Kerberos WG (krb-wg)
THURSDAY, March 4, 2004
0900-1130
Sapphire 1
======================================================
Chairs: Douglas Engert (not present), Jeffrey Hutzelman
Scribe: Richard Graveman
INTRODUCTION: Jeff Hutzelman
Jabber was used also by a majority of the WG, about 20 people.
There were no changes to the agenda.
DOCUMENT STATUS UPADTES
Several documents were sent to the IESG:
crypto framework (draft-ietf-krb-wg-crypto-07): New version was sent out before the
meeting. All issues should be resolved (Hartman). Waiting for the AD. IANA
considerations need to be checked.
clarifications (draft-ietf-krb-wg-kerberos-clarifications-05): BCN sent a new
version and believes all issues are resolved. Back to the AD. RH reported
three "discuss" votes.
GSSAPI-CFX (draft-ietf-krb-wg-gssapi-cfx-06): RH sent comments and the author
responded before the cut off. There was one more acknowledgment quest to be
dealt with editorial. Currently in WG LC.
KERBEROS-EXTENSIONS Status Update - Sam Hartman
In fairly good shape on protocol. The issue tracker will be changed. Text from
BCN has not arrived. Clarifications in 1510 may need change (Tom Yu).
Alternate Structure - Sam Hartman, for Tom Yu (draft-yu-krb-wg-kerberos-extensions-00)
Currently a skeletal individual submission. Alternative structure, but should describe
the same wire protocol. Motivations for change were clarity and removing
verbosity especially sections 3 and 5. Semantics embedded in Section 5; partial
material in Section 3 is hard to understand in one pass. Goals are to start with an
overview, get a good hierarchical structure, put semantics in one place with message
definitions, and remove implementation specific detail. Overview information will precede
message descriptions. Treat TGS and AS requests as specializations of the KDC request.
Work from common elements to differences. Describe Kerberos in terms of ASN.1 types.
The new layout is Overview, Basic Concepts, Individual Sections for the Three Protocols
of Kerberos, IANA, Security, etc.
The Overview explains TTP, use of symmetric crypto,
the three protocols. ASN.1 use, Principals, and Encrypted Data, Tickets.
· Credentials Acquisition
· Application Authentication
· Session key use
It describes common elements of KDC request handling; better discussion of keys;
clear up time handling (interactions of different time fields; who updates what).
Missing naming issues, transport, and typed holes. There are some protocol differences
(nonces, checksums, etc.).
Use text from Clarifications.
Questions and discussion:
Only a few people had read the proposal.
Is the structure better than the structure of Clarifications?
All who had an opinion thought this was an improvement. Easier for implementers.
Which structure should be adopted? Either approach requires work; we should therefore
pick one. Support for new structure. Detail still needs to be filled in. More discussion
on the ML.
REFERRALS (draft-ietf-keb-wg-kerberos-referrals-03)
Nothing was presented.
PKINIT (draft-ietf-cat-kerberos-pk-init-18)
Work Status and Open Issues - Chair
Submitted before cut off. Pretty much up to date with all identified issues. Still
on track for March WG LC and submission to IESG. Push from Vienna IETF to get this
done. CableLabs hosted interim meeting; list of issues posted and tracked. Several
items resolved on ML and in Minneapolis (confirmed on ML). All text received and
incorporated. Issues:
SubjectAltName for user cert
Comments from Brian included proposed text. Will be posted on the ML.
Client name canonicalization problems (Nico Williams)
Resolved recently. Believe there is consensus. Yes; clients need to include
a name in PKINIT AS requests. Name returned by the KDC must match.
OCSP
Better text now exists. One minor point is still being discussed on the ML.
Preauth Type Numbers
Earlier drafts had different message numbers. This is likely to be resolved
quickly. Brian asked whether RH wants to see this before it goes to LC. Yes;
it will be emailed to him when the text is stable.
Eric Rosenthal and Doc Evens will do an editorial review before WG LC.
PREAUTHENTICATION FRAMEWORK, Sam Hartman (draft-ietf-krb-wg-preauth-framework-00)
At the previous WG meeting, work was started. A first draft describes in more
detail how preauthentication works (at the client). An "authentication fashion"
exists at the client. The client keeps track of keys and other state. There may be
multiple round trips; the KDC is stateless. The mechanism needs to provide cookies
back to the KDC. The extensibility model needs to deal with the set of mechanisms and
key change. All of this has to e wrapped in the proposal. (The alternative is that one
knows already that the mechanism is supported.)
State management is also described. The issue is that much of this was never documented.
The remaining (hard) stuff includes:
1. The KDC cannot deal with combinations of methods. This will be out of scope.
2. Multi-round requests, signatures, and binding have to be done with a stateless KDC.
3. Nonces and signatures need more clarification.
Some items will be deferred until the Extensions model can be used. What depends just
on Clarifications will be included. Discussion will take place on the ML. The hope is to
make this useful with respect to Clarifications and not broken by Extensions.
Concluding discussion and open microphone:
1. The schedule needs to be revisited.
2. More information was requested on referrals.
The meeting adjourned at 10:04.
DECISIONS and ACTION ITEMS:
* WG: Remaining PKINIT issues (SubjectAltName, cname canonicalization,
OCSP text) to be resolved on list by March 15
* Brian Tung: send new PKINIT to reviewers by March 22
* various: PKINIT review done by March 29
* brian tung: send new PKINIT to I-D repository by April 2
* chairs: review kerberos-clarifications and send back to IESG if ready
* chairs: review kcrypto and send back to IESG if ready
Many of the members of the WG, about 20 where participating via Jabber,
and the Jabber logs can be found at:
http://www.xmpp.org/ietf-logs/krb-wg@ietf.xmpp.org/2004-03-03.html
(Personal thanks to Richard Graveman for writing up the minutes.)
|