Minutes IETF123: dmm: Mon 15:00
minutes-123-dmm-202507211500-00

Mobility aware Transport Network Slicing for 5G

• A major disccusion on whether to make the draft Standard or Info
  . One Q asking for explanation why OPSAD suggests this be a STD doc
  (because of YANG module)
  . Still, the DMM WG considers this an INFO doc
  . One more comment from AD: thanks to YANG model, it is supposed to
  be STD
  . So, suggestion from DMM: get rid of YANG model and make this doc
  INFO -- note this is not the final decision from the DMM WG
  . again the DMM AD confirms YANG model does not necessarily make
  STD. So, everything can stay.
• Author shared the delta (Rev#17 -> #19)
• For security section: planning to reference a similar (security)
  section of TEAS and solicit more feedback from sec. AD.

SRH Reduction for SRv6 End.M.GTP6.E Behavior

• shared ver#00 after WG adoption
• Some design discussion: The order of bit field
• Since this doc revolves around SR-header design, so suggestion is to
  ask for feedback from the SPRING WG.

MUP Architecture for DMM

• shared ver#00 after WG adoption
• first, went thru a few discussion points, mainly on MUP-C API
• also recap'ed & described 'the MUP architecture'; provided a list of
  field deployment cases utilizing MUP architecture.
• Q/A: talking about sessions under the 5G context, and compared with
  the sessions in MUP architecture; particularly asked around '5G
  session has many session related parameters and whether MUP sesssion
  might take some or all of them' -- reply is 'only a small set of
  (5G) session info/parameters are leveraged by MUP to influence
  routing'
• Discussions: all the mobile sessions will be established via regular
  mobile CP signalling. AFter the establishment of sessions, those
  sessions info will be shared (or downloaded) to the MUP-C for
  digestion. From then on, all the data-plane traffic will only go
  thru MUP (e.g., PEs), bypassing UPF (like in 5G case)

Mobile Traffic Steering

• author shared the latest revision of the draft
• Comments: this draft does address the gap regarding the end-to-end
  architecture
• Clarification Qs: is reference intf btwn MCS and MTS per-flow or
  aggregated-flow? Answer: it does not matter since the model can
  handle both per-flow and aggregation cases.
• how about QoS/slicing: can this MTS model handle QoS/slicing?
• Q/A: mobile system interacts with TN: Q centralizes on 3GPP/5G
  related discussion, and Ans is 'the draft tries to make a
  generalized end-to-end architecture, not just for 3GPP, though 3GPP
  is of course a good scenario for the draft'.