Examining cloud service models

No cloud computing discussion would be complete without mentioning Infrastructure as a Service (IaaS), Platform as a Service (PaaS), and Software as a Service (SaaS). Although other cloud service models exist, these are the most common cloud models offered by cloud providers and are the standard models per the National Institute of Standards and Technology (NIST). Let’s briefly recap these models before we get into Azure specifics:

• Infrastructure as a Service (IaaS):
  • As the name suggests, with this model, the cloud provider manages the underlying infrastructure. One of the most common examples of this implementation is cloud-based virtual machines – you can request virtual machines with certain specifications and the cloud provider has automation take care of provisioning them for you. You don’t need to worry about the underlying host or infrastructure. You will manage the virtual machine, including installing software, updates, and more.
  • This is the closest model to managing physical machines, without needing to manage any hardware. Provisioning virtual machines is much faster than having to procure and set up physical machines – as is scaling and removing them. You get the most flexibility and control with this model.
  • With the consumption-based model, you only pay for the resources you use (OpEx). For example, if you have a virtual machine and enable daily auto-shutdown, automation will take care of shutting the machine down each day at the desired time, allowing the machine to be stopped and deallocated, at which point the virtual machine itself no longer incurs any costs until it’s started again (other resources attached to the virtual machine may still incur separate costs).
• Platform as a Service (PaaS):
  • From a developer’s perspective, this is becoming the most popular model. With this model, developers can deploy their applications to a managed hosting environment.
  • The cloud provider manages all the underlying hardware and infrastructure as with IaaS, but with PaaS, they also manage any virtual machines, the operating systems, patching, and more. You, as a developer, just manage the application, along with the data and access to it.
  • The consumption-based model makes this a cost-effective model. Being able to provision an environment where you can host your application quickly, deploy and test your application, then potentially remove the environment (an Azure App Service, for example) without needing to worry about any of the details of the machines on which this environment is hosted greatly improves developer agility.
• Software as a Service (SaaS):
  • With this model, the cloud provider provides you with the software. You just manage who can use the software and which features.
  • One of the most common Microsoft-hosted examples of SaaS is Microsoft 365, which includes applications from what was previously called Office 365 – you are provided with the software; you just need to define who can access what features. The cloud provider handles everything else.
  • Typically, this is a pay-as-you-go model – monthly or annually, regardless of how much the software is used. The software is provided as-is. Because you don’t manage the software, if features of the software have limitations, you don’t have control over these features, unlike if you were managing the software itself.

The following diagram shows the service models compared to on-premises, showing which aspects you manage in each. There are many similar illustrations of this all over the internet, so you will have likely come across something similar at some point:

Figure 1.2 – Illustration of the different cloud service models and management responsibility

Increasing numbers of applications are being redesigned for hosting on PaaS services, allowing developers to focus on developing and have the infrastructure details not be a concern, greatly improving developer velocity. Whether that is the design decision or not, it’s important to understand the different options available. Now that we’ve had a recap of some cloud fundamentals, it’s time to start delving into Azure specifics.