Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
The shared responsibility model divides cloud security between provider and customer. The dividing line moves with the service type, and almost every cloud breach sits on the customer’s side of it.

What Is the Shared Responsibility Model?
The shared responsibility model is the agreement, set out by every major cloud provider, about which party secures which part of a cloud service. Amazon phrases it as security of the cloud versus security in the cloud: the provider protects the physical sites, the hardware, the hypervisor, and the managed service software, while the customer protects their data, identities, configurations, and applications. Microsoft and Google publish equivalent charts. The model is not a legal nicety — it is the reason most cloud incidents are the customer’s to prevent and the customer’s to report.
| Responsibility | On-premises | IaaS | PaaS | SaaS |
|---|---|---|---|---|
| Physical and network fabric | You | Provider | Provider | Provider |
| Hypervisor and host | You | Provider | Provider | Provider |
| Guest operating system and patching | You | You | Provider | Provider |
| Application code and dependencies | You | You | You | Provider |
| Service configuration | You | You | You | You |
| Identities and permissions | You | You | You | You |
| Data and its classification | You | You | You | You |
Read the bottom three rows carefully, because they never change. Whatever the service model, you own your data, who can reach it, and how the service is configured. Google describes a variant it calls shared fate, in which the provider supplies opinionated secure defaults, blueprints, and guardrails rather than simply drawing a line — a useful corrective to a diagram that can read as a disclaimer.
Incidents cluster around a small number of beliefs that the model quietly contradicts.
The 2019 Capital One breach is the canonical illustration of the whole model. A misconfigured web application firewall, an over-permissive role, and data accessible to that role produced a breach affecting more than 100 million people — every link in the chain on the customer side of the line.
Treat the provider’s diagram as a template, then produce your own version per service you consume, naming the internal team that owns each row. The exercise is more valuable than it sounds: it routinely surfaces responsibilities that everyone assumed belonged to somebody else, particularly logging, backup, key management, and vulnerability management for managed services.
Extend it to your suppliers. If a managed service provider administers your tenant, the division between you, them, and the cloud provider should be written into the contract rather than inferred. The same applies in reverse where you are the provider: your customers need to know which parts of tenant isolation and configuration are theirs to manage.
Data protection law does not recognise the shared responsibility model, and that is the point worth understanding. Under UK GDPR, you are generally the controller and the cloud provider is a processor acting on your instructions. Accountability for the security of processing stays with the controller, as does the duty to notify the ICO within 72 hours where a breach is likely to result in risk to individuals. “The configuration was the default” has never been accepted as mitigation.
The NCSC Cloud Security Principles are structured around the same division, asking what the provider asserts and what you must implement. For financial entities, DORA and the UK critical third parties regime push further still, requiring firms to map dependencies and keep accountability even where a designated provider is supervised directly.
If you want an independent view of which side of the line your current gaps fall on, our cloud security services assess configuration, identity, and exposure across AWS, Azure, and Google Cloud.
The principle is identical and the diagrams differ only in presentation and terminology. What genuinely varies is the detail within equivalent services: default encryption behaviour, how permission inheritance works, what logging is enabled without being asked, and which hardening the provider applies to managed components. Read each provider’s documentation for the specific services you use rather than relying on a general chart.
You are. Providers such as Microsoft and Google offer retention policies, versioning, legal hold, and recycle bins, which serve compliance and accidental deletion rather than recovery from a destructive incident. Their terms are explicit that customer data remains the customer’s responsibility. If your recovery requirement is restoring a defined state at a defined time, you need a third-party backup with its own retention.
To your customers and your regulator, yes. The provider owes you whatever its service level agreement promises, which is typically a service credit rather than compensation for your losses. Accountability for the availability of your service remains with you, which is why resilience planning, multi-region design, and tested recovery sit firmly on the customer side of the model.
It adds a third party to the division without removing your accountability. Write down precisely which controls the provider operates, which remain yours, and how evidence is supplied, then put that in the contract rather than leaving it to assumption. Gaps between a cloud provider’s responsibilities and a managed service provider’s scope are a common and well-exploited source of unmonitored systems.
You May Be Interested In These
We use a few cookies that are needed to make this website work, and they are always on. We would also like to set analytics cookies, to see which pages are useful; functional cookies, so that our live chat works; and advertising cookies, to measure our campaigns. We will not set any of those until you say yes, and you can change your mind at any time from the Cookie Settings link in the footer.
Choose the cookies you are happy for us to use. Strictly necessary cookies are always on; everything else is off until you turn it on.
These cookies keep the website working: they hold your security checks, your form submissions and the choice you make here. The site cannot run without them, so they are always on.
These let our live chat open and remember your conversation, and let the newsletter sign-up work. Switch them off and the chat window will not appear.
These tell us which pages are read and how people find us, so that we can improve the site. The reports are about groups of visitors, never about you personally.
These let us measure our advertising and show you our adverts on other websites, such as LinkedIn, Google and Meta. Switch them off and you will still see adverts, but they will be less relevant.