So far, we have talked about the cloud intuitively. But is there an official definition? Yes. The NIST (National Institute of Standards and Technology, a leading U.S. agency) published a definition that has become the global standard. According to NIST, something is truly “cloud” only if it meets five essential characteristics.

Knowing them will help you distinguish real cloud from simple “cloud marketing.”

You can get resources (a server, storage…) yourself, whenever you want, without talking to anyone. You don’t have to call a sales rep or wait for approval.

Analogy: a vending machine. You insert the coin, press the button, and get your drink instantly. You don’t need an employee to hand it to you.

Example: At 3 a.m. you need a server for a test. You log into the AWS console, click, and in 2 minutes you have it, without any AWS human intervention.

Resources are available over the network (usually the internet) and can be accessed from any device: laptop, mobile, tablet…

Analogy: your bank account. You access your money from the ATM, the mobile app, or the website. The money is the same, the access is universal.

Example: You manage your AWS servers from your laptop’s browser at the office and from the app on your phone while on the train.

The provider has a large shared “warehouse” of resources (thousands of servers) and dynamically distributes them among all clients. You don’t know exactly which physical machine you’re on, and it doesn’t matter.

Analogy: the power grid. The electricity in your house comes from a huge network shared by the whole city. You don’t have your own personal power plant; you use the part you need from a common resource.

This is what enables economies of scale: by sharing a gigantic infrastructure among millions of clients, the cost per client drops dramatically. Even though resources are shared, your data is isolated and protected from others.

You can increase or decrease capacity quickly, almost instantly, and often automatically. When there’s a spike in demand, you scale up; when it passes, you scale down.

Analogy: an accordion, or a rubber band. It stretches when you need it and returns to its original size when you don’t.

Real example: An e-commerce website on Black Friday automatically multiplies its servers at dawn when the flood of shoppers arrives, and reduces them at night when traffic drops. No one has to do it manually.

For many, this is the star feature of the cloud and what most distinguishes it from the traditional model.

Everything you use is measured (server hours, GB stored, requests…) and you pay exactly for it. It’s transparent: you can see your usage at any time.

Analogy: the electricity or water meter. You pay for the kilowatts or liters you actually consume, no more, no less.

Example: If your server was on for 30 hours this month, you pay for 30 hours. If you turn it off, you stop paying. AWS shows you a dashboard with the exact breakdown.

The pillars work together to create the “cloud” experience:

If a service is missing any of these pillars, it’s probably not real cloud, even if it’s called that. For example, a provider that takes two weeks to give you a server and charges a fixed fee doesn’t meet either on-demand self-service or measured service.

The five NIST pillars of cloud are:

1. On-demand self-service — you get it yourself, instantly.
2. Broad network access — from any device.
3. Resource pooling — shared infrastructure (economies of scale).
4. Rapid elasticity — grows and shrinks with demand.
5. Measured service — you pay exactly for what you use.

This is the official definition that distinguishes real cloud from marketing. In the next subchapter, we’ll land these concepts in the concrete advantages you get from using the cloud.