My Learning Journey: AWS Global Infrastructure Overview

The core idea is a hierarchy: Regions, Availability Zones, and edge locations, each solving a different problem.

A Region is a separate geographic area that contains multiple Availability Zones. What surprised me was realizing that choosing a Region isn't arbitrary — running your workload from a Region closer to your end users directly decreases latency. So Region choice is really a user-experience decision as much as a technical one.

Within each Region sit Availability Zones (AZs) — one or more discrete data centers, built specifically for fault isolation. I used to think of an AZ as basically "a data center," but that's not quite right: an AZ can span multiple data centers, though a single data center only ever belongs to one AZ. AZs within a Region are connected through high-speed, low-latency private links, which is what AWS recommends you lean on — provisioning your compute across multiple AZs rather than just one, so a failure in one doesn't take your whole application down.

That led into two terms I had been using loosely without really distinguishing: fault tolerant and elastic/scalable . Fault tolerant means the infrastructure has built-in redundancy — it can absorb a failure. Elastic and scalable means resources adjust dynamically as demand goes up or down. They're related, but they're solving different problems: one is about resilience, the other about flexibility.

Then there are edge locations — and this is where I had to correct an assumption. I initially thought edge locations were just mini-Regions, located in the same general areas. They're not. Edge locations exist in far more cities globally than Regions do, specifically so that services like CloudFront can cache and deliver content close to wherever the end user actually is, independent of which Region your infrastructure lives in.

Zooming out, Module 3 reinforced something practical: good architecture isn't just about getting an application running — it's about placing it intentionally. Which Region reduces latency for your users. How many AZs you're spreading across for resilience. Whether edge locations can take load off your origin entirely. That's infrastructure thinking, not just deployment thinking — and it's a mindset shift I'm glad this module pushed me toward.