It's been a while since I've written anything. Since graduating, I've fallen hard into the world of cloud computing — and rather than keep all the notes to myself, I want to start sharing them. This is the first episode of a series that takes you from "what even is the cloud?" to passing the AWS Cloud Practitioner exam, one bite-sized chapter at a time.
In this episode: what cloud computing actually is, why we moved off of "everyone owning their own servers," and a friendly introduction to the first six AWS services you'll meet.
What does the series cover?
By the end of these episodes you'll be comfortable with:
- What cloud computing is and why it matters
- The difference between On-Premise and Cloud
- AWS's core concepts and services, told as if they were characters
- The promises the cloud makes (and the small print)
- Enough vocabulary to pass the Cloud Practitioner exam
The whole thing is opinionated towards storytelling. The goal isn't to memorise definitions; it's to actually understand why each piece exists.
What is "the cloud," really?
The cloud is the ability to use computing power, storage, databases, and networking on demand, over the internet, without owning any of the physical hardware.
That sentence sounds dry, so let me rewind a couple of decades.
Life before the cloud (the 90s and early 2000s)
- Programs were installed from CDs.
- Files lived on your computer's hard drive.
- Companies bought their own servers and stuffed them in a room — sometimes literally a broom closet.
This setup had real problems:
- Hardware failures lost data. A drive died, and there went last quarter's invoices.
- Storage ran out. Filling the disk meant buying a new one and physically installing it.
- Traffic spikes broke everything. A successful marketing campaign would crash the website that benefited from it.
- Maintenance was expensive. Someone had to apply patches, replace failing parts, monitor temperature, and stay on call.
The result: high upfront cost, slow change, lots of operational stress.
Life with the cloud
Today, you open a console, pick what you need, and click. Servers live in someone else's data centre. You pay only for the time you used them, and when you don't need them, you turn them off.
The big providers — AWS (Amazon), Google Cloud, Azure (Microsoft) — built massive data centres around the world and turned compute into a utility.
Why the cloud makes sense
A short list of what changes when you move:
- Resources are elastic. Add capacity in seconds, remove it when you don't need it.
- You pay only for what you use. No "we bought a server for this big launch and now it sits at 5% utilisation."
- Scale handles itself (with the right configuration). Traffic spike at 3 AM? The cloud scales while you sleep.
- High availability is built in. Multiple data centres, automatic failover, backups.
- Access from anywhere. No VPN gymnastics to reach your own infrastructure.
What are the alternatives?
The cloud isn't the only option, and the Cloud Practitioner exam expects you to be able to compare:
- On-Premise — you own every server. Maximum control, maximum cost, maximum maintenance burden.
- Colocation — you own the servers, but they live in someone else's data centre. You rent space, power, and cooling; you still touch the hardware.
- Hybrid — sensitive workloads stay on-premise (or in a private cloud), elastic workloads run in the public cloud.
The cloud's promises
Five characteristics show up over and over again:
- On-demand self-service — provision resources without a human in the loop.
- Broad network access — reachable over standard internet.
- Multi-tenancy & resource pooling — many customers share the same underlying hardware safely.
- Rapid elasticity — scale up and down within minutes.
- Measured service — pay-per-use, with usage you can actually see.
Meet the cast: six AWS services
I find AWS services easier to remember if I picture them as characters in a small town. Here are the first six.
Amazon EC2 — the muscle
Elastic Compute Cloud. This is a virtual server. Your application runs here. EC2 comes in a hundred flavours (compute-optimised, memory-optimised, GPU-equipped, you name it), at sizes ranging from a tiny t2.nano to a u-24tb1.metal with 24 TB of RAM. You pay by the second or the hour.
Amazon S3 — the warehouse keeper
Simple Storage Service. Object storage with eleven 9s of durability (99.999999999%). You upload a file, you get a URL. It's how nearly every static asset on the modern web is served. Backups, photos, video, build artefacts — all happy living in S3.
AWS Lambda — the magic lamp
Serverless, event-driven compute. You write a function, AWS runs it when something triggers it (an HTTP request, a file upload, a queue message). When nothing's happening, Lambda costs you nothing — no idle server fee.
AWS IAM — the bouncer
Identity and Access Management. Decides who can do what. Users, roles, groups, policies all live here. IAM is the system that means a developer can deploy code but cannot read the customer database, even though they technically work for the same company.
Amazon RDS — the librarian
Relational Database Service. Managed databases. You pick the engine (Postgres, MySQL, MariaDB, SQL Server, Oracle, Aurora), and AWS handles backups, patches, replication, and failover. You focus on the data.
Amazon CloudFront — the global courier
A Content Delivery Network. Your static assets get cached at edge locations around the world. A user in Sydney requests an image stored in us-east-1, and CloudFront serves it from a server in Sydney instead of bouncing the request across the Pacific.
Cloud Practitioner exam: the questions you'll see
"Should this be On-Prem or Cloud?"
Typical question shape:
A company wants to avoid high upfront costs and scale with demand. Which solution is most appropriate?
Answer: cloud adoption. Reasoning: no hardware purchase, scales with demand, shifts to an operating-cost model.
CapEx vs. OpEx
- CapEx (Capital Expenditure) — large upfront purchase of hardware. Sits on the balance sheet, depreciates over years.
- OpEx (Operating Expenditure) — pay-as-you-go, expense-as-incurred.
This distinction comes up a lot on the exam, because it's the financial framing for why companies move to the cloud in the first place.
Deployment models
- Public Cloud — open to everyone, accessed over the internet. AWS, Azure, GCP.
- Private Cloud — dedicated to one organisation, often in their own data centre.
- Hybrid Cloud — both, connected and treated as one environment.
The Shared Responsibility Model
The single most exam-quoted concept:
AWS is responsible for security of the cloud. The customer is responsible for security in the cloud.
Translated:
- AWS handles: physical data centres, hardware, the hypervisor, the network fabric, the underlying software that runs the services.
- You handle: IAM configuration, data encryption, what's in your S3 bucket and who can read it, your application code, patching the OS on your EC2 instances.
Exam-day strategy
The Cloud Practitioner exam isn't a deeply technical exam — it's a conceptual one. The questions favour:
- Cost advantages (CapEx vs. OpEx)
- Elasticity & scalability
- High availability
- Global reach
- The shared responsibility model
- Knowing which service does what at a one-sentence level
If you can comfortably explain each of those to a non-technical friend, you're in good shape.
Up next
The next episode will build on this foundation: AWS Global Infrastructure (regions, availability zones, edge locations), more services from the AWS town (VPC, SNS, SQS), and the pricing models that come up on the exam.
Until then — welcome to the sky. The view's pretty good up here.