[11:31:36] --- wyllys has joined
[11:31:40] --- wyllys has left
[13:34:42] --- wyllys has joined
[13:34:45] --- wyllys has left
[15:08:50] --- lha has joined
[15:09:26] --- lha has left
[15:12:47] --- wyllys has joined
[15:12:52] --- wyllys has left
[16:46:12] --- krb-wg has joined
[16:51:48] --- bcneuman has joined
[16:58:36] --- warlord has joined
[17:10:07] --- Jeffrey Altman has joined
[17:10:24] <warlord> I hope there's a scribe because I'm in another room.
[17:10:25] --- tlyu has joined
[17:10:25] * Jeffrey Altman has changed the subject to: IETF64 Kerberos WG
[17:10:35] * Jeffrey Altman has changed the subject to: IETF64 Kerberos WG (Cypress Room)
[17:13:18] --- hartmans has joined
[17:14:12] --- kivinen has joined
[17:15:28] --- leifj has joined
[17:15:36] --- lha has joined
[17:15:55] <Jeffrey Altman> I will do my best to scribe
[17:16:32] <warlord> thank you
[17:16:42] <Jeffrey Altman> network is really bad here. beware
[17:16:42] <leifj> scribe
[17:16:46] <leifj> is me
[17:17:00] <leifj> jeff: agenda bashing
[17:17:53] <leifj> I'm assuming people not in the room are looking at slides
[17:18:20] <warlord> the (a) network is just fine.
[17:19:03] <leifj> jeff & nico discusses agenda issues
[17:19:38] <Jeffrey Altman> lzhu withdraws his proposed changes to enctype negotiation. Vista will handle initiator use of PROT_READY
[17:20:34] <leifj> jeff asks about the order of pkinit/sams stuff
[17:20:44] <leifj> sam goes first sais people
[17:20:55] <leifj> sam things his thing also could go on indefinitely
[17:21:01] <leifj> jeff: no changes to the agenda
[17:21:15] <leifj> jeff: any special presentation?
[17:21:42] <leifj> cliff: are you listening to the audio?
[17:21:56] <leifj> asks jeff
[17:22:12] --- wyllys has joined
[17:22:16] <bcneuman> No, I don't have adio, I was looking for the link.
[17:22:25] --- jhutz has joined
[17:23:30] <leifj> Jeffrey Altman is doing a special presentation
[17:24:01] <leifj> a gift of a specially bound and signed copy of 4120 and another present will be given to Cliff at some point
[17:24:17] <leifj> <round of applause for Cliff>
[17:24:23] <jhutz> In thanks and recognition for Cliff's service to the Kerberos community, we are presenting him with a bound copy of RFC4120 (which Jeff showed), and
[17:24:26] <jhutz> an ipod nano.
[17:24:29] <jhutz> There was a round of applause.
[17:24:58] <leifj> jeff: the audio from the session will be preloaded on the ipod nano
[17:24:59] <jhutz> cliff anything to say?
[17:25:02] <bcneuman> I appreciate the presentation. It was a long effort to get the draft done and I am sorry I can't be here this week.
[17:25:27] <leifj> jeff: document status
[17:25:52] <leifj> pkinit - in last call
[17:26:00] <leifj> few comments will be discussed later
[17:26:08] <bcneuman> And though I am absent, I do not intend my participation to stop. There is still much to be done, and the WG has stong leadershop to keep it all going.
[17:26:13] <leifj> larry: list of issues sent to jeff 10 min ago
[17:27:43] <leifj> discussion about the criterion for moving the document forward
[17:28:09] <leifj> jeff: extensions
[17:28:22] <leifj> tom stands up to talk about status of extensions
[17:28:39] <leifj> tom: new rev with suggested (some of them) changes
[17:28:59] <leifj> tom: need to talk about idn's
[17:29:30] <leifj> jeff: enctype neg
[17:29:38] <leifj> larry: add referrals
[17:30:18] <leifj> larry: ken and I have proposal about referrals - give referral draft to ken as a wg item
[17:31:08] <leifj> jeff: larrys anonymous will be talked about later
[17:31:56] <leifj> jeff: talks about an individual request to publish information describing ms rc4
[17:32:21] <leifj> jeff: does anyone oppose having something like that published?
[17:32:33] <leifj> ken: should it be in the form of kcrypto?
[17:32:51] <leifj> sam: comments on that...
[17:32:58] <bcneuman> I am all for publishing the ms rc4 - it documents a large base of what is there.
[17:33:56] <leifj> fyi - I voice-relay stuff people I know not in the room say
[17:34:19] <leifj> sam: presentation - Questioning Kerberos Assumptions
[17:35:18] <leifj> sam: is extensions flexible enough to meet our needs?
[17:36:02] <leifj> (slide1)
[17:36:13] <leifj> slide2
[17:36:34] <leifj> Why remap identifiers? shib and other saml-based id-systems remap identifiers
[17:36:57] <leifj> http://shibboleth.internet2.edu/
[17:38:00] <leifj> slide 3 - Why involve destination kdc? - protocol symmetry, group membership/authorization data, stoplists single point of control
[17:38:11] <jhutz> Presentations available via https://onsite.ietf.org/public/meeting_materials.cgi?meeting_num=64
[17:38:24] <leifj> slide 4
[17:38:26] <leifj> privacy
[17:38:33] <leifj> (thx jeff)
[17:39:13] <leifj> slide 5 - adding identifier mapping
[17:41:30] <leifj> slide 6 - saml
[17:42:30] <leifj> slide 7- saml & kerberos
[17:43:45] <leifj> slide 8 - how saml integration might work
[17:44:42] <leifj> slide 9 - saml : hard problems
[17:46:11] <leifj> slide 10 - mapping for accounts
[17:47:55] <leifj> slide 11 - account mapping: questions
[17:48:30] <leifj> slide 12+13- involving the destination kdc
[17:48:46] --- stefans has joined
[17:49:32] <leifj> slide 14: privacy
[17:49:44] <leifj> s/14/15/
[17:50:26] <leifj> slide 16: possible solutions
[17:50:42] <leifj> slide 17: privacy: questions
[17:51:07] <leifj> slide 18: questions
[17:51:33] <bcneuman> WRT: The concluding questions - on mapping of ID's I think there is a difference between mapping when one crosses a system bounardy, and mapping when one crosses realms within a system.
[17:52:23] --- raeburn@jis.mit.edu has joined
[17:52:25] <leifj> sam - I disagree
[17:52:27] <bcneuman> For example, I think that for straight cross realm kerberos, names should be preserved, with the path embedded in transited. But, I also feel that when a system requires name mapping to a name in a local realm, that such mapping should be possible.
[17:53:09] <bcneuman> Also, when a name in one system of authentication is used as a basis for authentication in another, then perhaps the name mapping becomes required. We dealt with some of this in the pkinit work.
[17:53:25] <leifj> <jeff relays Cliff>
[17:54:21] <bcneuman> As to involving the end KDC in cross realm authentication, I thought that it is involved. There is the issue of who makes the "authorization" decision, and we provided a means for the KDC to validate the transited path for the end node, but the final choice was on the end node.
[17:54:22] <leifj> simon: shishi solves the privacy issues
[17:54:32] <leifj> by always running over tls
[17:54:43] <leifj> sam - asks about using a reserved bit
[17:55:14] <leifj> sam, jeff and simon discusses the use of tls
[17:55:45] <leifj> simon - will change the use of the reserved bit
[17:56:35] <leifj> jeff: do you want to move the tls work to the wg?
[17:56:37] <leifj> simon: yes
[17:56:42] <krb-wg> http://josefsson.org/krb5starttls/
[17:56:45] <leifj> jeff: do we want to work on this?
[17:57:17] --- lha has left: Replaced by new connection
[17:57:33] <leifj> sam: kerberos depending on tls and/or pki is bad
[17:57:45] <leifj> sam: I may be changing my mind
[17:58:05] <leifj> sam: secure kerberos preauth mechs is *really* hard
[17:58:25] <leifj> sam: encrypted channel may be needed but is tls the only option?
[17:58:35] <leifj> sam: tls has different model for ciphers
[17:58:56] <leifj> sam: tls has the most restrictive registry for ciphers
[17:59:23] <leifj> sam: in favour of doing something like this but need to look at more options beyond tls
[18:00:12] <leifj> <jeff relays Cliffs q about cross-realm>
[18:00:19] <leifj> sam: will respond to Cliff offline
[18:00:53] --- lha has joined
[18:01:08] <leifj> jeff: do we want to include extensibility for simons stuff
[18:01:34] <leifj> nico: extension and tls are separate
[18:02:03] <leifj> sam: not opposed, udp symmetry might be needed still
[18:02:51] <leifj> jeff sam and simon discusses the extension bit simon uses wrt tls tcp and dtls
[18:03:32] <leifj> jeff: move discussion to the list
[18:04:02] <leifj> jeff: back to cross-realm kdc issue
[18:04:26] <jhutz> (did cliff find the audio?)
[18:04:54] <bcneuman> No, can someone give me apointer.
[18:05:00] <leifj> sam will scribe his note
[18:05:03] <hartmans> So, in the athena->cmu case I get the cmu tgt from the athena kdc
[18:05:17] <hartmans> There is no single stop or control point for all cmu tgts.
[18:05:33] <hartmans> I'm proposing that you get a cmu tgt from the cmu kdc using the cmu tgt you get from the athena kdc
[18:05:41] <bcneuman> But the CMU TGT is useless until one converts it to a service ticket by contacting the CMU tgt.
[18:05:41] <lha> cliff: http://videolab.uoregon.edu/events/ietf/
[18:06:06] <hartmans> Right, but you have to do that for every service.
[18:07:18] <Jeffrey Altman> issuing a TGT for the CMU realm allows an optimization to occur so that you do not have to perform the mapping operations or acquisition of the authorization data which may come from another network service.
[18:07:32] <bcneuman> I have audio.
[18:07:33] <Jeffrey Altman> With Sam's proposed change, you do that once
[18:07:54] <leifj> rlbob talks about saml
[18:08:43] <leifj> talks about where requirements for saml came from
[18:09:08] <leifj> rlbob: are the requirements reflected in saml the right ones?
[18:09:56] <leifj> rlbob: other areas in the ietf have come to the conclusion that saml does not always fit the bill
[18:10:53] <leifj> nico: useful work - should be wg items
[18:13:08] <leifj> leifj: vo-contributed attribute-assertions
[18:13:22] <leifj> rlbob: hard problem not solved by anyone
[18:13:25] <leifj> leifj: true
[18:15:10] <leifj> nico: we only have authorization data one way
[18:15:25] <leifj> nico: maybe in the realm of extensions
[18:15:41] <leifj> kurt: I have use-cases of mutual authorization data
[18:16:01] <leifj> sam: what does mutual authorization data mean?
[18:16:14] --- wyllys has left: Logged out
[18:16:45] <leifj> sam, nico and kurt discusses the precise nature of 'mutual auhorization data'
[18:17:34] --- stefans has left: Disconnected
[18:17:46] <bcneuman> There is one such assertion in a bit in the ticket that says the server is authroized to accept certain forwarding. This is a one off item and is a kludge, but it gives one example of what one might want to provide in a more fomalized method.
[18:17:46] <leifj> kurt: use-cases involved gssapi and ldap; ldap is used in server2server situations where authorization data needed both ways
[18:18:05] <bcneuman> I can also see this being used for some kind of anti-phishing activities.
[18:19:27] <leifj> tarek kamel: how do you know when authorization data extensions are critical?
[18:19:46] <leifj> sam: yes - negative acl's and other situations might introduce problems
[18:20:21] <leifj> <nico tarek and sam discusses this>
[18:22:03] <leifj> jaltman: discussion about this (negative acls) ran out of time in paris - more talk this time
[18:22:09] <leifj> in kitten (I assume)
[18:22:23] <leifj> jeff: other comments?
[18:22:25] <hartmans> Cliff, yes, I can see that sort of thing.
[18:22:29] <leifj> jeff: next topic
[18:22:44] <leifj> jeff: larrys anoymous draft
[18:23:31] <leifj> larry: discribes issue from <?> raised about the draft
[18:24:10] <leifj> larry thinks this issue is theoretical
[18:24:34] <leifj> larry gets oposition on this
[18:25:16] <bcneuman> Is this a bug, or a protocol issue.
[18:25:38] <bcneuman> Why would one use the same session key in two separate tickets issued by the KDC unless the client specifically requested it.
[18:25:43] <leifj> protocol issue!
[18:25:43] <tlyu> i think there is a protocol bug here
[18:26:04] <leifj> larry: take this to the list?
[18:26:09] <leifj> nico: other issues to discuss
[18:26:25] <leifj> jeff: take this after pkinit
[18:27:16] <leifj> jeff puts up larrys list of issues
[18:27:43] <leifj> jeff: pkix gave bad review on asn.1
[18:28:24] <leifj> jeff: pkix-folk also thought octect string wrappers around cms were unnecessary
[18:28:50] <leifj> jeff: in both cases these things were known and in the latter case was there for a reason
[18:29:04] <leifj> jeff: not reopening this - nothing new has come up
[18:29:13] --- stefans has joined
[18:29:57] <leifj> jeff: document now reflects rough consensus
[18:30:50] <leifj> jeff: comments should not be the only place where a normative reference occurs - this issue has been disucssed
[18:30:58] <leifj> jeff: larry has a counter argument
[18:31:39] <leifj> larry: minor problem - few if any places in the spec has this problem
[18:32:24] <leifj> larry: other documents (eg cms) also does this
[18:33:03] <leifj> larry: don't want to make change to address this issue
[18:33:12] <leifj> jeff: what does the wg say?
[18:33:41] <leifj> kurt: iesg has a statement on the use of formal lanaguages - support not replace the text
[18:33:46] <leifj> jeff: irrelevant
[18:34:00] <leifj> kurt: this means that normative refs should not be a comment
[18:34:40] <leifj> jeff and sam things this statement talks about abuse of ABNF
[18:35:07] <leifj> kurt: it may be so but it is still bad form
[18:35:33] <leifj> jeff: willbe another version of this document with typos
[18:36:01] <leifj> jeff: who will propose textual changes to address the issue with normative refs
[18:36:09] <leifj> jaltman: volouteer
[18:36:23] <leifj> sam: is this important enought to do?
[18:36:47] <leifj> jeff: does the wg think this is needed?
[18:37:21] <leifj> (humming)
[18:38:03] <leifj> jeff tries to gague interest - probably need to do the work
[18:38:13] <leifj> jaltman does the work
[18:38:37] <leifj> jeff: volounteer to compile the asn.1
[18:38:47] <leifj> lha will voluouteer using heimdal compiler
[18:39:19] <leifj> larry: always compile before submit!
[18:40:30] <leifj> jeff: other pkinit issues?
[18:40:53] <leifj> larry: love - please go through your issues
[18:41:05] <leifj> larry reads loves comments
[18:41:16] <leifj> and comments
[18:42:26] <leifj> love and larry disagreses on the use of local policy wrt to DH group parameters
[18:42:32] <leifj> love claims a downgrade attack
[18:42:49] <leifj> jeff: gists loves argument
[18:43:05] <leifj> larry: preconfiguration solves the problem
[18:43:13] <leifj> love: I don't want configuration on the clients
[18:43:44] <leifj> jeff: duel or this is just 'rough' consensus
[18:43:48] <leifj> love: I fold
[18:44:14] <leifj> next issue is discussed
[18:44:52] <leifj> lha :unsigned clear text strikes again --- hartmans
[18:45:06] <leifj> larry: generic problem...
[18:45:25] <leifj> lha: solution is simple - solve it now, don't wait for extensions
[18:45:34] <leifj> larry: no clear threat, wait for extensions
[18:46:05] <leifj> lha: it is a problem
[18:46:11] <leifj> larry: not specific to pkinit
[18:46:25] <leifj> jeff: do other people have anything to say?
[18:46:59] <leifj> jeff: humming on fix/no-fix about this?
[18:47:53] <leifj> jeff summariezes the question
[18:48:53] <leifj> either fix because this is an easy fix and the same fix for the same issue in the rsa-case already fixed, or don't fix becasue extensions will fix this later and this is not a threat now
[18:49:24] <leifj> sam: compromise - will will have to look at hashfunction upgrades, fix it then
[18:49:39] <leifj> love: I can live with that
[18:49:51] <leifj> nico: same as pushing to extensions!
[18:49:53] <leifj> sam: no sooner
[18:50:06] <leifj> larry: can do it very soon
[18:50:35] <leifj> humming for support for sams compromise
[18:50:57] <leifj> the yes have it
[18:51:12] <leifj> next issue
[18:52:26] <leifj> lha: issue is that q is not defined for most groups
[18:52:41] <leifj> larry: yes it is
[18:52:47] <leifj> jeff: resolve outside the room
[18:53:00] <leifj> next issue
[18:53:27] <leifj> larry: lha claims reference to section in external document is missing
[18:53:36] <leifj> lha: no I can't implement as it stands
[18:54:07] <leifj> jeff: let's not discussion the cost of IEEE standards
[18:54:33] <leifj> lha: need someone to verify the correctness of the specification
[18:54:50] <leifj> sam & tom : may need to use IEEE liason
[18:54:55] <leifj> next issue
[18:55:29] <leifj> this is a clarification of existing text
[18:55:44] <leifj> sam and jeff supports call for clarification
[18:56:23] <leifj> jeff: permitting both empty and absent sequences defeats the purpous of requireing DER
[18:56:47] <leifj> larry: yes but it won't make any difference
[18:57:02] <leifj> next issue
[18:57:19] <leifj> typographic
[18:57:21] <leifj> next issue
[18:57:49] <leifj> issue involves signed attributes
[18:57:59] <leifj> sam supports lha
[18:58:29] <leifj> jeff: one more issue
[18:58:58] <leifj> jaltman: introduction section needs new text - plese review on the list
[18:59:18] <leifj> sam: attacks on sha1 and md5
[18:59:52] <leifj> sam : the security community places empasis on algorithm agility
[19:00:10] <leifj> sam: sbelovin and rescorla has looked at this and the results are not good
[19:00:30] <leifj> sam: protocols tend not to be good at handling upgrade of algorithms
[19:00:55] <leifj> sam: pkinit should not suffer from this problem by the time it reaches the iesg
[19:01:15] <leifj> sam: hash neg upgrade strategy needs to exist
[19:01:50] <leifj> jeff: please review!
[19:02:06] <leifj> jeff: steve and rescorla will present on this at saag
[19:02:18] <leifj> jeff :milestones
[19:02:31] <leifj> 8 active milestones
[19:03:40] <leifj> jeff goes throug list of milestones on the web
[19:05:01] <leifj> milestones bumped 1 year
[19:05:37] <leifj> jeff: propose charter review for march 06
[19:05:58] --- nico has joined
[19:06:00] <leifj> larry: need milestones for ecc document
[19:06:17] <leifj> jeff: and need one for anonymous
[19:07:16] <leifj> jeff: propose issues on anonymous by next ietf
[19:07:34] <leifj> jeff: propose ecc lastcall by next month
[19:08:00] <leifj> jeff: simons stuff discussed on list before taking it on
[19:08:46] <leifj> ken: discusses an issue in enctype neg
[19:09:06] --- kivinen has left
[19:09:11] <leifj> ken and sam decide to not pursue this issue and go ahead with the document
[19:09:30] <leifj> enctype neg-04 should go ahead
[19:09:41] --- hartmans has left
[19:10:05] <leifj> jeff: we are done
[19:10:08] --- leifj has left
[19:10:13] --- bcneuman has left
[19:10:55] --- lha has left
[19:10:59] --- stefans has left: Disconnected
[19:11:20] --- tlyu has left
[19:11:33] --- stefans has joined
[19:12:13] --- warlord has left
[19:12:54] --- stefans has left: Disconnected
[19:14:28] --- jhutz has left
[19:15:12] --- raeburn@jis.mit.edu has left: Disconnected
[19:16:08] --- nico has left: Disconnected
[19:27:30] --- krb-wg has left: Disconnected
[19:49:40] --- stefans has joined
[19:54:17] --- stefans has left
[20:04:35] --- Jeffrey Altman has left
[20:17:02] --- Jeffrey Altman has joined
[21:41:46] --- Jeffrey Altman has left