Dev.to Security πŸ” Cybersecurity πŸ‘ 0 πŸ“– 3 min read

EAP-TLS Needs Certificate Identity Mapping, Not Just a Trusted CA

EAP-TLS Needs Certificate Identity Mapping, Not Just a Trusted CA EAP-TLS uses certificates to authenticate both the supplicant and the authentication server. In a typical 802.1X deployment, a client presents a certifi

EAP-TLS Needs Certificate Identity Mapping, Not Just a Trusted CA

EAP-TLS uses certificates to authenticate both the supplicant and the authentication server. In a typical 802.1X deployment, a client presents a certificate to the RADIUS server, and the RADIUS server presents its own certificate to the client. A trusted CA is the starting point. The authorization decision also needs a mapping from the certificate to the device or user that is allowed to enter the network.

That mapping is where many deployments become weaker than their diagrams. A certificate may be valid and issued by the right CA but belong to a retired laptop, a shared service account or a device in another group. If the policy only checks that the chain is trusted, the certificate becomes a portable network key. The identity in the certificate should be compared with an authoritative inventory before the VLAN or role is assigned.

Define which fields matter

The server certificate should be validated for the intended RADIUS endpoint, including the trusted root, subject alternative names and revocation status. The client certificate should be checked for the correct extended key usage, validity period and issuance policy. The identity mapping should specify whether the server reads the common name, a subject alternative name or another attribute, and how it resolves that value to a device or user record.

The certificate lifecycle should be linked to the identity lifecycle. When a device is retired, its certificate should be revoked or expired before the inventory is closed. When a user changes role, the network policy should change even if the certificate remains valid. A long-lived certificate with no renewal owner is a quiet path back into the network.

Separate user and machine authentication

Many organizations deploy both computer certificates and user certificates. The network policy should state which one is required for which access. A machine certificate may be appropriate for boot-time network access, while a user certificate may be required for a higher-trust segment. Lumping both into one group can grant a user the access intended only for the device.

Guest or onboarding networks need their own trust model. A device that has not yet received a certificate should not be allowed to select a privileged VLAN. The onboarding flow should issue a short-lived credential, bind it to the device inventory and record the approval. A fallback password method may be needed for legacy devices, but it should be limited and monitored.

Test revocation and failure behavior

Test an expired certificate, a revoked certificate, a certificate from a different CA and a certificate with an unexpected identity. Confirm that each failure is logged with a reason and that the client lands in a controlled network or no network. Then test a RADIUS service outage. The switch should not silently open the port or fall back to a weaker method.

Monitor certificate issuance, renewal failures and authentication failures by device class. A spike in failures after a CA change may be more useful than a generic alert on the NPS server. Keep the mapping rules under version control and review them with the identity and network teams.

EAP-TLS can provide strong mutual authentication. The outcome depends on the operational detail after the TLS handshake: the right certificate, the right identity, the right revocation check and a network policy that knows which device is allowed to use which path.

References

πŸ“° Read the original article on Dev.to Security

Originally published by Dev.to Security. Aggregated on AIWithGhost for educational purposes β€” full credit and traffic to the original publisher.