Skip to content
Tutorials Center
My learning

Topics

Lesson 1 of 57 ยท Cloud and Azure Basics

What Is Cloud Computing?

Learn what cloud computing means: IaaS, PaaS, SaaS and serverless, the shared responsibility model, pay-as-you-go pricing, public and hybrid clouds, CapEx vs OpEx.

Cloud computing means renting computing resources (servers, storage, databases, networks, even finished applications) from a provider over the internet, and paying for what you use instead of buying and running the hardware yourself. Instead of ordering a server, waiting weeks for delivery, racking it and keeping it cool, you ask for one with a few clicks or a single command, and it exists a minute later.

Microsoft Azure is one of the three largest public clouds. Before we look at Azure itself, this lesson gives you the vocabulary every cloud shares: the service models, who is responsible for what, how billing works and where the cloud can live. Nothing here needs an account. The runnable example is a small JavaScript stand-in that models a pay-per-use bill, so you can see why the cost model feels so different from buying hardware.

The core idea: on-demand, elastic, metered

Five properties make something "cloud" rather than just "a server somewhere":

  • On-demand self-service. You provision resources yourself, without filing a ticket with a person.
  • Broad network access. Everything is reachable over the network, from a browser, a command line or code.
  • Resource pooling. Many customers share the provider's huge pool of hardware, kept apart by virtualization and strict isolation.
  • Rapid elasticity. You can grow or shrink in minutes. A shop can run ten servers during a sale and one afterwards.
  • Measured service. Usage is metered, so you pay for hours, gigabytes or requests actually consumed.

The result is a different way to think about money. Buying servers is capital expenditure (CapEx): a large payment up front for assets that you must size for your peak demand and keep for years. Renting in the cloud is operational expenditure (OpEx): a smaller recurring cost that follows your usage. CapEx wastes money when the hardware sits idle; OpEx can cost more per unit when you run something flat-out all year, which is why reserved discounts exist (covered in Pricing, Cost Management and Budgets).

IaaS, PaaS and SaaS

Cloud services come in layers. The higher you go, the more the provider manages for you and the less you control.

flowchart TB
    subgraph IaaS["IaaS: you manage the OS and up"]
      A1["Virtual machine<br/>you patch the OS"]
    end
    subgraph PaaS["PaaS: you manage code and data"]
      A2["Web app platform<br/>you deploy code"]
    end
    subgraph SaaS["SaaS: you manage your users and settings"]
      A3["Finished app<br/>email, office suite"]
    end
    IaaS --> PaaS --> SaaS

Moving from IaaS to SaaS hands more work to the provider and gives you less control.

  • Infrastructure as a Service (IaaS) gives you virtual machines, disks and networks. You choose the operating system and you are responsible for patching it, installing software and securing it. In Azure this is Virtual Machines.
  • Platform as a Service (PaaS) gives you a managed runtime. You hand over your code, and the provider runs the servers, the operating system and the web server. App Service and Azure SQL Database are PaaS.
  • Software as a Service (SaaS) is a finished application you simply use, such as Microsoft 365. You manage your users and settings, nothing else.
  • Serverless is a style of PaaS where you do not think about servers at all. You supply a function or a container, the platform runs it when something happens (a web request, a timer, a message) and you are billed per execution. Azure Functions is the standard example. Servers still exist; you just never see them.

A rough rule: choose the highest layer that still does what you need, because every layer you do not manage is a layer you do not have to patch, monitor or staff.

The shared responsibility model

Moving to the cloud does not move all the work to the provider. Security and operations are shared. The provider is always responsible for the physical datacenter, the hardware and the virtualization layer. You are always responsible for your data, your accounts and who can reach what. The layers in between depend on the service model:

Responsibility On-premises IaaS PaaS SaaS
Data and information You You You You
Accounts and identities You You You You
Applications You You You Provider
Operating system You You Provider Provider
Network controls You Shared Shared Provider
Physical hosts and datacenter You Provider Provider Provider

Notice that the first two rows never move: you own your data and your identities in every model. That is why most real-world cloud breaches come from customers' misconfigurations (a storage container left public, a leaked password) and not from the provider's datacenter. Security Best Practices and Defender for Cloud returns to this.

Public, private and hybrid clouds

Where the cloud runs is a separate question from what kind of service it is:

  • A public cloud (Azure, AWS, Google Cloud) is owned by the provider and shared by many customers, each isolated from the others.
  • A private cloud is cloud-style infrastructure used by a single organization, in its own datacenter or hosted for it exclusively.
  • A hybrid cloud combines the two (or on-premises servers and a public cloud) and connects them, for example an old database that must stay in a company office feeding web apps in Azure. Private Endpoints, VPN and ExpressRoute shows how the connection works.
  • Multicloud means using more than one public cloud provider.

Reasons companies keep some workloads on-premises include data-residency laws, very low latency to local equipment and large existing investments. Reasons they move to the public cloud include speed, global reach and not running datacenters at all.

Why the bill behaves differently

In a datacenter, your cost is fixed: the servers cost the same whether the site is busy or idle. In the cloud, cost is a function of usage. Three typical meters are compute time (per second or hour a machine runs), storage (per gigabyte per month) and data leaving the provider's network (per gigabyte, often called egress). Every service has its own meters, and the sum of them is the bill.

The first example models a small web shop with three meters. The numbers are invented round figures chosen to keep the arithmetic readable, not real Azure prices.

JS A pay-per-use bill

Press Run (or Ctrl + Enter) to see the console output here.

No output.

The sale day costs far more than the quiet day, but the quiet day costs almost nothing. Owning hardware sized for the peak would cost the "always at peak" figure every day, used or not. That elasticity is the main economic argument for the cloud.

What is real and what is the stand-in

The service models, the shared responsibility table and the CapEx/OpEx idea are real concepts you will meet in every Azure document. The bill calculator above is a stand-in: real Azure prices vary by region, size and commitment, and are published in the pricing calculator.

The second stand-in classifies a few workloads into the service model that fits them best, using a simple set of questions.

JS Which service model fits?

Press Run (or Ctrl + Enter) to see the console output here.

No output.

Check your understanding

Quick check ยท 4 questions

of 4 answered You scored / 4 ยท perfect!

  1. 1. Who is responsible for patching the operating system of an IaaS virtual machine?

    With IaaS the provider runs the hardware and virtualization; the operating system and everything above it belong to you.
  2. 2. Which item stays the customer's responsibility in IaaS, PaaS and SaaS alike?

    Data and identities are yours in every model. The provider always handles the physical datacenter and hosts.
  3. 3. A company pays a small amount every month that follows its usage instead of buying servers. This is mainly:

    Pay-as-you-go cloud billing is operational expenditure. Buying hardware up front is capital expenditure.
  4. 4. What best describes a hybrid cloud?

    Hybrid means combining public cloud resources with private or on-premises ones and connecting them.

Summary

  • Cloud computing rents on-demand, elastic, metered resources over the network.
  • IaaS gives you machines, PaaS gives you a managed platform, SaaS gives you a finished app; serverless runs your code per event with no server management.
  • Responsibility is shared: you always own your data and identities.
  • Public, private, hybrid and multicloud describe where and how clouds are combined.
  • Pay-as-you-go turns fixed CapEx into usage-driven OpEx, which rewards elasticity and punishes forgotten resources.

Next: Azure at a Glance.