There is still a problem in this document that needs to be fixed. Section 3.1 of the document explicitly states that there is no requirement for subject names to be globally unique in the RPKI. As I have pointed out several times before, this is a violation of X.509, which says that "CAs are unambiguously defined by a distinguished name" and that a directory name is "A construct that singles out a particular object from all other objects. A name shall be unambiguous (that is, denote just one object); however, it need not be unique (that is, be the only name which unambiguously denotes the object)."

While some people have expressed a concern that a few issuers may create name collisions even if there is a requirement that names be unambiguous, no one has provided an argument as to why such a requirement cannot be specified in the SIDR documents. Name collisions are a security risk, and so should be avoided to whatever degree is possible. While one may not be able to prevent malicious (or incorrectly designed) issuers from creating collisions, one can easily instruct those who are following the specifications to not create them.

The problem of name collisions is further compounded by the fact that draft-ietf-sidr-cp-10.txt (and other SIDR WC documents) explicitly permits name collisions, but provides no warnings about the risks/problems that name collisions create. If two CAs issue certificates and CRLs under the same name then an X.509 compliant path validation library may accept certificates even though they have been revoked or may reject (as revoked) certificates that are valid. Yet, there are no warnings in any of the SIDR WG documents about the security risks created by name collisions. In fact, the documents do just the opposite: they provide (incorrect) rationalizations for why there is no need to avoid name collisions in the RPKI.

draft-ietf-sidr-cp needs to require subject names to unambiguously identify a single entity in the RPKI. There is no reason that such a requirement cannot be imposed. As I demonstrated in a previous email (http://www.ietf.org/mail-archive/web/sidr/current/msg01739.html), this is a requirement that is very easy to implement and does not require "Certification authorities to coordinate PKI-wide unique Subject names." So, a name uniqueness requirement can be imposed just as easily as these requirements:

* The distinguished name for every CA and end entity consists of a single Common Name (CN) attribute with a value generated by the issuer of the certificate. Optionally, the serialNumber attribute MAY be included along with the common name (to form a terminal relative distinguished name set), to distinguish among successive instances of certificates associated with the same entity.

* The Subject name in each certificate MUST be unique relative to all Subject names certified by the same Issuer.

* Each CA operating within the context of this PKI MUST require each Subject to demonstrate proof-of-possession (PoP) of the private key corresponding to the public key in the certificate, prior to issuing the certificate.


As I have stated before, just because the SIDR WG documents require names to be unambiguous (as is required by X.509), that does not preclude draft-ietf-sidr-res-certs from indicating that there is a risk that name collisions will occur.

Dave

_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to