Cloud computing fundamentals | 14 min read

What are the main cloud service models?

WRITTEN BY Leigh Ryan
cover image

Part 1 of our series on fundamental cloud computing concepts was all about cloud deployment models. Quick recap: your choice of deployment model (public cloud, private cloud, hybrid cloud) dictates where your cloud infrastructure and applications will be hosted—in a public cloud vendor's data center, in a self-hosted data center, or in a combination of both.

Today's topic, cloud service models, relates to how that data is hosted and handled by your cloud provider.

Already familiar with the concept of cloud service models? For a more in-depth look at the ins and outs of infrastructure as a service (IaaS) and platform as a service (IaaS), read IaaS vs. PaaS: Which one is best?

In this article, we'll:

  • Take a quick look at what the term cloud service model actually entails
  • Provide a bite-sized definition of the four main ones—IaaS, PaaS, FaaS and SaaS
  • Compare the main characteristics of each model side by side in a handy table (jump to the recap table)

What is a cloud service model?

Wikipedia's list of as-a-Service offerings
Source: Wikipedia

Over the past decade or so, the rise of cloud computing has spawned an entire industry of "somethings as a service" (highlights from Wikipedia's list include farming as a service and drone as a service).

Keeping it simple, the -as-a-service part of all those initialisms is just another way of saying 'available on-demand from a third party thanks to the cloud'.

In the context of an organization's IT and development environments, cloud computing makes the building blocks for any type of application instantly available to purchase and use, without having to consider the underlying physical infrastructure required to run them.

Choosing a cloud service model to host a given application means deciding which of that application's building blocks is going to be managed by the cloud vendor, and which parts are left up to you, the subscriber.

The general consensus is that there are three main service (or cloud delivery) models: infrastructure, software, and platform as a service.

A fourth model, function as a service—FaaS, sometimes used interchangeably with the term serverless—is also becoming increasingly prevalent. We'll cover it quickly below, but the whole serverless/functions/Lambda movement is in a league of its own and definitely deserves its own series of posts.

On-prem vs. the cloud

To better understand each of the main cloud service models, let's start by looking at the way an organization's computer infrastructure, management, and development was typically handled in the days before the cloud.

For applications hosted in a traditional on-premises environment (an internal email server or a corporate intranet, for example), all of the following elements had to be managed by internal IT:

  • Physical data center infrastructure
  • Security
  • Networking
  • Storage
  • Servers
  • OS
  • Database management
  • BI services
  • Development tools
  • Middleware
  • Hosted applications

In cloud computing, a third party—the vendor—manages some or all of the above elements. The ideal balance depends on the nature of the application: some custom solutions will require more granular control; other times, it makes more sense to offload everything to the vendor.

Cloud services can be classified as either infrastructure, function, platform, or software as a service, depending on which of the elements listed above are managed by the vendor, and which are controlled by the subscriber.

The first model, infrastructure as a service, has the fewest vendor-managed components; on the other end of the spectrum, SaaS offerings are 100% vendor-managed, from hosting and data storage all the way to the application itself.


Infrastructure as a service (IaaS): the blank slate

Infrastructure as a service is the most flexible of the three cloud service models, but it's also the one with the fewest concrete applications on its own. Why? Simply because, like a blank canvas or a box of parts, it requires a fairly broad skillset to make use of stand-alone IaaS efficiently.

In IaaS, the cloud provider is responsible for the basics: servers, networking, security, storage. Working in IaaS isn't really different than working on-prem, at least functionally--the only real difference is that someone else is in charge of maintaining and securing the physical components of your environment.

IaaS: who's in charge of what?

Cloud vendor manages

  • Physical data center infrastructure
  • Security
  • Networking
  • Storage
  • Servers

You manage

  • Operating system
  • Database management
  • BI services
  • Development tools
  • Middleware
  • Hosted applications

Platform as a service (PaaS): the timesaver

PaaS delivers all the benefits and functionality of IaaS as well as a framework on which to quickly develop different types of cloud applications. Platform as a service offerings comprise a bundle of managed services or pre-coded components designed to speed up the development process from start to finish.

The fastest-growing cloud delivery model according to Gartner, PaaS aims to streamline app development by providing all the capabilities needed to build, deploy, test and update applications within a single, integrated environment. PaaS is ideal for getting applications to market quickly while allowing development teams to focus less on low-level coding and tedium, and more on efficiency and output.

PaaS: who's in charge of what?

Cloud vendor manages

  • Physical data center infrastructure
  • Security
  • Networking
  • Storage
  • Servers
  • Operating system
  • Database management
  • BI services
  • Development tools
  • Middleware

You manage

  • Hosted applications

Software as a service (SaaS): the whole package

This cloud service model is probably the most straightforward, and certainly the most widespread. Just look at Microsoft's Office suite. Not so long ago, software had a strong physical facet: if you wanted to roll out the latest version of Word, PowerPoint and the like to your users, you needed CDs and valid license keys, then install the software manually on each machine.

Now, all you need to do is sign your organization up for to Office 365 to provide your users with on-demand access to all of the apps included in the suite. No more on-premises Exchange servers or manual updating to worry about. Everything is managed by the vendor.

SaaS: who's in charge of what?

Cloud vendor manages

  • Physical data center infrastructure
  • Security
  • Networking
  • Storage
  • Servers
  • Operating system
  • Database management
  • BI services
  • Development tools
  • Middleware
  • Hosted applications

You manage

  • All that sweet time you're saving

Function as a service (FaaS): a new frontier?

Microsoft includes function as a service in its list of major cloud service models. Another term that's sometimes used interchangeably with FaaS is 'serverless', which does a pretty good job describing what makes this model so radically different from the first three: it lets you build and serve apps without a traditional development environment—i.e., without servers.

With FaaS, developers can build apps faster than ever before by focusing their efforts on coding the different 'pieces' of the app, then mapping them to triggers. The FaaS framework takes care of connecting all the parts as they should. It's a bit like IFTTT or Zapier, but for rapid cloud application development and deployment.

The pieces of code and the triggers that tie them together are referred to as Azure Functions in the Microsoft cloud, and Lambda functions in the AWS realm. If you're interested in this new approach to rapid app development, check out the open-source Serverless Framework, which acts as the backbone for managed FaaS services such as those in Azure or AWS. The resources and documentation sections are full of examples and best practices to get you up to speed.


IaaS vs. PaaS vs. FaaS vs. SaaS at a glance

You're now somewhat familiar with the basics of each main cloud service model (infrastructure as a service, platform as a service, function as a service, and software as a service). Here's a table comparing each model's main characteristics. We'll expand on most of these points in upcoming posts, so stay tuned!

PaaS vs. SaaS vs. FaaS vs. IaaS - cloud service models compared - ShareGate

Want to learn more about the cloud computing and its potential for businesses? Written by Azure MVP Jussi Roine and published by ShareGate, Modern Business Powered by Microsoft Azure is a full-length book offering a foundational overview of the cloud, with a focus on Microsoft Azure, the fastest-growing public cloud platform in the industry.