CDNI Working Group Minutes IETF-95, Buenos Aires, Argentina - Chaired by Francois Le Faucheur, and Ray and Brandenburg (on behalf of Kevin Ma) - Meeting notes captured by Kent Leung and Francois Le Faucheur, edited by Francois Le Faucheur - Audio Recording at: https://www.ietf.org/audio/ietf95/ietf95-atlanticob-20160407-1400.mp3 - Slides accessible at: http://www.ietf.org/proceedings/95/cdni.html Thursday, April 7, 2016, 14:00-16:00, Room Atlantico B ======================================================= - about 50 people in the room, plus 8 on MeetEcho (including co-chair Kevin Ma) Introduction and Agenda (WG chairs) --------------------------------------------------------------- - Introduction by the WG chairs, and Note Well statement. - Change in WG personnel: * change in Assigned Area Director: Alexey Melnikov replacing Barry Leiba. * Thanks expressed to Barry Leiba on behalf of CDNI WG as well as ICE WG - Agenda review, no request to change agenda - RFC published since previous IETF meeting : RFC 7736: CDNI Media Type Registration (Informational) - Document Update and progress against the charter milestones * cdni-logging: still in “IESG Review.” All DISCUSSes cleared. * cdni-metadata: Ready for WG Last Call? * cdni-redirection: under “IESG Review”. Comments being addressed. * cdni-control-triggers: under “IESG Review”. Comments being addressed. * cdni-footprint-capabilities-semantics: Semantics under “IESG Review”. Comments being addressed. * cdni-uri-signing: o WG document on URI signing: Ready for WG Last Call? o URI signing for HAS: IPR update - Documents beyond the charter: * CDNI handling of HTTPS Delegation: new rev reflecting Yokohama discussions. CDNI Logging, draft-ietf-cdni-logging-24: Francois Le Faucheur --------------------------------------------------------------- - Remaining ABNF corrections included (with help from Pete Resnik) - See slides: o clarifications regarding c-groupid mapping and implementation requirement level (MAY) o resolved comments from Alissa Cooper (as discussed on mailing list) - ready for publication CDNI Metadata, draft-ietf-cdni-metadata-12: Kevin Ma ------------------------------------------------------------ - See slides: o updates and changes since last rev o one open question regarding removing “.v1” or not , but no comment either way from the room. - Chair/Francois proposes that document moves to WG Last Call and asks if anyone objects. No-one objects. - Chair/Francois recommends a WG Chair review be conducted at the same time. FCI Semantics : Kevin Ma draft-ietf-cdni-footprint-capabilities-semantics-12 ------------------------------------------------------------------------------- - 4 revs since last IETF meeting - See slides for changes - Rev -12 was submitted to IESG review - all IESG comments today have been addressed - will work on upcoming IESG comments - Chair/Francois asks about I-Ds for actual FCI interfaces. Kevin indicates that Kevin and Jan will provide new rev for IETF-96 Routing Request Redirection for CDN Interconnection, draft-ietf-cdni-redirection-17: Ray van Brandenburg ---------------------------------------------------------------------------------------------------------- - 4 new revs since last IETF - See slides for changes - was submitted to IESG review - currently addressing IESG review comments, will continue to do so as comments come - Several IESG comments about privacy concerns around cdni-redirection - Kevin Ma: change the tone of document but allow all info to be set in doc, the info does not need to be encrypted (heavy) given TLS is used - Francois: Talk about tradeoff of privacy vs. info, even with TLS; should the dCDN get all the info? - Kevin Ma: The info may not be stored, used in memory, maybe a note on this info is in transit - Ray: issue is given the info to a 3rd party - Francois: not sure if that's the issue; need to understand the specific concern for RRI in order to be able to address them - Sanjay Mishra : is the fundamental concern about the fact that uCDN redirects the request to dCDN? Ray/Francois answered (concern is not redirection itself which is inherent to CDNI model, but concern is uCDN asking dCDN and providing a lot of info, so dCDN can respond pointing request to a specific surrogate) - Kent: is the concern when Client IP for first request may be different than client IP for fetching content from surrogate - Ray/Francois: no, not that - AD/Alexei: How about having a chat with security reviewer ? - Chair/Francois: Can we deal with this privacy IESG concern the same way we dealt with it for cdni-logging, because that was very effective? this involved a mini-Design-Team with CDNI authors as well as two designated privacy/security experts. - AD/Alexey: Please send me some background and I will work with privacy/security experts to set up the same approach - Chair/Francois: I will send all details CDNI Control Interface, draft-ietf-cdni-triggers-12: Rob Murray ------------------------------------------------------- - 3 new revs since last IETF- See slides for changes - Rob will include the latest minor tweak in TLS text from cdni-logging - started addressing some IESG review comments - exAD/Barry Leiba: general comments: o respond to IESG comments quickly, either thanking them or inform them that you’ll respond later o For DISCUSSes: need to engage AD and discuss, not go back and change draft only o No response to Ben's comment, and to Alissa’s comments o Barry’s “Yes” position will disappear (since he is no longer AD) o Alexey will review for one additional position o note that there are 3 CDNI docs in IESG, so a lot of work in the pipe for ADs CDNI URI Signing, draft-ietf-cdni-uri-signing-07: Kent Leung ------------------------------------------------------------- - 2 versions since last IETF - See slides for changes - Specific provisions for privacy (encryption of IP address/prefix) - Leif Hedstrom conducted expert review. His comments will be addressed in next rev. - Leif Hedstrom: Apple is developing an implementation of the uri-signing spec. Target is to make it available in Apache Traffic Server (tentatively in May 2016). - Ray and Brandenburg: planning to also have an implementation of the uri-signing spec within 6 weeks - Next steps: o Address WG comment on path method for URI Signing o Address Brian Weis's review comments o Address Ray's implementation feedback o Address Leif's implementation feedback - Chair/Francois: after these steps document will be ready for WG Last Call. Does anybody object to issuing WGLC then (tentatively in 2nd half of May)? no-one objects. - Leif and Phil agreed to review draft in WGLC URI Signing for HAS draft-brandenburg-cdni-uri-signing-for-has-02: Ray van Brandenburg -------------------------------------------------------------------------------------- - History: At Dallas IETF meeting, WG decided to remove segment-specific extensions from uri-signing WGI document because of KPN IPR statement. Segment-specific extensions were re-published as draft-brandenburg-cdni-uri-signing-for-has. - MPEG remains interested in a solution that is compatible with CDNI URI Signing o Latest liaison (March 8) at https://datatracker.ieL.org/liaison/1464/- Several new versions of draft-brandenburg-cdni-uri-signing-for-has- Ray had discussions with KPN about IPR terms - See slide: KPN is proposing some potential alternative IPR terms and ask whether the CDNI WG would consider re-incorporating the segment-specific extensions into the WG document with such IPR terms. - exAD/Barry: has the actual IPT statement to IETF been updated with these new proposed terms? - Ray: no it has not been updated yet. - exAD/Barry: an IETF WG should not be involved in negotiating IPR terms (proposed new IPR terms). Please update the IPR statement, then WG can decide what is the next step - Alexey : Clarify that IPR is for draft-brandenburg-cdni-uri-signing-for-has only, not main URI Signing draft, so the IPR should also be updated in terms of the related document - Chair/Francois: no decision can be made today, but when/if IPR is updated, then we can discuss on the mailing list HTTPS and delegation of encrypted traffic, draft-fieau-https-delivery-delegation-02: Frederic Fieau --------------------------------------------------------------------------------------------------- - Latest changes: o Added a section on HTTPS delivery delegation requirements o Rephrased HTTPS section o Added a section on a Lurk interface in CDNI - See slides for details on changes - Ray: Should draft go to LURK BOF (if it becomes a WG) - Sanjay: LURK is only a BOF (no decision to turn into a WG yet), another BOF is coming - Frederic: some work needed in CDNI as it is CDNI specific - Francois: track what things are needed in CDNI in document, but main work should be in LURK (if/when created) - Barry: the only use case being considered in LURK BOF is CDN - ? : there may be other use cases - Sanjay: why the work should not be here? - Francois: the technical details of delegation should be handled in LURK, this WG should be a consumer of the LURK solution (e.g. defining how it affects CDNI interfaces e.g. metadata interface/FCI can support LURK parameters needed for CDNI) - Kevin Ma: I agree with Francois - Barry: CDNI people should actively participate in LURK - Francois: Frederic and Kevin are involved already in LURK - Thomas Fossati: dCDN needs certificate from CP/uCDN to authenticate Closing Remarks --------------- - AD/Alexey: We want to close the CDNI documents already under IESG review before submitting more to IESG review. Meeting closes--------------