Cloud and service models¶
This morning you probably played a song on Spotify, opened Classroom and saved something to Google Drive. None of it ran on your phone, not really. Where was the song? Where is your homework right now? And who keeps the machine that stores it switched on at three in the morning?
The answer is "in the cloud", and it is less mysterious than it sounds. A hotel in Calvià that receives ten times more bookings in August than in January cannot buy ten servers for August and throw nine away in September. It rents. As a DAM developer, almost everything you build will end up running on somebody else's computer, so you need to know exactly what you are renting.
What you will learn
- Define cloud computing with its five characteristics, and tell it apart from on-premise
- Tell the three service models apart: IaaS, PaaS and SaaS
- Say who manages what in each model, and which part is always yours
- Classify real services you use, with proof from the provider's own website
- Publish your own web page with GitHub Pages, and say which model it is
- Avoid the trap: "the cloud" is not in the sky, it is a building full of somebody else's servers
The cloud is renting from the utility¶
Cloud computing is on-demand access, over the network, to computing resources (servers, storage, databases, applications) that a provider shares among many customers, that can grow or shrink quickly, and that you pay for according to use.
In the city-builder game of this module, the cloud is renting water and electricity from a utility instead of building your own plant. You open the tap and pay the bill. The alternative is called on-premise: the company buys its own servers and keeps them in its own building.
The US standards institute NIST wrote the definition everyone quotes (SP 800-145, 2011). It lists five characteristics. A service that lacks them is not really cloud:
| Characteristic | What it means | Example |
|---|---|---|
| On-demand self-service | You sign up and create resources yourself, without waiting for anyone to install anything. | Creating a server from a web form in a couple of minutes |
| Network access | You reach it from any place and device with a connection. | Opening Classroom at school, at home or on your phone |
| Shared resources | The same hardware serves many customers, kept separate logically. | Thousands of companies in the same data centres |
| Elasticity | Resources can be added or removed quickly as demand changes. | The hotel's website in August and in January |
| Pay per use | Use is measured, and you pay for what you consume. | Like the electricity bill |
Back to the hotel. On-premise, it buys a server, puts it in an air-conditioned room and pays somebody to look after it: a big investment on day one, and if it falls short in August, bad luck. In the cloud there is no initial investment, it pays monthly according to use, and in August it asks for more capacity.
Careful: \"the cloud is somewhere up there\"
The cloud is physical. It is a data centre: a building full of servers, cables, power supplies and cooling, owned by a provider such as Amazon, Microsoft, Google or Hetzner. An external hard disk or a server in the office cupboard is not cloud: nobody else shares it, it does not grow on demand and nobody bills you by use.
Three service models: IaaS, PaaS and SaaS¶
You can rent at three different heights. The higher you go, the more the provider does for you.
A service model says how much of the system the provider gives you ready-made, and how much is left for you to manage.
| IaaS · infrastructure | PaaS · platform | SaaS · software | |
|---|---|---|---|
| What you get | Virtual machines, network and storage | A ready environment where you deploy your code | A finished application, ready to use |
| What you manage | Operating system, software, application and data | The application and the data | Only your data and settings |
| Who it is for | System administrators | Developers | End users |
| Examples | Amazon EC2, Azure Virtual Machines, Google Compute Engine, Hetzner Cloud | Google App Engine, Heroku, Azure App Service, GitHub Pages | Gmail, Google Classroom, Microsoft 365, Spotify, Kahoot |
- IaaS (Infrastructure as a Service): you rent a virtual machine, a computer simulated by software on the provider's hardware. It is like renting an empty plot with water and power connected: you build the house.
- PaaS (Platform as a Service): you bring your code and the platform runs it. Like renting a furnished flat: you bring your things.
- SaaS (Software as a Service): you just use the application. Like staying in a hotel.
Who manages what¶
Stack the layers of any system from the building up to your data. For each model, mark what is the customer's job (C) and what is the provider's (P):
| Layer | On-premise | IaaS | PaaS | SaaS |
|---|---|---|---|---|
| Data | C | C | C | C |
| Application | C | C | C | P |
| Runtime (Java, Node, database...) | C | C | P | P |
| Operating system | C | C | P | P |
| Virtualisation | C | P | P | P |
| Servers, disks and network | C | P | P | P |
| Building, power and cooling | C | P | P | P |
Look at the first row. Your data are always your responsibility, whatever the model. The higher you climb, the less administration you do, but the less control you have. Providers call this the shared responsibility model.
flowchart LR
ON["<b>On-premise</b><br>(build your own<br>power station)"] --> I["<b>IaaS</b><br>(rent the plot<br>with power)"]
I --> P["<b>PaaS</b><br>(rent a<br>furnished flat)"]
P --> S["<b>SaaS</b><br>(stay in<br>a hotel)"]
Careful: the borders are not walls
Some services sit between two models. Firebase, for example, is a platform you build on (PaaS), but it also offers finished pieces such as user login. What is assessed is the reason: say what the customer manages and what the provider manages, and the label follows.
GitHub Pages: a platform you can use today¶
GitHub Pages is a free GitHub service that publishes a static website (HTML, CSS and images that do not change per visitor) straight from a code repository.
You put the files in a repository; GitHub stores them, serves them to the world and keeps the servers running. You never see a server. That is the definition of PaaS, and you will use it all unit: in Deliverable 3 your group publishes its dossier as a GitHub Pages site.
Three facts to know before you start (checked on 25/10/2026 in GitHub's documentation):
| Fact | What it means for you |
|---|---|
| GitHub's terms require you to be 13 or older | Everyone in this course can open an account; it is free and needs no card |
| With a free account, the repository must be public | Everything you upload can be seen by anybody: no personal data, no phone numbers; your first name is enough |
| Pages is not for shops or for collecting passwords | A dossier, a portfolio or a project page is fine |
In class: what level does it play at?¶
Part 1 · Classify real services. In the Classroom template:
| Step | Do this | You should get |
|---|---|---|
| 1 | Write down four services you used this week (phone or computer) and add these four: Amazon EC2, Heroku, Google App Engine, Hetzner Cloud | A list of eight services |
| 2 | Classify each one as IaaS, PaaS or SaaS and write one sentence on why: what does the customer manage, and what does the provider? | Eight labels with a reason |
| 3 | For each, paste a link to the official page (product or pricing page, not a blog or an AI) and quote the sentence that proves it | Eight pieces of evidence |
| 4 | For your project company, list three systems it has or probably has (website, bookings, ERP, email, plant control...) and, for each, the model you would propose and a real provider that offers it | The draft of Section A of D3 |
Part 2 · Your first site on GitHub Pages.
| Step | Do this | You should see |
|---|---|---|
| 1 | Sign in to GitHub, or create a free account at github.com/signup (email, password, username, a check and a code sent to your email) | Your GitHub home page |
| 2 | Click + (top right) → New repository. Name it exactly yourusername.github.io. Visibility Public. Turn on Add README. Create repository |
Your new repository with a README.md |
| 3 | Add file → Create new file. Name it index.html. Write a minimal page: a heading with your first name and the sentence "This website is a ... service because ...". Commit changes |
index.html in the repository |
| 4 | Settings → Pages. Under Build and deployment, Source: Deploy from a branch; Branch: main and / (root). Save |
A message that your site is being published |
| 5 | Wait (it can take up to 10 minutes) and open https://yourusername.github.io |
Your page, visible from any computer in the world |
If you cannot or do not want to open a GitHub account, make the page with Google Sites and your school account, and say in the sentence that Google Sites is SaaS, not PaaS. It does not affect your individual mark: in the deliverable, the website is group work.
Hand in on Classroom: the table of part 1 with its evidence, your three company systems and the link to your site. No generative AI: this activity is assessed (class work, 30%).
For your project¶
This lesson starts Deliverable 3 · Cloud architecture, due Monday 8 February 2027 at 23:59, with individual defences in session 14 that same day.
| Today's work | Becomes | Criterion |
|---|---|---|
| Three company systems with a model and a real provider | Section A: service models | a |
| Your GitHub account and first site | The group's website, published by the Technician | Group part |
Key takeaways¶
- Cloud computing is renting computing resources on demand, over the network, shared, elastic and paid by use. It runs in real data centres. The opposite is on-premise.
- IaaS gives you machines, PaaS gives you a place to deploy your code, SaaS gives you the finished application.
- The higher the model, the less you administer and the less you control. Your data are always your responsibility (the shared responsibility model).
- End users mostly meet SaaS; developers live mostly in PaaS and IaaS.
- GitHub Pages publishes a static website from a repository: a PaaS you can use for free, with a public repository.
Quick check¶
1. Which of these is cloud computing?
- An external hard disk where the company copies its files every Friday
- A server bought by the company and kept in its own office
- Storage rented from a provider, reached over the internet and billed by the GB used
- A USB stick shared by the whole office
The cloud is on demand, over the network, shared, elastic and paid by use. The disk, the office server and the stick have none of those features; the office server is on-premise.
2. A company rents a virtual machine and installs the operating system, the database and its application on it. Which model is it?
- IaaS
- PaaS
- SaaS
- On-premise
The provider gives only the infrastructure (the virtual machine, network and storage). Everything from the operating system up is the customer's job: that is IaaS.
3. In which model does the customer manage only their data and settings?
- IaaS
- PaaS
- SaaS
- In none: the customer always manages the operating system
In SaaS the application is finished and the provider runs everything below it. The customer still owns the data, which is their responsibility in every model.
4. You upload the files of your website to a repository and GitHub Pages publishes it without you managing any server. Which model is it?
- IaaS
- PaaS
- SaaS
- It is not cloud, because it is free
You bring the code; the platform stores and serves it. That is PaaS. Being free does not stop it being cloud: the provider still runs the data centre.
5. Which statement about the shared responsibility model is true?
- In SaaS, the provider is responsible for the customer's data
- In IaaS, the provider updates the operating system of your virtual machines
- Whatever the model, the customer is responsible for its data
- On-premise, the provider manages the building and the power
The data row is always the customer's. In IaaS the operating system is also yours; on-premise there is no provider at all.
6. A classmate says: \"The cloud is just the internet; my files are in the air.\" What is the best answer?
- True: the cloud has no physical location
- The files are in real data centres owned by a provider; the internet is only the road to them
- True, but only for SaaS services
- False: cloud files are always stored on your own device
The cloud is physical: buildings full of servers, cables and cooling. The network is how you reach them. That is why, in lesson 4, it will matter in which country they are.
7. A bakery in Inca wants online ordering without hiring a programmer, and a start-up in Palma wants to deploy its own app without managing servers. Which models fit each one best?
- IaaS for the bakery, SaaS for the start-up
- SaaS for the bakery, PaaS for the start-up
- PaaS for both
- On-premise for both, to save money
The bakery needs a finished application (SaaS). The start-up has its own code and wants somebody else to run the servers: that is PaaS.
You now know what you can rent. Next you will see who owns the power station, what the cloud is used for and how much the bill is: Deployment and functions.