Explore our new Cybersecurity Terms Glossary, explained by our experts.

View the Glossary View the Glossary
PRIVIA

Cybersecurity Terms Glossary

What Is RADIUS?

RADIUS is the protocol network devices use to ask a central server whether a user may connect. It dates from 1991 and its original cryptography shows it.

Illustration for the cybersecurity term RADIUS
Home / Blog / Cybersecurity Terms / What Is RADIUS?
17 November 2026

What Is RADIUS?

What Is RADIUS?

RADIUS—Remote Authentication Dial-In User Service—was written in 1991 for a problem that no longer exists: letting a bank of modems check dial-up users against a central database instead of each holding its own list. The problem disappeared; the protocol did not. Today it authenticates enterprise Wi-Fi, VPN concentrators, wired 802.1X switch ports, and network device logins across nearly every organisation.

RFC 2865 and 2866 define it. Microsoft NPS, FreeRADIUS, Cisco ISE, and Aruba ClearPass are the implementations you will meet.

The Three A’s

The Cryptography Is the Problem

RADIUS runs over UDP and protects almost nothing. Only the User-Password attribute is obscured, using an MD5-based construction with a shared secret. Every other attribute—username, assigned VLAN, access list, accounting data—travels in clear text. Response integrity relies on a Response Authenticator, also built on MD5.

That design held up for three decades mainly because the traffic was assumed to stay on a trusted management network. In July 2024 researchers published Blast-RADIUS, catalogued as CVE-2024-3596, which exploits an MD5 chosen-prefix collision to forge a valid Access-Accept from an Access-Reject. An attacker positioned on the path between the network access server and the RADIUS server can turn a rejected login into an accepted one without knowing the shared secret or any user credential.

The mitigations are specific. Require the Message-Authenticator attribute on all packets, which most vendors shipped as an enforced default in patched releases. Move RADIUS traffic onto a dedicated, isolated management network. And where possible move to RADIUS over TLS, known as RadSec and specified in RFC 6614, which wraps the whole exchange in TLS on TCP port 2083 and eliminates the class of problem rather than patching one instance of it.

Shared Secrets and Other Operational Debt

Each network access server shares a secret with the RADIUS server, and in most estates those secrets are short, reused across every device, and unchanged since installation. A secret recovered from the configuration of one access point in an unlocked cupboard therefore unlocks the ability to impersonate any device in the estate.

Set a unique random secret of at least 22 characters per device, store them in a vault, and rotate them when network staff leave. While you are there, restrict which source addresses the RADIUS server will accept requests from, and make sure accounting logs are forwarded somewhere they survive a device reboot.

What It Authenticates, and How Well

RADIUS itself does not perform the authentication; it carries a method. For Wi-Fi and wired 802.1X that method is EAP, and the choice between EAP-TLS and PEAP-MSCHAPv2 decides far more about your security than RADIUS does. A rogue access point attack against a client that does not validate the RADIUS server certificate harvests domain passwords regardless of how well the RADIUS deployment itself is configured.

For administrative logins to switches and routers, RADIUS is widely used but TACACS+ is the better fit, because it separates authentication from authorisation and can authorise individual commands rather than granting a privilege level wholesale.

Where RADIUS fronts a VPN, add multi-factor authentication at the RADIUS layer. Most platforms support a challenge step that prompts for a one-time password or a push approval, and a VPN protected by a password alone remains one of the most commonly exploited entry points in UK incidents—frequently following a credential-stuffing run against an appliance with no second factor and no lockout.

A Short Deployment Checklist

Patch for CVE-2024-3596 and enforce Message-Authenticator. Use unique long shared secrets. Isolate RADIUS traffic or move to RadSec. Prefer EAP-TLS over password-based EAP methods. Build redundancy, because a single RADIUS server failure means nobody connects to anything. Decide the behaviour when the server is unreachable deliberately, since failing open and failing closed are both defensible and only one of them matches your risk appetite. And forward accounting logs to your SIEM, where they answer the question every investigation starts with: which device was that.

At Privia Security, we test network authentication infrastructure end to end in our network security services.

FAQ – Frequently Asked Questions

What Is Blast-RADIUS and Do We Need to Act on It?

Blast-RADIUS, tracked as CVE-2024-3596 and published in July 2024, uses an MD5 chosen-prefix collision to forge a valid accept response from a reject, allowing an attacker on the network path to authenticate without credentials. Apply vendor patches, enforce the Message-Authenticator attribute on all packets, isolate RADIUS traffic, and move to RADIUS over TLS where supported.

What Is the Difference Between RADIUS and TACACS+?

RADIUS combines authentication and authorisation in one response, encrypts only the password field, and runs over UDP. TACACS+ separates the three functions, encrypts the entire payload, and runs over TCP, which lets it authorise individual commands on a device. RADIUS suits network access for users; TACACS+ suits administrative access to network equipment.

Should RADIUS Traffic Be Encrypted?

Yes. Classic RADIUS obscures only the password attribute and leaves usernames, assigned VLANs, access lists, and accounting data in clear text. RADIUS over TLS, known as RadSec and defined in RFC 6614, wraps the whole exchange in TLS over TCP. Where that is not yet supported, carry RADIUS on an isolated management network at minimum.

How Should Shared Secrets Be Managed?

One unique, randomly generated secret of at least 22 characters per network access server, stored in a vault rather than a spreadsheet, and rotated when network staff leave or a device is decommissioned. Reusing one secret across the estate means that recovering it from any single accessible device allows an attacker to impersonate every other one.

Related Terms

You May Be Interested In These

What Is Firmware Security?

Cybersecurity Terms

17 November 2026

What Is Firmware Security?

Read More
What Is Alert Fatigue?

Cybersecurity Terms

17 November 2026

What Is Alert Fatigue?

Read More
What Is Alert Triage?

Cybersecurity Terms

17 November 2026

What Is Alert Triage?

Read More