Re: [tram] Eric Rescorla's Discuss on draft-ietf-tram-stunbis-16: (with DISCUSS and COMMENT)

[Gonzalo Camarillo <[e-mail]>]

Authors of stunbis (Marc and Gonzalo S),

could you please respond to this? Thanks!

Cheers,

Gonzalo

On 27-Dec-18 03:05, Eric Rescorla wrote:
> We really need to bring this to a close before both Spencer and I step
> down, I should think.
> 
> -Ekr
> 
> 
> On Thu, Nov 22, 2018 at 4:30 AM Gonzalo Camarillo
> <[e-mail] <mailto:[e-mail]>>
> wrote:
> 
>     Eric, Spencer, thanks for the status update!
> 
>     Authors, could you please let us know when you plan to get to this?
> 
>     Gonzalo
> 
>     On 19-Nov-18 23:04, Spencer Dawkins at IETF wrote:
>     > Hi, Gonzalo, 
>     > On Mon, Nov 19, 2018 at 9:06 AM Eric Rescorla <[e-mail]
>     <mailto:[e-mail]>
>     > <mailto:[e-mail] <mailto:[e-mail]>>> wrote:
>     >
>     >     Not really. I have not received any response to my mail of Oct 23.
>     >     As before, I'm happy to have a call, but I believe we're reaching
>     >     the limits of what can be accomplished by email.
>     >
>     >
>     > Eric would know best (it being his Discuss), but this is also my
>     > understanding. 
>     >
>     > I can add this to the agenda of the IESG informal telechat on November
>     > 29, if there hasn't been a call before then, but I don't have a reason
>     > to wait until then, if it's possible to talk sooner.
>     >
>     > Thanks,
>     >
>     > Spencer
>     >  
>     >
>     >
>     >     -Ekr
>     >
>     >
>     >     On Mon, Nov 19, 2018 at 6:35 AM Gonzalo Camarillo
>     >     <[e-mail]
>     <mailto:[e-mail]>
>     >     <mailto:[e-mail]
>     <mailto:[e-mail]>>> wrote:
>     >
>     >         Hi Spencer,
>     >
>     >         what is the status of this? Are the authors and the document
>     >         shepherd
>     >         working with the relevant ADs on the discusses?
>     >
>     >       
>      https://datatracker.ietf.org/doc/draft-ietf-tram-stun-pmtud/ballot/
>     >
>     >         Cheers,
>     >
>     >         Gonzalo
>     >
>     >         On 24-Oct-18 16:40, Spencer Dawkins at IETF wrote:
>     >         > Hi, Marc, 
>     >         >
>     >         > On Wed, Oct 24, 2018 at 3:44 AM Marc Petit-Huguenin
>     >         <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         > <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>> wrote:
>     >         >
>     >         >     Hi Spencer,
>     >         >
>     >         >     I sent my answers to Eric Rescorla questions on
>     2018-10-07
>     >         following
>     >         >     an in-person meeting with my co-author, but never got a
>     >         response
>     >         >     back.  Because there was no change proposed by Eric
>     I went
>     >         ahead and
>     >         >     published -19 a couple of weeks after that, with the
>     text
>     >         agreed in
>     >         >     response to Adam's and Benjamin's comments.
>     >         >
>     >         >   
>     >       
>       https://www.ietf.org/mail-archive/web/tram/current/msg02635.html
>     >         >
>     >         >
>     >         > Ah - I wonder if that was what had happened. 
>     >         >
>     >         > It sounds like you did the right thing, and that Eric
>     has now
>     >         responded,
>     >         > which is also the right thing to do. 
>     >         >
>     >         > Thanks for helping me understand.
>     >         >
>     >         > Spencer
>     >         >  
>     >         >
>     >         >     On 10/23/18 7:28 PM, Spencer Dawkins at IETF wrote:
>     >         >     > Hi, Marc,
>     >         >     >
>     >         >     > I see that a -19 has been submitted, but didn't see a
>     >         reply from
>     >         >     Eric in
>     >         >     > this thread. Do you think that you've converged?
>     >         >     >
>     >         >     > (I saw an offer of a conference call, so thought an
>     >         out-of-band
>     >         >     > conversation might have happened)
>     >         >     >
>     >         >     > Thanks,
>     >         >     >
>     >         >     > Spencer
>     >         >     >
>     >         >     > On Sun, Oct 7, 2018 at 9:35 AM Marc Petit-Huguenin
>     >         >     <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>> wrote:
>     >         >     >
>     >         >     >> Hi Eric,
>     >         >     >>
>     >         >     >> Please see inline.
>     >         >     >>
>     >         >     >> On 09/10/2018 03:25 AM, Eric Rescorla wrote:
>     >         >     >>> On Sat, Sep 8, 2018 at 2:31 PM, Marc Petit-Huguenin
>     >         >     <[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>>
>     >         >     >>> wrote:
>     >         >     >>>
>     >         >     >>>> Hi Eric,
>     >         >     >>>>
>     >         >     >>>> Apologies for the delay in getting back to that.
>     >         >     >>>>
>     >         >     >>>> I think that there is some misunderstanding in what
>     >         STUNbis is
>     >         >     trying to
>     >         >     >>>> do, so please see my comments inline.
>     >         >     >>>>
>     >         >     >>>> On 06/18/2018 10:43 AM, Eric Rescorla wrote:
>     >         >     >>>>> Hi folks,
>     >         >     >>>>>
>     >         >     >>>>> I've reviewed the new version, but I don't think
>     >         that the biddown
>     >         >     >>>>> discussion makes much sense.
>     >         >     >>>>>
>     >         >     >>>>> To recap, there are two hashes here:
>     >         >     >>>>>
>     >         >     >>>>> - The hash which you use to store the password
>     with
>     >         (currently
>     >         >     mostly
>     >         >     >>>> MD5)
>     >         >     >>>>> - The hash you use to compute the MAC
>     (currently SHA-1).
>     >         >     >>>>>
>     >         >     >>>>> First, let's stipulate that MD5 isn't a great
>     choice
>     >         here,
>     >         >     though SHA-1
>     >         >     >>>>> isn't a great choice
>     >         >     >>>>> either for pwd hashing You want Argon or the like.
>     >         With that said,
>     >         >     >>>> there's
>     >         >     >>>>> no sensible
>     >         >     >>>>> biddown attack on that hash because it's a
>     >         per-server feature,
>     >         >     not a
>     >         >     >>>>> per-transaction
>     >         >     >>>>> feature. So, as long as the server has MD5-hashed
>     >         passwords, the
>     >         >     >>>> situation
>     >         >     >>>>> is bad.
>     >         >     >>>>
>     >         >     >>>> In no place in STUNbis we are proposing to use
>     SHA-1
>     >         for password
>     >         >     >>>> encryption, so I am not sure where that come from. 
>     >         What we
>     >         >     propose is:
>     >         >     >>>>
>     >         >     >>>
>     >         >     >>> You're right, it's SHA-256, but my criticisms apply
>     >         equally there.
>     >         >     >>
>     >         >     >> SHA-256 was what the WG adopted.  Our attempts to add
>     >         other passwords
>     >         >     >> encryption mechanisms were denied.  It is true that
>     >         Argon2 was
>     >         >     not in that
>     >         >     >> list (in our defense Argon2 was not known before July
>     >         2015), but
>     >         >     I do not
>     >         >     >> see why the WG would have accepted that one over the
>     >         others. 
>     >         >     Anyway it is
>     >         >     >> too late to fix this, as it is my understanding that
>     >         the WG does
>     >         >     not have
>     >         >     >> enough energy to reach consensus on a new password
>     >         algorithm. 
>     >         >     Someone can
>     >         >     >> just write a draft adding Argon2 as password
>     >         encryption, as we
>     >         >     will have a
>     >         >     >> IANA registry for that.
>     >         >     >>
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>> - Establish a registry for new password algorithms
>     >         (section
>     >         >     18.5), so
>     >         >     >>>> algorithms like Argon2 could be added later
>     (but note
>     >         that our own
>     >         >     >>>> proposals to add more password algorithms were
>     >         rejected by the
>     >         >     working
>     >         >     >>>> group).
>     >         >     >>>> - Add a new password algorithm to that registry,
>     >         namely SHA-256.
>     >         >     >>>> - Register MD5 as an initial password algorithm for
>     >         backward
>     >         >     >> compatibility
>     >         >     >>>> purpose.
>     >         >     >>>>
>     >         >     >>>> As for the biddown protection itself, it is my
>     >         recollection that it
>     >         >     >>>> happened more or less like that:
>     >         >     >>>>
>     >         >     >>>>
>     >         >     >>>> INT. MEETING ROOM - DAY
>     >         >     >>>>
>     >         >     >>>> One of the co-editors of STUNBis stands at the
>     >         microphone:
>     >         >     >>>>
>     >         >     >>>>                            CO-EDITOR
>     >         >     >>>>                  We added SHA-256 protection for
>     >         passwords
>     >         >     >>>>                  in STUNBis.
>     >         >     >>>>
>     >         >     >>>>                            SOMEONE (V.O.)
>     >         >     >>>>                  As MD5 still need to be supported,
>     >         you need to add
>     >         >     >>>>                  protection for bid-down attacks.
>     >         >     >>>>
>     >         >     >>>> CLOSE-UP on CO-EDITOR ROLLING HIS EYES
>     >         >     >>>>
>     >         >     >>>>                            CO-EDITOR
>     >         >     >>>>                  OK, I'll work on that.
>     >         >     >>>>
>     >         >     >>>> Four to eight months has passed.
>     >         >     >>>>
>     >         >     >>>> INT. ANOTHER MEETING ROOM - DAY
>     >         >     >>>>
>     >         >     >>>> The same co-editor of STUNBis stands at the
>     microphone:
>     >         >     >>>>
>     >         >     >>>>                            CO-EDITOR
>     >         >     >>>>                  We added a nice mechanism to
>     prevent
>     >         bid-down
>     >         >     >>>>                  attacks on passwords.  Any
>     comments?
>     >         >     >>>>
>     >         >     >>>>                            THE ROOM
>     >         >     >>>>                  (silence)
>     >         >     >>>>
>     >         >     >>>>                            CO-EDITOR
>     >         >     >>>>                  Moving on...
>     >         >     >>>>
>     >         >     >>>
>     >         >     >>> I don't see how any of this is relevant to my
>     >         technical points
>     >         >     above.
>     >         >     >>
>     >         >     >> My point was that we, the co-editors, did not
>     decide on
>     >         adding
>     >         >     bid-down
>     >         >     >> protection, someone asked us to do so and no-one
>     in the
>     >         WG saw
>     >         >     any problem
>     >         >     >> with that.  The reasons that person wanted that
>     are not
>     >         known to
>     >         >     us but, as
>     >         >     >> you insist, here's one reason I can think of:
>     >         >     >>
>     >         >     >> It is a fact that, for operational reasons, a
>     password
>     >         database
>     >         >     cannot be
>     >         >     >> re-encrypted at once.  Also for operational reasons,
>     >         the MD5 password
>     >         >     >> cannot be immediately removed from the database
>     as soon
>     >         the user
>     >         >     submitted
>     >         >     >> a new one.  In fact, and for quite some time, both
>     >         encrypted
>     >         >     variants of
>     >         >     >> the same password may have to be kept in that
>     database,
>     >         because a
>     >         >     single
>     >         >     >> user may use a mix of devices, some of them that
>     use an
>     >         RFC 5389
>     >         >     client,
>     >         >     >> some that use a STUNbis client.  It is up to the STUN
>     >         server owner to
>     >         >     >> decide how long the migration to STUNbis will
>     take and
>     >         when it
>     >         >     will be
>     >         >     >> acceptable to reject all RFC 5389 (i.e. MD5)
>     clients (that
>     >         >     migration time
>     >         >     >> can be purposely reduced to 0 seconds but that's the
>     >         choice and
>     >         >     >> responsibility of the owner of the server).
>     >         >     >>
>     >         >     >> Meanwhile we still need to be sure that if the STUN
>     >         client is
>     >         >     implementing
>     >         >     >> STUNbis it unconditionally gets the additional
>     >         protection of the new
>     >         >     >> password encryption algorithm.  That's where the
>     biddown
>     >         >     protection kicks
>     >         >     >> in, by preventing an online attacker to have the
>     server
>     >         >     misidentifying a
>     >         >     >> STUNbis client as an RFC 5389 client, by
>     preventing an
>     >         online
>     >         >     attacker to
>     >         >     >> have the client misidentifying a STUNbis server as a
>     >         RFC 5389
>     >         >     server, and
>     >         >     >> having both them use the MD5 encrypted password
>     instead
>     >         of the
>     >         >     SHA-256
>     >         >     >> encrypted password, all of that easily done by
>     >         stripping the
>     >         >     unprotected
>     >         >     >> 401 response of the new STUNbis PASSWORD-ALGORITHMS
>     >         attribute.
>     >         >     >>
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>>> The second topic is the hash used to compute the
>     >         MAC. However,
>     >         >     I don't
>     >         >     >>>> see
>     >         >     >>>>> how
>     >         >     >>>>> this gives you sensible biddown protection because
>     >         that hash
>     >         >     is also
>     >         >     >> used
>     >         >     >>>>> to compute
>     >         >     >>>>> MAC over the negotiation: an attacker who has
>     >         compromised a
>     >         >     MAC which
>     >         >     >> the
>     >         >     >>>>> server
>     >         >     >>>>> supports will quite likely be able to forge a MAC
>     >         over the
>     >         >     transcript
>     >         >     >> as
>     >         >     >>>>> well. This is,
>     >         >     >>>>> I suppose, potentially useful as a defense against
>     >         some other
>     >         >     weakness
>     >         >     >>>>> (e.g.,
>     >         >     >>>>> version #), but I don't really see how the current
>     >         design
>     >         >     helps against
>     >         >     >>>>> attacks on the
>     >         >     >>>>> MAC.
>     >         >     >>>>
>     >         >     >>>> There is no biddown attack protection for the
>     MAC, as
>     >         stated in
>     >         >     Section
>     >         >     >>>> 16.3:
>     >         >     >>>>
>     >         >     >>>> "The bid-down protection mechanism described in
>     this
>     >         document
>     >         >     is new,
>     >         >     >>>>  and thus cannot currently protect against a
>     bid-down
>     >         attack that
>     >         >     >>>>  lowers the strength of the hash algorithm to
>     HMAC-SHA1."
>     >         >     >>>>
>     >         >     >>>> What we put in place is a plan for *future*
>     versions
>     >         of STUN to get
>     >         >     >>>> biddown protection for the MAC.  That's it, no
>     more,
>     >         no less.
>     >         >     >>>>
>     >         >     >>>
>     >         >     >>> Yes, but I don't believe that this will provide
>     bid-down
>     >         >     protection for
>     >         >     >> the
>     >         >     >>> MAC in the future for the reasons I indicate above.
>     >         >     >>>
>     >         >     >>> If you think this does something useful, please
>     show me an
>     >         >     example attack
>     >         >     >>> and how this fixes it. Note that it's not generally
>     >         useful to
>     >         >     just bid
>     >         >     >> down
>     >         >     >>> the MAC itself unless the MAC you bid down to is
>     weak
>     >         enough to
>     >         >     exploit
>     >         >     >> in
>     >         >     >>> some other way.
>     >         >     >>
>     >         >     >> I do not know what weaknesses will be discovered
>     in the
>     >         future. 
>     >         >     I am also
>     >         >     >> pretty sure that the cost of not using that mechanism
>     >         is very
>     >         >     close to 0.
>     >         >     >> What I am sure of is that the cost of reengineering a
>     >         new biddown
>     >         >     >> protection mechanism if we ever need it will be
>     high. 
>     >         We already
>     >         >     went
>     >         >     >> through the pains of designing one for the password
>     >         algorithm, so
>     >         >     why not
>     >         >     >> extend it so it can be used in the aftermath of the
>     >         next Snowden
>     >         >     facepalm
>     >         >     >> moment?
>     >         >     >>
>     >         >     >>> Again, happy to have a call to walk though this
>     if that
>     >         >     >>> helps.
>     >         >     >>>
>     >         >     >>> -Ekr
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>
>     >         >     >>>>>
>     >         >     >>>>> You might think that there was a MAC which was
>     easier to
>     >         >     reverse to
>     >         >     >> find
>     >         >     >>>>> the original
>     >         >     >>>>> password, but the defense you have here
>     doesn't help
>     >         with that
>     >         >     because
>     >         >     >>>> the
>     >         >     >>>>> on-path attacker can do a bid-down and use the
>     >         client as a MAC
>     >         >     oracle
>     >         >     >> for
>     >         >     >>>>> any MAC the client supports.
>     >         >     >>>>>
>     >         >     >>>>> So, I still don't understand what this is supposed
>     >         to do. Happy to
>     >         >     >> have a
>     >         >     >>>>> call if you
>     >         >     >>>>> think that helps
>     >         >     >>>>>
>     >         >     >>>>> -Ekr
>     >         >     >>>>>
>     >         >     >>>>>
>     >         >     >>>>>
>     >         >     >>>>>
>     >         >     >>>>> On Mon, Jun 18, 2018 at 3:40 AM, Gonzalo
>     Camarillo <
>     >         >     >>>>> [e-mail]
>     <mailto:[e-mail]>
>     >         <mailto:[e-mail]
>     <mailto:[e-mail]>>
>     >         >     <mailto:[e-mail]
>     <mailto:[e-mail]>
>     >         <mailto:[e-mail]
>     <mailto:[e-mail]>>>> wrote:
>     >         >     >>>>>
>     >         >     >>>>>> Marc, Eric,
>     >         >     >>>>>>
>     >         >     >>>>>> what is the status of this discussion?
>     >         >     >>>>>>
>     >         >     >>>>>> Thanks,
>     >         >     >>>>>>
>     >         >     >>>>>> Gonzalo
>     >         >     >>>>>>
>     >         >     >>>>>> On 04/05/2018 2:35 AM, Eric Rescorla wrote:
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> On Mon, Apr 23, 2018 at 11:16 AM, Marc
>     >         Petit-Huguenin <
>     >         >     >>>> [e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>
>     >         >     >>>>>>> <mailto:[e-mail]
>     <mailto:[e-mail]> <mailto:[e-mail]
>     <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>>> wrote:
>     >         >     >>>>>>>
>     >         >     >>>>>>>     On 04/22/2018 05:22 PM, Eric Rescorla wrote:
>     >         >     >>>>>>>     > On Sun, Apr 22, 2018 at 2:02 PM, Marc
>     >         Petit-Huguenin <
>     >         >     >>>>>> [e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>
>     >         >     <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>
>     >         <mailto:[e-mail] <mailto:[e-mail]>
>     <mailto:[e-mail] <mailto:[e-mail]>>>>>
>     >         >     >>>>>>>     > wrote:
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >>>>      For a request or indication
>     message,
>     >         the agent
>     >         >     MUST
>     >         >     >>>>>> include the
>     >         >     >>>>>>>     >>>>      USERNAME,
>     MESSAGE-INTEGRITY-SHA256, and
>     >         >     >> MESSAGE-INTEGRITY
>     >         >     >>>>>>>     >> attributes
>     >         >     >>>>>>>     >>>>      in the message unless the agent
>     >         knows from an
>     >         >     external
>     >         >     >>>>>> indication
>     >         >     >>>>>>>     >>>>      which message integrity
>     algorithm is
>     >         supported
>     >         >     by both
>     >         >     >>>>>> agents.  In
>     >         >     >>>>>>>     >>>>      this case either
>     MESSAGE-INTEGRITY or
>     >         >     >>>>>> MESSAGE-INTEGRITY-SHA256 MUST
>     >         >     >>>>>>>     >>>>      be included in addition to
>     >         USERNAME.  The HMAC
>     >         >     for the
>     >         >     >>>>>> MESSAGE-
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>> This text appears to conflict with S
>     7.3 of
>     >         >     5245-bis, which
>     >         >     >>>> says:
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> [text missing]
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> Hmm, no, RFC245bis is still referencing
>     >         RFC5389, so the
>     >         >     >>>> procedure
>     >         >     >>>>>> for
>     >         >     >>>>>>>     >> stunbis does not apply.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     > I hear what you're saying, but
>     publishing two
>     >         >     documents at the
>     >         >     >>>>>> same time
>     >         >     >>>>>>>     > which
>     >         >     >>>>>>>     > make contrary recommendations about
>     the same
>     >         basic
>     >         >     protocol is
>     >         >     >>>>>> un-good.
>     >         >     >>>>>>>
>     >         >     >>>>>>>     Sure, but wouldn't it be simpler to have
>     >         rfc5245bis
>     >         >     using stunbis
>     >         >     >>>>>>>     and have them updating their text, more than
>     >         adding some
>     >         >     tortuous
>     >         >     >>>>>>>     text in stunbis?
>     >         >     >>>>>>>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >>>      The STUN usage must specify which
>     >         transport
>     >         >     protocol is
>     >         >     >>>>>> used, and
>     >         >     >>>>>>>     >> how
>     >         >     >>>>>>>     >>>>      the agent determines the IP
>     address
>     >         and port
>     >         >     of the
>     >         >     >>>>>> recipient.
>     >         >     >>>>>>>     >>>>      Section 8 describes a DNS-based
>     >         method of
>     >         >     determining
>     >         >     >> the
>     >         >     >>>>>> IP
>     >         >     >>>>>>>     >> address
>     >         >     >>>>>>>     >>>>      and port of a server that a usage
>     >         may elect to
>     >         >     use.
>     >         >     >> STUN
>     >         >     >>>>>> may be
>     >         >     >>>>>>>     >> used
>     >         >     >>>>>>>     >>>>      with anycast addresses, but only
>     >         with UDP and
>     >         >     in usages
>     >         >     >>>>>> where
>     >         >     >>>>>>>     >>>>      authentication is not used.
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>> Why this restriction? You should be able
>     >         to use
>     >         >     anycast with
>     >         >     >>>>>> long-term
>     >         >     >>>>>>>     >>> AUTH for (say) TURN.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >>
>     >         https://www.ietf.org/mail-archive/web/behave/current/
>     >         >     >>>>>> msg03582.html
>     >         >     >>>>>>>   
>     >          <https://www.ietf.org/mail-archive/web/behave/current/
>     >         >     >>>> msg03582.html>
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> I think that the decision of
>     permitting Anycast
>     >         >     should be left
>     >         >     >>>> to
>     >         >     >>>>>> each
>     >         >     >>>>>>>     >> STUN Usage.  Basic STUN Usage does
>     not use
>     >         >     authentication and
>     >         >     >>>> use
>     >         >     >>>>>> only a
>     >         >     >>>>>>>     >> one round trip for the Binding
>     transaction, so
>     >         >     Unicast can be
>     >         >     >>>>>> used.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >> OTOH, TURN and ICE should probably say
>     >         something
>     >         >     about that,
>     >         >     >> so
>     >         >     >>>> I
>     >         >     >>>>>> added a
>     >         >     >>>>>>>     >> new bullet point in Section 13:
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >>    o  If Anycast addresses can be
>     used for
>     >         the server
>     >         >     in case
>     >         >     >>>> TCP
>     >         >     >>>>>> or
>     >         >     >>>>>>>     >>       TLS-over-TCP, or authentication
>     are used.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     > Are you leaving this text in? That seems
>     >         very confusing.
>     >         >     >>>>>>>
>     >         >     >>>>>>>     In isolation yes, but I think it makes sense
>     >         which the text
>     >         >     >> before
>     >         >     >>>>>>>     the bullet points:
>     >         >     >>>>>>>
>     >         >     >>>>>>>        A STUN usage defines how STUN is actually
>     >         utilized --
>     >         >     when to
>     >         >     >>>> send
>     >         >     >>>>>>>        requests, what to do with the responses,
>     >         and which
>     >         >     optional
>     >         >     >>>>>>>        procedures defined here (or in an
>     extension
>     >         to STUN)
>     >         >     are to be
>     >         >     >>>>>> used.
>     >         >     >>>>>>>        A usage also defines:
>     >         >     >>>>>>>
>     >         >     >>>>>>>     [...]
>     >         >     >>>>>>>
>     >         >     >>>>>>>        o  If Anycast addresses can be used
>     for the
>     >         server in
>     >         >     case TCP
>     >         >     >>>> or
>     >         >     >>>>>>>           TLS-over-TCP, or authentication
>     are used.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> What is the need for the restriction at all.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>     >>>>      transaction over UDP or
>     >         DTLS-over-UDP is also
>     >         >     >> considered
>     >         >     >>>>>> failed if
>     >         >     >>>>>>>     >>>>      there has been a hard ICMP error
>     >         [RFC1122].  For
>     >         >     >> example,
>     >         >     >>>>>> assuming
>     >         >     >>>>>>>     >> an
>     >         >     >>>>>>>     >>>>      RTO of 500ms, requests would
>     be sent
>     >         at times
>     >         >     0 ms, 500
>     >         >     >>>>>> ms, 1500
>     >         >     >>>>>>>     >> ms,
>     >         >     >>>>>>>     >>>>      3500 ms, 7500 ms, 15500 ms, and
>     >         31500 ms.  If the
>     >         >     >> client
>     >         >     >>>>>> has not
>     >         >     >>>>>>>     >>>>      received a response after
>     39500 ms,
>     >         the client
>     >         >     will
>     >         >     >>>>>> consider the
>     >         >     >>>>>>>     >>>>      transaction to have timed out.
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>> I note that these recommendations
>     now seem
>     >         crazily
>     >         >     long. I
>     >         >     >>>>>> assume the
>     >         >     >>>>>>>     >>> WG had consensus on this, but I
>     wanted to
>     >         note it.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> Not just the WG, also the IESG that
>     >         approved RFC 5389
>     >         >     too as,
>     >         >     >>>> but
>     >         >     >>>>>> for the
>     >         >     >>>>>>>     >> addition of "or DTLS-over-UDP", this
>     is the
>     >         same text
>     >         >     than in
>     >         >     >>>> RFC
>     >         >     >>>>>> 5389.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     > Yes, I know. My point is that while they
>     >         might have bee
>     >         >     >> sensible
>     >         >     >>>>>> when
>     >         >     >>>>>>>     > 5389 was published they *now* seem crazily
>     >         long. Did
>     >         >     the WG
>     >         >     >>>>>> explicitly
>     >         >     >>>>>>>     > decide not to update them?
>     >         >     >>>>>>>
>     >         >     >>>>>>>     No, nobody ever suggested that there was an
>     >         issue there.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> OK, well, this seems like it should be
>     considered,
>     >         then,
>     >         >     because this
>     >         >     >>>>>>> doesn't
>     >         >     >>>>>>> match modern practice.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>     > This would be good to explain.
>     >         >     >>>>>>>
>     >         >     >>>>>>>     New text:
>     >         >     >>>>>>>
>     >         >     >>>>>>>        [...]
>     >         >     >>>>>>>        containing the subjectAltName of that
>     >         certificate. 
>     >         >     The test
>     >         >     >> on
>     >         >     >>>>>> the
>     >         >     >>>>>>>        MESSAGE-INTEGRITY-SHA256 attribute
>     >         indicates that the
>     >         >     >>>> transaction
>     >         >     >>>>>> is
>     >         >     >>>>>>>        authenticated and that the client
>     >         implements this
>     >         >     >> specification
>     >         >     >>>>>> and
>     >         >     >>>>>>>        so can process the ALTERNATE-DOMAIN
>     attribute.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> All right.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     >>>>      o  What authentication and
>     >         message-integrity
>     >         >     mechanisms
>     >         >     >>>>>> are used.
>     >         >     >>>>>>>     >>>>
>     >         >     >>>>>>>     >>>>      o  The considerations around
>     manual vs.
>     >         >     automatic key
>     >         >     >>>>>> derivation
>     >         >     >>>>>>>     >> for
>     >         >     >>>>>>>     >>>>         the integrity mechanism, as
>     >         discussed in
>     >         >     [RFC4107].
>     >         >     >>>>>>>     >>>>
>     >         >     >>>>>>>     >>>>      o  What mechanisms are used to
>     >         distinguish STUN
>     >         >     >> messages
>     >         >     >>>>>> from other
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>> Why is this required? It seems like
>     that's
>     >         a generic
>     >         >     STUN
>     >         >     >>>>>> feature.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> That text is identical to the text in RFC
>     >         5389.  RFC
>     >         >     5764/7983
>     >         >     >>>> is
>     >         >     >>>>>> one such
>     >         >     >>>>>>>     >> mechanism, but there is nothing that
>     >         prevent another
>     >         >     protocol
>     >         >     >>>>>> stack to use
>     >         >     >>>>>>>     >> a different mechanism (inference,
>     shim, etc...)
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     > But ultimately no matter what the
>     other protocol
>     >         >     provides for
>     >         >     >>>>>> demux, STUN
>     >         >     >>>>>>>     > has its
>     >         >     >>>>>>>     > own demux.
>     >         >     >>>>>>>
>     >         >     >>>>>>>     In fact I think that the reason for that
>     item
>     >         was because
>     >         >     >>>>>>>     FINGERPRINT can also be used to demux STUN
>     >         traffic, but
>     >         >     it is
>     >         >     >>>>>>>     optional.  So an STUN Usage needs to tell if
>     >         FINGERPRINT is
>     >         >     >>>>>>>     mandatory (like in ICE).
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> This should be explained.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>>>      that is not readily subject to
>     >         offline dictionary
>     >         >     >>>> attacks.
>     >         >     >>>>>>>     >>>>      Protection of the channel itself,
>     >         using TLS or
>     >         >     DTLS,
>     >         >     >>>>>> mitigates
>     >         >     >>>>>>>     >> these
>     >         >     >>>>>>>     >>>>      attacks.
>     >         >     >>>>>>>     >>>>
>     >         >     >>>>>>>     >>>>      STUN supports both
>     MESSAGE-INTEGRITY and
>     >         >     >>>>>>>     MESSAGE-INTEGRITY-SHA256,
>     >         >     >>>>>>>     >>>>      which is subject to bid down
>     attacks
>     >         by an on-path
>     >         >     >>>>>> attacker.
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>> By an on-path attacker who can forge
>     >         HMAC-SHA1 in
>     >         >     real-time?
>     >         >     >>>>>>>     That's a
>     >         >     >>>>>>>     >>> pretty serious adversary, so you should
>     >         clarify here
>     >         >     >>>>>>>     >>>
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >> New text:
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >>    STUN supports both
>     MESSAGE-INTEGRITY and
>     >         >     >>>>>> MESSAGE-INTEGRITY-SHA256,
>     >         >     >>>>>>>     >>    which is subject to bid down
>     attacks by
>     >         an on-path
>     >         >     attacker
>     >         >     >>>>>> that
>     >         >     >>>>>>>     >>    would strip the
>     MESSAGE-INTEGRITY-SHA256
>     >         attribute
>     >         >     leaving
>     >         >     >>>>>>>     only the
>     >         >     >>>>>>>     >>    MESSAGE-INTEGRITY attribute and
>     >         exploiting a potential
>     >         >     >>>>>>>     vulnerability.
>     >         >     >>>>>>>     >>    Protection of the channel itself,
>     using
>     >         TLS or DTLS,
>     >         >     >>>> mitigates
>     >         >     >>>>>>>     these
>     >         >     >>>>>>>     >>    attacks.  Timely removal of the
>     support of
>     >         >     >> MESSAGE-INTEGRITY
>     >         >     >>>>>> in a
>     >         >     >>>>>>>     >>    future version of STUN is necessary.
>     >         >     >>>>>>>     >>
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>     > I still don't understand the capabilities
>     >         you seem to
>     >         >     believe
>     >         >     >> the
>     >         >     >>>>>>>     attacker
>     >         >     >>>>>>>     > has.
>     >         >     >>>>>>>     > Can you describe the exact attack.
>     >         >     >>>>>>>     >
>     >         >     >>>>>>>
>     >         >     >>>>>>>     1. Vulnerability is found in HMAC-SHA1
>     >         >     >>>>>>>     2. Client Alice still supports M-I and
>     >         M-I-256, does not
>     >         >     know
>     >         >     >> what
>     >         >     >>>>>>>     version of STUN server Bob supports and so
>     >         send both.
>     >         >     >>>>>>>     3. On-path attacker removes M-I-256.
>     >         >     >>>>>>>     5. stunbis server Bob thinks that Alice
>     is an
>     >         RFC 5389
>     >         >     client and
>     >         >     >>>>>>>     continue with that protocol.
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>> This seems like an extremely weak attack. In
>     >         general, any
>     >         >     protocol of
>     >         >     >>>>>>> this type is as strong
>     >         >     >>>>>>> as the weakest integrity algorithm it
>     supports, so
>     >         it's not
>     >         >     that the
>     >         >     >>>>>>> protocol has a downgrade
>     >         >     >>>>>>> attack, but rather that the minimum algorithm
>     >         supported is
>     >         >     one you
>     >         >     >>>> don't
>     >         >     >>>>>>> trust as much
>     >         >     >>>>>>> as you might like.
>     >         >     >>>>>>>
>     >         >     >>>>>>> -Ekr
>     >         >     >>>>>>>
>     >         >     >>>>>>>
>     >         >     >>>>>>>     --
>     >         >     >>
>     >         >
>     >         >
>     >         >
>     >         >     --
>     >         >     Marc Petit-Huguenin
>     >         >     Email: [e-mail]
>     <mailto:[e-mail]>
>     >         <mailto:[e-mail]
>     <mailto:[e-mail]>> <mailto:[e-mail]
>     <mailto:[e-mail]>
>     >         <mailto:[e-mail]
>     <mailto:[e-mail]>>>
>     >         >     Blog: https://marc.petit-huguenin.org
>     >         >     Profile: https://www.linkedin.com/in/petithug
>     >         >
>     >
>