Threats relating to IPv6 multihoming solutions
Summary: Needs a YES.
(Ted Hardie) Discuss
None of these issues are reasons to block this document, but the combination was enough to put this into DISCUSS rather than Comment. After we've had a chance to discuss them, I anticipate shifting to no-ob or yes. In section 2., the draft says: address - an IP layer name that contains both topological significance and acts as a unique identifier for an interface. Is it important that it be a unique identifier? I know of cases where the same IP address is bound both to an ethernet interface and a tunnel interface. Those interfaces are unique at some layer, just not at the IP layer. I'm not sure the distinction is worth raising here, in the terminology section, but I also wasn't sure how important "unique" was to the authors. In section 3.1, the draft says: Applications that use security mechanisms, such as IPsec or TLS, with mutual authentication have the ability to "bind" the FQDN to the cryptographic keying material, thus compromising the DNS and/or the routing system can at worst cause the packets to be dropped or delivered to an entity which does not posses the keying material. The binding isn't always to the FQDN; it can be to an address. Could this be rephrased as: Applications that use mutually authenticating security mechanisms, such as IPSEC or TLS, have the ability to bind an address or FQDN to cryptographic keying material. Compromising the DNS and/or routing system can result in packets being dropped or delivered to an attacker in such cases, but since the attacker does not possess the keying material the application will not trust the attacker. This section also uses "class of applications", where I believe they are talking about "class of security configurations for applications". The same application (or class of applications) could employ each of these. Do the authors consider "leap of faith" trust arrangements to fall into their 5th category? They certainly lack strong identity, but they seem to provide assurance that a particular peer is the same peer over a number of interactions; there may be no assurance as to who that peer is with reference to an external system, but that may not be required. 4.1.2 touches on this, but it wasn't clear to me whether these trust arrangements fell under 5 or were some other category. In Section 7, the draft says: A possible approach that solutions might investigate is to defer verification until there appears to be two different hosts (or two different locators for the same host) that want to use the same identifier. Does this mean simultaneous hosts/locators? I'm concerned that a combination here of a DoS and impersonation could make simultaneous appearance rare: A is talking to B; X DoS'es A while impersonating A to B; B sees a "new A", but no longer has the ability to verifiy the first A.
(Scott Hollenbeck) Discuss
Section 3.1 asks this unresolved question: [TBD: Does one-way authentication, without mutual authentication, add a different class of applications?] This should be cleaned up before IESG review.
Missing IANA Considerations.
(Russ Housley) Discuss
In addition to the TBD already identified by Scott, I have a few other concerns. This document is, perhaps necessarily, somewhat incomplete. It is difficult to do a real analysis without pinning down specific protocol details, which this draft cannot do given the topic it is trying to cover. It includes a general analysis of redirection attacks based on multihoming solutions that allow locator changes. As such, it should point out that additional work is needed, especially in the following areas: 1) If we were to make the split between locators and identifiers, then we probably want to tie most application security mechanisms to the identifier, not to the locator. Therefore, work is needed to understand how attacks on the identifier mechanism affect security, especially, in my mind, attacks on any mechanism(s) that bind locators to identifiers. 2) Does multicast make matters worse? It usually does. 3) Connectionless transport protocols need more attention. They are already difficult to secure, even without a locator/identifier split.
Comments include ones from SecDir review by Rob Austein. The title and abstract are a bit confusing unless one already understands the subject of the document. The document is really pretty specific to issues surrounding locator/identifier split and redirection attacks. The document says: : : identifier - an IP layer identifier for an IP layer endpoint : (stack name in [NSRG]). The transport endpoint is a : function of the transport protocol and would : typically include the IP identifier plus a port : number. There might be use for having multiple : interfaces per stack/per host. : s/interfaces/identifiers/ ? Probably not, but the text is a bit confusing, since the very next sentence says that identifiers aren't associated with an interface. Please clarify. In section 4.1, please be blunt in pointing out the man-in-the- middle potential in all of the redirection attacks.
(David Kessens) Yes
The draft doesn't have an IANA Considerations section, which should be a null one if present. Also Appendix A will need to be removed.