[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: SCTP and IPsec issues


>>>>> "Angelos" == Angelos D Keromytis <angelos@cis.upenn.edu> writes:
    Angelos> The alternative (setting up a number of SAs, for each of the
    Angelos> valid destination addresses) does not work because of SA state
    Angelos> issues (at least the replay counter has to be synchronized
    Angelos> across all these SAs) and the obvious memory waste.

    Angelos> Recommendation: the architecture document should state that SAs
    Angelos> are identified by the extended triplet.

  As an alternative, it would be nice if the architecture said that the
mapping of SPD->SA was an N:1 mapping, then no textual change would be
necessary to support this.

    Angelos> 3) The SCTP source/destination port numbers should be included
    Angelos> as SPD selectors.

    Angelos> Recommendation: include this in the architecture
    Angelos> document. No-brainer.

  But, should the document be rev'ed each time a new protocol comes along
that has a new set of selectors?

    Angelos> 4) The multiple source/destination addresses need to appear in
    Angelos> the Phase 2 ID payloads in IKE. There are two ways of doing
    Angelos> this:

    Angelos> a) Allow multiple source/destination ID payloads in messages 1
    Angelos> and 2 of IKE Quick Mode.  b) Introduce a new type of ID, that
    Angelos> allows for recursive inclusion of other IDs. Thus, Phase 2 would
    Angelos> still only send one ID payload each for source/destination; each

  I strongly favour (b). 
  This is needed for proper per-host keying support for ICMP.
  
  It would also be nice if phase 2 SAs could be referenced in a consistent
way such that additional selectors could be *added* to an existing SA.
  
    Angelos> Recommendation: do (b), as we think it is less intrusive for
    Angelos> implementations than (a). However, recursion is limited to only
    Angelos> one-deep (i.e., you can't have a recursive ID inside a recursive
    Angelos> ID).

  My preference is to define three "recursion" types: AND, OR and NOT.
  Permit at most *three* levels of such logic.

    Angelos> 5) The initiator of an IKE exchange may not know all the remote
    Angelos> host's addresses (used for SCTP), and thus may not be able to
    Angelos> propose an accurate Phase 2 destination ID. Solutions:

    Angelos> a) The Responder may drop the Phase 2 exchange and start a Phase
    Angelos> 2 exchange on its own (using the same Phase 1 SA), using
    Angelos> information from the old exchange, providing the correct set of
    Angelos> addresses. The advantage of this approach is that it requires no
    Angelos> protocol changes (although implementations have to be able to
    Angelos> deal with this initial rejection).  b) The Responder could
    Angelos> modify the Responder ID received in message 1 of Quick Mode, and
    Angelos> send it back in message 2. The Initiator then needs to verify
    Angelos> that this is according to policy (as opposed to doing a simple
    Angelos> bytestring comparison, as is possible today). This requires no
    Angelos> wire changes, but a small change in the processing at the
    Angelos> Initiator. This is a potentially more powerful and lightweight
    Angelos> approach (everything is still done in one exchange, and the
    Angelos> Initiator does not have to deal with a Phase 2 exchange
    Angelos> rejection).

  This begins to sound like negotiation.

] Train travel features AC outlets with no take-off restrictions|gigabit is no[
]   Michael Richardson, Solidum Systems   Oh where, oh where has|problem  with[
]     mcr@solidum.com   www.solidum.com   the little fishy gone?|PAX.port 1100[
] panic("Just another NetBSD/notebook using, kernel hacking, security guy");  [
Follow-Ups: References: