Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
Multi-cloud security is the problem of applying one security standard across providers whose identity models, logging, and controls have almost nothing in common.

What Is Multi-Cloud Security?
Multi-cloud security is the discipline of holding one consistent security standard across two or more public cloud providers. The difficulty is not that any individual provider is weak. It is that AWS, Microsoft Azure, and Google Cloud express identity, networking, logging, and encryption in genuinely different terms, so a control that is trivial in one is a project in another, and the gap between them is where incidents live.
Few teams choose multi-cloud on a whiteboard. It arrives through acquisition, through a data science group that wanted BigQuery, through a SaaS vendor that only runs on Azure, or through a board that asked for provider concentration to be reduced. Deliberate reasons exist too — negotiating leverage, regional availability, a service that genuinely has no equivalent elsewhere — but the security team usually inherits the result rather than designing it.
| Area | AWS | Azure | Google Cloud |
|---|---|---|---|
| Permission model | IAM policies on identities and resources | Azure RBAC roles over scopes | IAM bindings on resource hierarchy |
| Default inheritance | None between accounts | Downward from management group | Downward from organisation and folder |
| Audit trail | CloudTrail | Entra ID and Activity logs | Cloud Audit Logs |
| Workload credentials | Instance profiles, IRSA | Managed identities | Service accounts, workload identity federation |
Three practical consequences follow. Permission inheritance means a role granted at an Azure management group reaches subscriptions created months later, which has no AWS equivalent. Log formats, field names, and retention defaults differ enough that a detection rule written once must be rewritten twice. And the same words mean different things: a Google Cloud service account is an identity you can impersonate, while an AWS role is assumed, with different audit evidence in each case.
Write the policy in outcome terms, then implement it per provider. “No storage object is readable without authentication” is portable; “S3 Block Public Access is enabled” is not. From there, a few things do most of the work:
Resilience is the argument most often used for multi-cloud, and it deserves scrutiny: running in two clouds only helps if you have tested failing over, and the duplication itself adds attack surface. For UK financial services the question is now formal. The critical third parties regime created under the Financial Services and Markets Act 2023, with final rules published by the Bank of England and FCA in November 2024, allows designated providers — cloud platforms among them — to be supervised directly, and firms remain accountable for mapping and testing their important business services. Entities in scope of DORA face parallel obligations on concentration and exit planning.
Knowing where the seams are between your providers is worth more than another dashboard. Our cloud security services assess each estate against its own native controls and then test what connects them.
Generally no, at least not without significant investment. Two providers mean two identity models, two logging stacks, two sets of native controls, and two bodies of expertise your team must maintain. The resilience benefit is real for specific failure scenarios, but the day-to-day effect is a wider attack surface and thinner coverage on whichever platform receives less attention.
Use both. Native tooling such as AWS Security Hub, Microsoft Defender for Cloud, and Google Security Command Centre has the deepest signal for its own platform and is usually cheaper. A third-party posture or CNAPP platform gives you one policy set, one severity scale, and one inventory across providers. Most organisations feed native findings into the cross-provider platform rather than choosing one.
Use workload identity federation. Each major provider can trust an OIDC token issued by another, so a job running in Google Cloud can assume an AWS role, or a GitHub Actions workflow can obtain short-lived Azure credentials, without any stored secret. Reserve static keys for the few systems that genuinely cannot federate, rotate them automatically, and alert whenever a new one is created.
They are different problems that often appear together. Hybrid cloud combines on-premises infrastructure with a public provider, and the hard part is usually identity bridging and network trust between the two. Multi-cloud means two or more public providers, where the hard part is reconciling different control models. An estate can be both, which is the point at which standardised policy matters most.
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.