- Edge locations cache content near users (CloudFront).
- Choose a Region by compliance, latency, service availability and cost.
- “11 nines” is S3 durability, not overall AWS availability.
06Quick quiz
1. An Availability Zone is…
2. ap-south-1 is which Region?
3. To survive the failure of one data centre, you should…
4. 99.999999999% (“11 nines”) describes…
5. First thing to check for a bank’s data location?