The Shared Responsibility Confusion Problem: What Microsoft Secures in Azure and What You Still Have to Protect

The Azure shared responsibility model is the most misunderstood idea in cloud security. Plenty of teams move workloads into Azure believing Microsoft now ‘looks after security’, and then discover, usually after an incident, that a large slice of the job never left their desk. That gap between what people assume and what is actually true is the confusion problem, and it is behind a surprising number of avoidable cloud breaches.

This guide explains the split in plain English: what Microsoft secures, what stays with you, and how the line moves depending on the type of service you use. If you’re studying for the Microsoft Azure Fundamentals exam, you’ll also see why this topic appears in the AZ-900 certification course and how to learn it properly.

Azure shared responsibility model chart showing what Microsoft secures and what the customer protects across SaaS, PaaS and IaaS

Simplified from Microsoft’s published shared responsibility model.

What Is the Azure Shared Responsibility Model?

Think of renting a serviced office. The building owner looks after the structure, the power, the lifts and the front-door security. You still decide who gets a key, whether the filing cabinets are locked and what happens to your documents. The Azure shared responsibility model works the same way: Microsoft secures the cloud platform itself, and you secure what you put in it and how you configure it.

Microsoft publishes its own explanation on Microsoft Learn, and it’s worth reading alongside this article. The key point is that responsibility is never all-or-nothing. It varies by service type, and some duties are always yours no matter what you buy.

What Microsoft Secures in Azure

Under the Azure shared responsibility model, Microsoft always takes care of the foundations that you could never realistically manage yourself:

  • Physical datacentres: buildings, power, cooling, fire protection and physical access controls.
  • Physical hosts: the servers and hardware that run your workloads.
  • Physical network: the cabling and core networking inside and between datacentres.
  • The cloud platform: hypervisors, the underlying fabric and the security of Azure’s own services.

Microsoft also holds a wide range of independent compliance certifications for its platform. That’s reassuring, but it’s where many misunderstandings begin, which we’ll get to shortly.

What You Still Have to Protect

Regardless of whether you use virtual machines, databases or Microsoft 365, some responsibilities are always on the customer’s side of the Azure shared responsibility model:

  • Your data: classifying it, encrypting it where required, controlling who can read it and making sure it can be recovered.
  • Accounts and identities: strong passwords, multifactor authentication, sensible admin roles and removing access when people leave.
  • Devices and endpoints: the laptops, phones and servers that connect to your cloud environment.
  • Access and permissions: applying least privilege so people only see what they genuinely need.
  • Configuration: storage settings, firewall rules, public endpoints and everything else you can switch on or off.

Notice that most of that list is about people and settings, not technology failures. Industry reporting on cloud incidents regularly points to misconfiguration and weak identity controls rather than flaws in the platform itself.

How Responsibility Shifts Between IaaS, PaaS and SaaS

This is the part that trips people up. As you move from infrastructure to software services, Microsoft takes on more, but never everything.

Infrastructure as a Service (IaaS)

Azure Virtual Machines are the classic example. Microsoft runs the hardware and hypervisor, but you manage the guest operating system, patching, installed software, network rules and the data inside. If a VM isn’t patched, that’s on you.

Platform as a Service (PaaS)

With services such as Azure App Service or Azure SQL Database, Microsoft handles the operating system and much of the platform. You’re still responsible for your application code, data, identities and how access is configured.

Software as a Service (SaaS)

In a SaaS product like Microsoft 365, Microsoft runs nearly the whole stack. You remain responsible for your data, user accounts, devices and sharing settings. A perfectly secure platform can still leak files if links are shared publicly.

A simple rule helps: the further you move towards SaaS, the less infrastructure you manage, but identity and data are always yours.

Five Azure Shared Responsibility Model Mistakes That Lead to Breaches

1. Assuming Microsoft backs up everything

Platform resilience isn’t the same as a recoverable backup of your data. If someone deletes a critical database or ransomware encrypts a file share, you need your own backup and restore plan, tested regularly.

2. Trusting default settings

Defaults are designed for ease of use, not for your specific risk profile. Review storage access, network exposure and logging for every new resource instead of accepting what’s pre-selected.

3. Treating Microsoft’s compliance certificates as your own

Microsoft’s certifications cover its platform. They don’t automatically make your workload compliant. You still have to configure, monitor and document your own controls.

4. Leaving multifactor authentication optional

Stolen credentials are one of the easiest ways into any cloud tenant. Requiring multifactor authentication, particularly for administrators, closes a major gap. Microsoft Entra is where these identity policies live.

5. Believing PaaS means no patching

Microsoft patches the platform, but your own application code, libraries and dependencies are still your job. Vulnerable code running on a secure platform is still vulnerable.

A Practical Azure Shared Responsibility Model Checklist

Use this as a starting point for any new Azure workload:

  1. Classify your data so you know what needs the strongest protection.
  2. Turn on multifactor authentication for every user, and enforce it for admins first.
  3. Apply least privilege using Azure role-based access control, and review roles quarterly.
  4. Secure your network with sensible firewall rules and avoid exposing management ports to the internet.
  5. Enable monitoring and alerts and review recommendations in Microsoft Defender for Cloud.
  6. Use policy to prevent drift, for example with Azure Policy, so risky configurations get flagged or blocked.
  7. Back up and test restores because an untested backup is only a hope.
  8. Document who owns what so nobody assumes someone else is handling it.

The Azure Shared Responsibility Model for Australian Businesses

Australian organisations often align their cloud controls with guidance such as the ACSC’s Essential Eight, available from the Australian Cyber Security Centre. Whether you’re a start-up in Brisbane, a government team in Canberra or an enterprise in Sydney or Melbourne, the same principle applies: knowing where Microsoft’s responsibility ends is the first step to meeting yours.

That’s why many IT teams, from Perth and Adelaide through to Hobart, Darwin and the Gold Coast, build this knowledge into their onboarding. Understanding the Azure shared responsibility model helps cloud administrators, help desk staff, security analysts and even non-technical managers make better decisions.

Why the Azure Shared Responsibility Model Is Part of AZ-900

The Azure Fundamentals exam covers core cloud concepts, and the shared responsibility model is one of the ideas you’re expected to understand. You can review the exam outline on the official Microsoft Azure Fundamentals certification page.

If you’d rather learn with an instructor, our AZ-900 courses are hands-on and available in several formats: live online, in-house for your team, full-time on weekdays, or part-time on weeknights and weekends. You can explore the Microsoft Azure training course and the AZ-900 course details to compare options.

Looking for training in your city? We run classes for Melbourne, Sydney, Brisbane, Adelaide, Canberra and Perth, and our locations page lists every option. Want to know why people choose us? Read why choose our AZ-900 training or learn about us.

Azure Shared Responsibility Model FAQs

What does the Azure shared responsibility model mean?

It means security in Azure is divided between Microsoft and the customer. Microsoft secures the physical infrastructure and platform, while you protect your data, identities, devices and configuration.

Is the customer always responsible for data?

Yes. Whether you use IaaS, PaaS or SaaS, you remain responsible for your data, your user accounts and how access is granted.

Who is responsible for patching in Azure?

It depends on the service. For virtual machines (IaaS), you patch the operating system and software. For PaaS, Microsoft handles the platform, and you handle your own application code.

Does the Azure shared responsibility model appear on the AZ-900 exam?

Yes, it’s part of the cloud concepts you’re expected to understand. Practising real scenarios, such as deciding who owns a given task, is the best way to remember it.

Where can I get answers to other AZ-900 questions?

Our AZ-900 certification FAQs cover exam format, preparation and course options, and the blog has more study guides.

Stop Guessing Who Owns Azure Security

Clouds aren’t insecure by default, and they aren’t secure by default either. They’re secure when both sides do their part. Once you clearly understand the Azure shared responsibility model, you can stop assuming and start assigning, which is where real protection begins.

Ready to build the skills? Contact our team to discuss AZ-900 training for you or your organisation, or browse the course details to get started.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top