What is zero trust security?

In the old model, the network worked like a building with a guarded entrance. Once inside the door, people could move around freely. The problem is that an attacker who gets in through the entrance can also move around freely. Zero trust reverses that assumption: nobody is let in on trust, whether outside or inside the walls.

Zero trust is a security model built on a simple rule: trust nothing, verify everything. Every user and every device must prove who it is and that it has the right to reach the specific resource it is asking for, every time.

The idea behind zero trust

The model assumes that threats can exist anywhere, including on the inside. A stolen login, an infected laptop or a disgruntled employee means that guarding the boundary with the outside world is not enough. The old idea that everything on the inside is safe no longer holds when cloud services, subcontractors and remote working have blurred where the inside even begins.

Instead, the system assesses every attempt to reach a service based on its context: who is asking, from which device, from where and what information is involved. If the picture does not fit, the attempt is blocked, even if the password is correct.

Least privilege

At the heart of zero trust is giving everyone exactly as much access as they need, and no more. A finance assistant needs the payroll system, not the entire server estate. If an attacker still gets hold of that account, they can only reach the payroll system, not everything else.

Strong authentication is part of this. Multi-factor authentication, where the password is supplemented with something more, carries more weight than a password alone. The network is also divided into smaller zones, known as microsegmentation, so that anyone who does get into one part does not automatically reach the rest.

Introducing zero trust step by step

Zero trust is not a product that comes in a box, but a way of working that develops over time. Most organisations start in three steps.

First, the organisation maps what it has: which users, devices, services and data, and how they are connected. Next, the biggest risks are dealt with one by one, using procedures and tools that detect and stop anomalies. Finally, the approach is extended to more parts of the organisation. A common framework for this work is the US standard NIST SP 800-207, which describes the principles in detail. The point of going step by step is that the organisation can start where the risk is greatest, instead of rebuilding everything at once.

What the model protects against

Zero trust is most useful in environments where the old perimeter has broken down. It is aimed in particular at:

  • Remote and hybrid working, where devices connect from outside the office.
  • Cloud services, where data is stored outside the organisation's own network.
  • Insider threats, where someone on the inside misuses their access.
  • Stolen credentials, where the right password in the wrong hands would otherwise give free rein.

What they have in common is that protection can no longer rely on where someone is. Zero trust shifts the question from location to identity and context.

Zero trust beyond IT

The mindset does not stop at computers. The same idea, that everyone should only be able to reach what they genuinely need, is familiar from physical security. An access control system that only opens the doors a person actually needs follows the same logic in practice. A card that works everywhere is convenient but risky, whereas one that only unlocks the right doors limits the damage if it is lost.

For an organisation that is already reviewing its information security and cybersecurity, zero trust is less a new gadget and more a way of thinking. Regulations such as the EU's NIS2 Directive point in the same direction, towards tighter control of who can access what.