Explore our new Cybersecurity Terms Glossary, explained by our experts.
Cybersecurity Terms Glossary
A trust boundary is the line where data or a request moves between zones with different levels of trust. Almost every serious vulnerability lives on one of those lines.

What Is a Trust Boundary?
A trust boundary is the line in a system where data, a request, or control passes between two areas that trust each other to different degrees. The browser and your API sit on opposite sides of one. So do your application and its database, your network and your supplier’s, and user mode and the kernel. Draw these lines on a design and you have drawn a map of where security controls have to live, because everything crossing a boundary has to be treated as untrusted until it is checked.
The work at every crossing is roughly the same, and the order matters.
Client-side validation is a usability feature, never a control. It lives on the wrong side of the line and an attacker simply does not run it.
The expensive failures happen when a boundary exists in reality but not on the diagram. The Capital One breach disclosed in July 2019 is the textbook case: a server-side request forgery flaw let an attacker make the application fetch a URL of their choosing, and that request reached the cloud instance metadata service. The application and the metadata endpoint were on opposite sides of a trust boundary that nobody had drawn, and the credentials returned led to roughly 106 million records.
Modern architectures keep creating new lines. A CI/CD pipeline that builds a pull request from a fork executes untrusted code with access to build secrets. A large language model application that reads a web page or a document has just taken untrusted input into its instruction stream, which is the mechanism behind prompt injection. A browser extension sits between your users and every site they visit. Each is a boundary, and each needs the same six steps as a web form.
In a data flow diagram, trust boundaries are drawn as dotted lines cutting across the flows. STRIDE is then applied most intensively to the elements and flows that touch them, because that is where spoofing, tampering, and information disclosure concentrate. If a design review produces nothing else, getting the team to agree on where the boundaries are has already been useful—disagreement about that is itself a finding.
A related discipline is deciding what happens when the check at a boundary cannot run: see fail-secure and fail-safe. And at network level, segmentation turns a conceptual boundary into an enforced one, while zero trust takes the argument further by removing implicit trust from the internal network altogether.
Finding boundaries that are documented but not enforced is one of the most productive parts of an assessment. Our web application security services test every crossing we can reach, including the ones the design did not mention.
A firewall enforces one, but the boundary is the conceptual line between zones of differing trust, not the device. The same boundary may be enforced by a firewall, a security group, a service mesh policy, or an authorisation check in code. Boundaries also exist where no device sits at all, such as between two processes on one host.
Yes. Most serious intrusions involve an attacker who already has a foothold inside the network moving sideways, so internal boundaries decide how far they get. Service-to-service authentication, per-object authorisation, and segmentation between environments all limit lateral movement. Treating the internal network as trusted is the assumption zero trust architectures were created to remove.
Sketch the data flows and ask, at each arrow, whether the two ends are controlled by the same party with the same level of assurance. Wherever the answer is no—different owner, different privilege level, different tenant, different network zone, or a human in the loop—you have a boundary. Mark it and check which of the six boundary controls are present.
Because it runs on the untrusted side of the boundary. An attacker uses a proxy or a script instead of your interface and sends whatever they like straight to the API. Client-side checks are worth keeping for usability and quick feedback, but every rule that matters must be re-applied on the server, where you control the code.
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.