This chapter explains the three core cloud service models — Infrastructure as a Service, Platform as a Service, and Software as a Service — and when each is used. AZ-900 tests this under objective 1.2, one of the most frequently tested topics on the exam.
Jump to a section
A simple way to picture IaaS, PaaS, and SaaS Explained
Cooking dinner from scratch means buying raw ingredients and doing every step yourself — full control, but full effort. A meal kit gives you pre-measured ingredients and a recipe, so you skip the shopping and planning but still do the cooking. Ordering takeout means someone else handles everything, and you just eat. These three approaches map closely to IaaS, PaaS, and SaaS: IaaS gives you the raw building blocks (like buying ingredients) and you assemble everything yourself; PaaS gives you a pre-built platform (like a meal kit) so you focus on your application instead of infrastructure; SaaS gives you a finished, ready-to-use application (like takeout) where you don't build or maintain anything at all.
Infrastructure as a Service (IaaS)
IaaS provides the fundamental building blocks of computing — virtual machines, storage, and networking — over the internet. The customer manages the operating system, applications, and data, while Azure manages the physical infrastructure underneath. This gives the most control and flexibility, but also the most responsibility.
Platform as a Service (PaaS)
PaaS provides a ready-to-use platform for building and running applications, without the customer needing to manage the underlying operating system or infrastructure. Azure App Service, for building and hosting web applications, is a common example. The customer focuses on their application code and data; Azure manages everything below that layer.
Software as a Service (SaaS)
SaaS provides a complete, ready-to-use application delivered over the internet, typically accessed through a web browser. Microsoft 365 is a widely used example. The customer doesn't manage infrastructure, a platform, or even the application itself — they simply use it, configuring only their own data and settings.
Choosing between them
The right model depends on how much control versus convenience a workload needs. A team that needs deep control over the operating system and networking (for example, running a specialized legacy application) might choose IaaS. A team building a new web application that doesn't want to manage servers might choose PaaS. A team that just needs email and office productivity tools would use a SaaS product rather than building anything themselves.
Assess how much infrastructure control is needed
If the workload requires deep OS-level control or custom configurations, IaaS is likely the right fit; if not, a higher-level model reduces management burden.
Consider development vs. consumption needs
If a team is building a custom application, IaaS or PaaS applies depending on how much infrastructure they want to manage. If the need is simply to use existing software, SaaS applies instead.
Weigh control against operational overhead
Each step from IaaS to PaaS to SaaS trades some control for reduced management responsibility — there's no universally 'best' choice, only the right fit for the specific need.
Recognize that a real environment often uses all three
Most organizations combine service models — for example, using SaaS for email, PaaS for a custom web app, and IaaS for a workload that needs specific OS-level configuration.
A company running a legacy application that requires a very specific operating system configuration uses IaaS virtual machines, since it needs OS-level control that PaaS wouldn't allow. A separate team at the same company builds a new customer-facing web application using PaaS, so they can focus entirely on their code rather than managing servers or patching operating systems.
Meanwhile, the whole company uses a SaaS product for email and office collaboration tools rather than building or hosting anything themselves. This mix — IaaS for one need, PaaS for another, SaaS for a third — is typical; organizations rarely use only one service model exclusively.
Objective 1.2 is one of the most heavily tested areas on AZ-900 — expect multiple questions asking you to identify which service model fits a described scenario, or to match a responsibility to the correct model.
A common wrong answer is confusing PaaS and SaaS — remember PaaS is a platform for *building* applications, while SaaS is a finished, ready-to-use application. Another common trap is assuming IaaS always means "no management by Azure at all" — Azure still manages the physical infrastructure and virtualization layer even under IaaS.
Stable terms: IaaS, PaaS, SaaS. Knowing a clear, correct example of each (VMs for IaaS, App Service for PaaS, Microsoft 365 for SaaS) is one of the most reliable ways to answer scenario-based questions correctly.
Memory trick: IaaS = you manage everything above the physical hardware. PaaS = you manage just your app and data. SaaS = you manage only your own settings and data within a finished product.
IaaS provides infrastructure building blocks (VMs, storage, networking); the customer manages the OS, applications, and data.
PaaS provides a ready-to-use platform for building applications; Azure manages the OS and runtime, the customer manages their app and data.
SaaS provides a complete, ready-to-use application; the customer manages only their own data and settings.
Most organizations use a mix of all three service models across different workloads, rather than committing to just one.
These come up on the exam all the time. Here's how to tell them apart.
IaaS
Customer manages OS, applications, and data
Most control, most management responsibility
Example: Azure Virtual Machines
Good fit when deep infrastructure control is needed
PaaS
Azure manages OS and runtime; customer manages app and data
Less control, less management responsibility
Example: Azure App Service
Good fit for building applications without managing servers
Mistake
PaaS and SaaS are basically the same thing.
Correct
PaaS is a platform for building and running your own applications; SaaS is a complete, finished application you simply use. They serve very different purposes.
Mistake
IaaS means Azure has no involvement in management.
Correct
Azure still manages the physical infrastructure and virtualization layer under IaaS — the customer manages the operating system, applications, and data on top of that.
Mistake
One service model is always 'better' than the others.
Correct
Each model trades control for reduced management overhead differently. The right choice depends entirely on the specific workload's needs, not a universal ranking.
IaaS gives you infrastructure building blocks (like virtual machines) that you configure yourself. PaaS gives you a ready-made platform to build and run applications on, without managing the underlying OS. SaaS gives you a complete, finished application to simply use. Each shifts more management responsibility to Azure as you move from IaaS to SaaS.
Azure Virtual Machines is a classic IaaS example. Azure App Service (for hosting web applications) is a common PaaS example. Microsoft 365 (email, Word, Excel, delivered as a ready-to-use service) is a widely recognized SaaS example.
No — most use a mix, choosing the model that fits each specific workload. A company might use IaaS for a legacy application needing OS control, PaaS for a new custom application, and SaaS for standard productivity tools.
You've just covered IaaS, PaaS, and SaaS Explained — now see how well it sticks with free AZ-900 practice questions. Full explanations included, no account needed.
Done with this chapter?