A friend asks you to hold a birthday party for forty people next Saturday. You have two choices. You could buy a hall, chairs, a sound system and a kitchen, use them for one evening, and then own all of it for the rest of your life. Or you could book a venue for the evening, use everything that is already set up, and walk away when the party ends. Almost nobody buys a hall for one party. Yet for decades, that is roughly how companies ran their computers. They bought the whole building.
The cloud is the ‘book the venue for the evening’ option for computing. You rent computers that someone else has already bought, powered, cooled and connected. When you are done, you stop paying. That is the whole idea. Everything else in this series is detail hung on that one frame.
So what is the cloud, really?
Strip away the marketing and the cloud is a simple arrangement. A company called a cloud provider builds enormous buildings full of computers, called data centres. They run the power, the air conditioning, the network cables and the security guards. You, the customer, ask for a slice of that, a server here, some storage there, over the internet. You get it in seconds, you pay for the time and space you use, and you release it when you are finished.
The closest everyday parallel is renting an apartment instead of buying a house. When you buy a house you pay a huge amount up front, you fix the leaking roof yourself, and the house is yours whether you live in it or not. When you rent, you move in the same day, the landlord fixes the roof, and you can leave at the end of the month. Cloud computing is renting the apartment. Running your own servers in your own building is buying the house. Neither is wrong, but they feel completely different to live with.
There is also a respected, neutral definition that you will see quoted in interviews and study guides. A United States standards body, the National Institute of Standards and Technology, defines cloud computing as on-demand network access to a shared pool of computing resources that can be quickly given out and taken back with little effort. Read that slowly and you will spot the same renting idea in formal clothes. On-demand means you ask and you get it. Shared pool means the provider keeps a large stock and hands you a piece. Quickly given out and taken back means no waiting weeks for hardware to arrive.
The word ‘cloud’ itself is almost an accident. For years, when engineers drew diagrams of their own systems, anything out on the internet that they did not control was sketched as a fuzzy cloud shape. The name stuck. So when someone says ‘it runs in the cloud’, they mean ‘it runs on computers out there that we rent, reached over the internet’. There is no actual cloud and nothing floats in the sky. Your files sit on a hard drive in a building you could, in principle, drive to.
Why renting computers caught on
Renting only wins if it solves a real problem, and it does. The same standards body lists five traits that make something genuinely cloud rather than just a server in a rented room. You do not need to memorise them word for word, but each one maps to a frustration that owning your own hardware used to cause.
The five traits, in plain words
First, self-service: you click or run a command and the server appears, no purchase order, no waiting for a person to approve it. Second, broad access: you reach it from a laptop, a phone or another server, anywhere with a connection. Third, a shared pool: the provider owns a huge stock of machines and slices pieces to many customers, which is why it is cheaper than each company buying its own. Fourth, elasticity, which is the ability to grow and shrink fast. If your website gets busy on sale day, you add servers in minutes and remove them at night. Fifth, measured use: everything is metered like electricity, so you can see exactly what you consumed and pay for that and nothing more.
Here is the contrast that made companies switch, laid out side by side.
| Question | Own it yourself | Rent the cloud |
|---|---|---|
| How fast to start? | Weeks: order, ship, install | Minutes: click and go |
| What do you pay up front? | A large lump sum | Little or nothing |
| Busy day, need more? | Buy more hardware, hope it arrives | Add capacity in minutes |
| Quiet night, need less? | Idle hardware, still paid for | Shrink and stop paying |
| Who fixes the building? | You and your team | The provider |
What actually happens when you use the cloud
Say you open a photo app on your phone and your pictures load. Behind that simple moment, a short journey happens. Your phone sends a request over the internet. It reaches one of the provider’s data centres, grouped into a region, which is a cluster of buildings in one part of the world. Inside, a rented server handles your request, pulls your photos from rented storage, and sends them back. The round trip takes a fraction of a second, and you never see any of it.
The reason this matters is that the same pattern sits under almost everything you already use. Streaming a show, sending an email, checking a bank balance, saving a document so it appears on your other laptop: each one is your device talking to rented computers in a data centre somewhere. You have been a cloud user for years. The only new part is that, in a cloud job, you will be on the other side, setting up those rented servers and storage instead of just using them.
Why this matters in your first job
In your first weeks, a senior engineer might say ‘spin up an instance in the Mumbai region and put the files in a bucket’. That sentence is exactly the journey above: an instance is a rented server, a bucket is rented storage, and the region is which set of buildings to use. In an interview, ‘what is the cloud’ is a real opening question, and a calm answer about renting computers you reach over the internet, with the five traits, beats a memorised paragraph every time.
Who runs these clouds
A handful of large companies, often called hyperscalers, run most of the public cloud. The three you will meet most are Amazon Web Services, Microsoft Azure and Google Cloud. Together they hold well over half of the market, with Amazon Web Services in front, Microsoft Azure second and Google Cloud third by most recent counts in early 2026. The exact percentages move every quarter, so treat them as ‘these three are the big ones’ rather than fixed numbers.
For a fresher, the good news is that they rhyme. All three rent you servers, storage and networking. They use different names for the same ideas, which makes the field look more crowded than it is. Learn the concept once and you can find its twin on any provider. Here are a few names lined up so the pattern is visible.
| The idea | Amazon Web Services | Microsoft Azure | Google Cloud |
|---|---|---|---|
| A rented server | EC2 instance | Virtual Machine | Compute Engine |
| Bulk file storage | S3 | Blob Storage | Cloud Storage |
| Your private network | VPC | Virtual Network | VPC |
Do not feel pressure to pick a favourite yet. Many teams use more than one provider, and the skills transfer. This series stays even across all three so you build understanding, not loyalty. Part 9 looks properly at how to think about the choice. For now, ‘three big providers, same core ideas, different labels’ is enough.
Who is responsible for what
Renting an apartment splits the chores. The landlord keeps the building standing, the water running and the front gate locked. You lock your own door, decide who you let in, and look after what is inside. Cloud works the same way, and the idea has a name you will hear constantly: the shared responsibility model. The provider secures the cloud itself, the buildings, the hardware, the core network. You secure what you put in it, your data, who has access, and how you set things up.
A small bill, broken down
Numbers make the renting idea concrete. Imagine you put up a small personal website. You rent one modest server and a little storage, and you switch the server off when you are asleep because nobody is visiting at 3 a.m. Cloud billing is by the hour for the server and by the gigabyte for storage, so a part-time setup costs part-time money.
Worked example
Say a small server rents for about 2 cents an hour and you run it 10 hours a day, the busy stretch when people actually visit. That is 10 hours times 30 days, so 300 hours a month. 300 hours times 2 cents is 600 cents, which is 6 dollars a month for the server.
Add 10 gigabytes of storage at roughly 2 cents per gigabyte per month. That is another 20 cents. Your bill lands near 6 dollars and 20 cents for the month.
Now the lesson: if you had left the server on all 24 hours instead of 10, you would pay for 720 hours, about 14 dollars and 40 cents, more than double, for a website nobody was visiting overnight. Same server, very different bill. That single habit, turning off what you are not using, is the heart of cloud cost control, and we go deep on it in Part 13.
Those exact rates are rounded for the maths and shift over time and by region, so do not quote them as gospel. The shape of the lesson is what lasts: you pay for time and space used, and waste comes from leaving things running. There is also a way to practise this for free. Each big provider offers a free tier, a set of small allowances meant for learning. For example, Amazon Web Services gives new accounts a pool of starter credits and some always-free service allowances. The details change often, so check the current page before you rely on a specific figure. Part 17 is built entirely around learning on these free tiers without a surprise bill.
Where to go from here
If you remember one thing, remember this: the cloud is renting computers you reach over the internet, paid by the hour and the gigabyte, with the work of running the building handed to the provider. Every grand-sounding term ahead, instances, buckets, regions, identity, containers, serverless, is a detail bolted onto that frame. You already understand the frame.
A useful first habit costs nothing. For one day, notice how often you touch the cloud without thinking: every photo that syncs, every message that arrives on two devices, every show that streams. Each is your device renting a moment of someone else’s computer. Once you see it, the field stops feeling abstract.
The honest reasons companies actually made this switch, including the ones the brochures skip, are where we go next. Keep the renting picture in your head and Part 2 will click into place.
Try this now: Write down, in one sentence and your own words, the difference between buying a house and renting an apartment. You just explained the cloud. Keep that sentence; you will reuse it in your first interview.
References
- NIST Special Publication 800-145, The NIST Definition of Cloud Computing
- AWS Free Tier (official overview of free allowances)
- Google Cloud global locations: regions and zones
- Microsoft Azure global infrastructure: geographies and regions
About The Author
DrJha