Re: [DMM] DMM WG Adoption Poll (2) for "Mobile User Plane Evolution" - draft-zzhang-dmm-mup-evolution-06

I joined this mailing list recently so I’m not familiar with the full history of this draft. Sorry for SPAM in case this has been discussed earlier.

I noticed few fundamental issues with “ANUP”:

  *   Mobility management and session management are fundamentally 2 different functions and must not (cannot) be merged
     *   Most mobile use-cases require mobility across eNB/gNB. Even FWA has inherent requirements for mobility due to fallback or interrupted LoS with mmWave. Same with private wireless as it requires X2 or S1 handovers. LIPA is no exception – see below.
     *   Session management is anchored permanently with 4G. 5G allows session mobility but not widely used/deployed.
  *   Even if LIPA PDN is enabled on a UE (e.g for local IP access at home or office), rest of the PDNs require mobility (walking away from home eNB) so merging mobility management and session management is not feasible.
  *   LIPA PDN requires SGW function to be done in the MNO core. There is no way to do “local” processing between HeNB and LGW
     *   MNO has to control paging (otherwise major security risk) to activate radio link from HeNB so the user plane function does a hair-pin to SGW and returns to LGW
     *   Same with 5GC with I-SMF/I-UPF activated
  *   LIPA didn’t become a reality due to limited use-case (local IP access only). No major benefit due to widespread use of WiFi access.
     *   Major security issue for service providers. No monetary gain to provide such a service

“ANUP” can be considered as a product architecture where HeNB (serving Internet, voice, + LIPA) and LGW-U/UPF (LIPA only) are collocated but are 2 independent functions.

PGW-U/UPF functions for Internet and voice will always reside in MNO network so HeNB will handover to eNB/gNB as soon as UE walks away from HeNB.

Unless the focus is on a very specific case of single PDN with no mobility (“private MEC” or “satellite to satellite” only), it seems ANUP like function is unnecessary at this point and crossing over into 3GPP territory. Proposing to merge N1/N2 and N4 signaling will be an overkill to solve a very niche use-case.

Sandeep

From: dmm <dmm-bounces@ietf.org> On Behalf Of David Lake
Sent: Friday, October 20, 2023 12:50 AM
To: Miya Kohno <miya.kohno@gmail.com>; Satoru Matsushima <satoru.matsushima@gmail.com>
Cc: dmm <dmm@ietf.org>; draft-zzhang-dmm-mup-evolution@ietf.org
Subject: Re: [DMM] DMM WG Adoption Poll (2) for "Mobile User Plane Evolution" - draft-zzhang-dmm-mup-evolution-06

CAUTION: This is an external email. Please be very careful when clicking links or opening attachments. See the URL nok.it/ext for additional information.

Miya

“ANUP solely addresses user plane convergence and tries to liaise with 3GPP, but I'm afraid this is a half-baked attempt.”

Absolutely agree and there are many aspects of the current 3GPP design which are not very ‘IETF-like’ that we should be considering (removal of IMS and total separation of the voice/text service from the transport network would be the first item I’d consider – then move onto the use of tunnels, then look at authentication, then at a truly global-mobile solution… etc.)

So this draft is good but I agree it only really addresses the ‘here-and-now’ and we have much in the control plane that needs reworking.

What I think we need in IETF (and it is not in a chartered group such as DMM – that scope is known) is a study of the current ‘mobile’ infrastructure with reference to how today’s consumers are using it predominantly for Internet access (a good thing for IETF and that includes voice and text which is mostly OTT now) and whether the current economic and architectural models fit that use case.

Where I struggle is that is probably not an IETF WG or IRTF RG.

David

From: dmm <dmm-bounces@ietf.org<mailto:dmm-bounces@ietf.org>> on behalf of Miya Kohno <miya.kohno@gmail.com<mailto:miya.kohno@gmail.com>>
Date: Friday, 20 October 2023 at 02:02
To: Satoru Matsushima <satoru.matsushima@gmail.com<mailto:satoru.matsushima@gmail.com>>
Cc: dmm <dmm@ietf.org<mailto:dmm@ietf.org>>, draft-zzhang-dmm-mup-evolution@ietf.org<mailto:draft-zzhang-dmm-mup-evolution@ietf.org> <draft-zzhang-dmm-mup-evolution@ietf.org<mailto:draft-zzhang-dmm-mup-evolution@ietf.org>>
Subject: Re: [DMM] DMM WG Adoption Poll (2) for "Mobile User Plane Evolution" - draft-zzhang-dmm-mup-evolution-06
Hello,

Though this draft is mostly agreeable, drastic improvement also requires the control plane work.
In fact, RAN Core convergence has already been discussed.
e.g.
https://ieeexplore.ieee.org/document/10073488
https://ieeexplore.ieee.org/document/10001281

ANUP solely addresses user plane convergence and tries to liaise with 3GPP, but I'm afraid this is a half-baked attempt.

The MUP[*] assumes no change to 5GC, and I think IETF DMM should focus on this at least for the time being.

Architecture: https://www.ietf.org/archive/id/draft-mhkk-dmm-srv6mup-architecture-05.html
Motivation: https://www.ietf.org/archive/id/draft-kohno-dmm-srv6mob-arch-07.html

Cheers,
Miya

On Thu, Oct 19, 2023 at 6:42 PM Satoru Matsushima <satoru.matsushima@gmail.com<mailto:satoru.matsushima@gmail.com>> wrote:
Dear DMMers,

This email starts a two-weeks DMM WG adoption poll (2) for ""Mobile User Plane Evolution" - draft-zzhang-dmm-mup-evolution-06.

https://datatracker.ietf.org/doc/draft-zzhang-dmm-mup-evolution/

Please review the draft and post any comments on this mail thread prior to Friday, November 3rd, 2023.

Regards,
Sri, Satoru

_______________________________________________
dmm mailing list
dmm@ietf.org<mailto:dmm@ietf.org>
https://www.ietf.org/mailman/listinfo/dmm