Minutes IETF98: perc

Meeting Minutes Privacy Enhanced RTP Conferencing (perc) WG
Title Minutes IETF98: perc
State Active
Other versions plain text
Last updated 2017-04-02

Meeting Minutes

Privacy Enhanced RTP Conferencing
Session 2017-03-29 9:00 - 11:30
Chairs :- Richard Barnes,  Suhas Nandakumar

Scribe: Dan Burnett
Jabber: Matt Miller


* draft-ietf-perc-double, draft-ietf-perc-srtp-ekt-diet
  - Cullen presented the draft. no major open issues
  - Authors to submit updated draft with editorial fixes
    to enhance readability.
  - expected next version by 1st May

* draft-ietf-perc-private-media-framework
  - No major open issues.
  - Authors to submit updated drafts with
    ++ key hierarchy explanation in perc context
    ++ packet format details for a double rtp packet
    ++ update references to dtls-id draft
  - expected next version by 1st May

* draft-ietf-perc-dtls-tunnel, draft-jones-perc-dtls-id
  - Conference Id replacement with dtls-id was discussed and approved
  - Authors to update tunnel to refer to Martin Thomson's tls-id draft (in
  mmusic) - expected next version by 1st May

* draft-roach-perc-webrtc
  - Adam discussed call flows for using perc in webrtc context
  - No major issues on overall direction of the draft
  - Next updates include:
      ++ consider including MD initiated callflow
      ++ consider removal of Offer/Answer signing and scope it to just
      fingerprint ++ possibly add new co-author(s)

* draft-grozev-perc-double-rtx, draft-grozev-perc-ssrc
  - Authors discussed Rtx considerations within perc and plausible use of
  framemarking - Discussion on SSRC mutability and OHB extensions were
  proposed. - No conclusions made on either topics. Tabled till next interim

* Action Items and Next Steps
  - Authors to submit updated version by 1st May
  - Chairs to setup virtual interim before June to help meet the milestones
  - WG to discuss double usage with Rtx/Fec at the interim

Raw Notes (Dan Burnett)

Chairs  Introduction
No milestone changes since IETF 97.
No changes to agenda.

Paul Jones      DTLS Tunnel     draft-ietf-perc-dtls-tunnel
Discussion about the final point on Tunnel Affinity slide (slide 4) -- concerns
with assuming same-name key servers can all receive messages for a given DTLS
association. Cullen suggests defining id that names kd pool and md pool as
context. Slide 6: EKR suggests type before length.

Paul Jones      DTLS Id in TLS/DTLS     draft-jones-perc-dtls-id
Slide 5: MD needs to be able to read the dtls_id.  No in-room concerns with
using this in place of the conference id, etc.  Plan to refer to Martin's
document in MMUSIC.

Adam Roach      Perc in WebRTC Context  draft-roach-perc-webrtc
Slide 8: Adam notes that some of this content will obviously change as a result
of the id discussion we just had. Slide 9: Adam notes that these flows are
illustrative only and not normative or comprehensive.  We convey a name back to
the client so he can do a cert check. Slide 16: Long discussion.  Adam will
rework how the MD asserts the KD's identity. Slide 30: It was pointed out that
this is like a Peer Connection that does not do media.  The permission and
acces needs are essentially the same. Slide 32: Show of hands in room for
continuing work like this?  Good support in the room.

Cullen Jennings Ekt on Diet     draft-ietf-perc-srtp-ekt-diet
Code point change slide:  no one obects to the change from 3 to 2 for aeskw_256.
(some decision that I didn't catch about what EKR wanted here, Richard to fill

Cullen Jennings SRTP Double Encryption  draft-ietf-perc-double
EKR wants to bring up structure of OHB.  Cullen thinks the client should always
insert. EKR wants us to be precise on the syntax/structure as well as
describing what the purpose/allowed uses of it are.  Cullen will restrict the
use to be for statistics reports.

David Benham    Perc Media Framework    draft-ietf-perc-private-media-framework
No discussion or decisions.

Emil Ivov       Perc RTX Considerations TBD
Slide 6: An important point raised is that this is not just about padded (fake)
packets, but rather about how to handle repair streams that may need to be
intercepted at the hop level. But Emil says we still need to indicate/handle
discardable packets. Lots more discussion about how repair streams work. 
Decision is to have a focused call on this topic. Slide 9: A point was made
that SFUs could also switch among the timestamps rather than creating a new one.

TBD     TBD/Chair Wrap up       TBD
Chairs ask which docs need more work before Last Call.  The answer is most of
them.  Target given by chairs is that authors will update their docs by May so
we can Last Call by June.