Infinite length chain of RRDP repositories

Vulnerability

Routinator prior to version 0.10.2 does not impose a limit on the depth of a Certificate Authority (CA) chain during RRDP repository validation [1][3]. In RPKI, a CA can select which RRDP repository publishes its data. By repeatedly generating a child CA that itself points to another RRDP repository, an attacker can create a chain of de-facto infinite length. Routinator will process this chain indefinitely, causing the validation run to never complete [1][3].

Exploitation

An attacker with the ability to act as a malicious CA can construct a chain of CAs pointing to different RRDP repositories, each new CA being a child of the previous one, creating an infinite loop. No special network position or authentication is required beyond the ability to publish valid RPKI objects. Routinator will follow the chain until it exhausts resources or runs indefinitely, never finishing the validation cycle [1][3].

Impact

Successful exploitation causes Routinator to never complete a validation run. As a result, the validator continues to serve stale data (if previously started) or, on initial startup, fails to serve any data at all [1][3]. This effectively denies service to Relying Parties that depend on up-to-date RPKI data for routing decisions.

Mitigation

NLnet Labs released Routinator version 0.10.2 on 2021-11-09, which introduces a configurable max-ca-depth variable that limits the CA chain length and prevents infinite processing [2][3]. Users should upgrade to 0.10.2 or later. No workarounds exist for prior versions. This CVE is not listed on the CISA KEV catalog.