RE: [EXTERNAL] Improving ND security

["Templin (US), Fred L" <[e-mail]>]

Hi, about SEND I was just doing a maintenance update of the AERO draft and noticed
that the subject is already covered there in full detail in Section 3.22:

https://datatracker.ietf.org/doc/draft-templin-intarea-6706bis/

So, for those who have participated in this discussion, please refer to the AERO spec
and realize that the only thing missing in OMNI is an appropriate reference citation
to AERO. However, a point that could be considered by the OMNI design team is
whether the SEND text in AERO should instead be moved into the OMNI draft.

Please review and comment,

Fred

> -----Original Message-----
> From: ipv6 [mailto:[e-mail]] On Behalf Of Michael Richardson
> Sent: Monday, August 03, 2020 5:39 PM
> To: Ted Lemon <[e-mail]>; v6ops list <[e-mail]>; 6man <[e-mail]>
> Subject: Re: [EXTERNAL] Improving ND security
> 
> 
> Ted Lemon <[e-mail]> wrote:
>     > On Jul 31, 2020, at 1:17 PM, Templin (US), Fred L <[e-mail]> wrote:
>     >> Ted, I think SEND would solve the neighbor cache resource exhaustion attack since a NCE is
>     >> only created on receipt of an authentic (SEND-protected) IPv6 ND message.
> 
>     > OK, but an attacker can just generate a bazillion key pairs, right? So
>     > it can still exhaust the cache, at the cost of some computation work.
> 
> If you do a TUFO trust model, yes.
> If you do something else, because you are in a different environment than a
> IETF meeting network^W^W^Wcoffee shop, then the potential win is quite large.
> 
> Even if you do a TOFU trust model, keys which were first used prior to the
> attack, could continue to be be used, protecting the cache.
> 
> This pushes the exhaustion attack from one part of the control plane (where
> ND messages are processed), to another part of the control plane (where keys
> are trusted on first use).
> Is that a good thing?  Maybe.
> 
> --
> Michael Richardson <[e-mail]>, Sandelman Software Works
>  -= IPv6 IoT consulting =-
> 
>