Satoru:

 

Many thanks for the comments. And sorry for the slow reply since I have been 
travelling internationally and just came back.

 

Regarding your question: quoted “Does the above text mean ANUP can be 
implementable with no change with 3GPP spec modifications?”

[Reply]
You have referenced the section 3.2 “…...As IETF technologies continue to 
evolve, they can be automatically applied in mobile networks without any 
changes in 3GPP architecture/specification…” – what we have stated in the 
preceding sentence (of the quoted words) is ‘…3GPP for radio access and IETF 
for the rest…’. The corresponding picture (in section 3.2) explains the 
splitting of the two parts.
Because of the splitting of the 3GPP-part and the IETF-part, we suggest the 
IETF technologies will be applied to the IETF-part, which is separate from the 
3GPP-part. Thus, the evolution of new IETF technologies is somewhat 
self-contained and will not have direct overflowing impact on the 3GPP 
architecture/specification which will evolve independently by themselves.
We understand your confusion and we will make this paragraph clearer in a new 
revision.
In summary: 
Based on our draft, the ANUP implementation will have impact to the 3GPP spec. 
That is exactly what we have elucidated (and as you quoted below) “… This 
section discusses potential MUP evolution in 6G mobile networks.  It does 
involve changes in 3GPP architecture and signaling, so the purpose is to share 
the ideas in IETF/wireline community first..’
 

As a final note: the 3GPP last release for 5G (i.e., Rel-19) is close to 
completion by the end of the year. The 6G SID in 3GPP SA1 has been approved 
last month and the study-phase work has started. Even if the consensus is to 
not introduce significant architecture changes in 6G, there will exist 
evolutions in many parts of the wireless network (i.e., RAN, CN-CP, CN-UP, 
etc.). The 3GPP is studying the 6G architecture now. We believe this draft (of 
ANUP) does bear the merit for the coming 6G evolution. So, we request the DMM 
WG to consider the adoption.

 

Thank you and BR,

 

-Tianji

 

 

 

From: Satoru Matsushima <[email protected]>
Date: Monday, October 14, 2024 at 2:30 AM
To: Tianji Jiang <[email protected]>
Cc: "[email protected]" <[email protected]>
Subject: Re: [DMM] IETF draft: MUP/ANUP -- For continuous discussion & comments

 

Thank you Tianji, for your followup on your draft.

 

I understand that you don't have an ANUP implementation. That's fine. And also 
I agree that rough consensus doesn't require running code to a Informational 
state document and WG operation in IETF as RFC2418.

In the context of the WG adoption call for your draft, we needed to seek 
technical accuracy as ones pointed out incorrect descriptions in the draft 
especially on 3GPP Cplane. Asking if a running code exists was one of my take 
to see if it is correct or not. 

 

But I still want to see if it is implementable or not. Because section 3.2 
stated:

 

...As IETF technologies continue to evolve, they can be automatically applied 
in mobile networks without any changes in 3GPP architecture/specification.

 

[..snip..]

 

   *  Simplified signaling - instead of seven-steps of separate N2/N4
      signaling from separate AMF/SMF to separate AN/UPF and N11
      signaling between AMF and SMF to set up the N3 tunneling for a PDU
      session, a two-step signaling between a new single control plane
      entity to the single integrated ANUP is enough - see Section 4.2
      for details.

 

Does the above text mean ANUP can be implementable with no change with 3GPP 
spec modifications?

 

On the other hand, I found another text as follow:

 

3.  MUP Evolution for 6G


   This section discusses potential MUP evolution in 6G mobile networks.
   It does involve changes in 3GPP architecture and signaling, so the
   purpose is to share the ideas in IETF/wireline community first.

 

This statement sounds reasonable and consistent with your conclusion, but 
contradicts the section 3.2 text. I was confused about this. So I would 
appreciate if you could correct me.

 

Best regards,

--satoru

 

On Sat, Oct 12, 2024 at 12:05 PM Tianji Jiang <[email protected]> 
wrote:

 

Hi, DMM Team:

 

For our I.D. 
(https://datatracker.ietf.org/doc/draft-zzhang-dmm-mup-evolution/), the DMM WG 
chairs initiated the WG adoption call in IETF-118 (Nov., 2023). And since then, 
we have made two presentations in IETF-119 and IETF-120. As of now, there have 
been many comments discussing the merits vs. concerns of the draft. This email 
is summarizing what we have in order to solicit the continuous discussions as 
well as push the WG adoption of the I.D.

 
WG Comments & questions vs. our replies & AIs, corresponding to the up-to-date 
revised draft, after the adoption-call and the IETF-118/-119/-120:
 
(From adoption-call) Comment#1: user-plane vs. control-plane
The draft seems to include inconsistent C-Plane discussions compared to what 
the authors argued. 
[Reply]:
We agree with one comment: CP does play a pivotal role in realizing the 
original DMM principles. The full potential of ANUP can be achieved upon the 
consideration of both CP & UP.
The work on user-plane & control-plane in the draft centers around the 
MUP/ANUP, i.e., the potential advantages with the integration of gNB & UPF. 
The discussions on ‘control-plane’ in the draft do not lay on the broader-sense 
aspects of 5G CP (e.g., RM, MM, SM, Paging, QoS settings, etc.). Nor do we 
intend to do so. Please see our replies to the following comment#2.
 
(From adoption-call) Comment#2: Regarding ‘SMF/N4, AMF/N2, Paging/RNA-AMF, etc.’
Sec.2 of "draft-zzhang-dmm-5g-destiributed-upf" assumed SMF exists, which could 
contradict with the ANUP proposal for simplifying signalling (N4 still 
required). In terms of paging, RNA paging over Xn has no IP addr info. If it 
has, it is different from the current. And it also assumes AMF exists after 
paging. It seems to contradict with ANUP concept (N2 still required too). If 
AMF doesn’t exist, it is different from the current.
Basically, this comment concerns this I.D. might step into too much into the 
3GPP territory, e.g., talking about architectures of the SMF/N4, AMF/N2, 
Paging/RNA-AMF, etc.
[Reply]:
First, we want to be crystal clear: as set forth at the beginning of the draft, 
the I.D. does not intend to do the 3GPP 5G/6G work in IETF. So, while we do 
list some critical 5GS CP features that might be related to this I.D., e.g., 
paging, CP signalling, etc., we have been acknowledging that the complete 
studies must be done in 3GPP SDO. 
Without doubt, 5GS is extremely complicated. While the MUP/ANUP might have 
certain advantageous impacts to the current 5GS CP, more will be explored along 
the time, with the main objective at the 6G evolution, once the draft is 
adopted.
In summary, we don’t intend to cross into the 3GPP territory. Just, if we do 
come up with novel thoughts during the normative work of the ANUP, we will 
liaise to 3GPP for possible advisory references.
[Some actions via I.D. revisions]:
Remove: 2.1 (O-RAN), 6.3 (Mobility HO), 6.4 (Paging)
o   This change is to address the concerns from more than one commenter that 
the I.D. should not step into the 3GPP domain. These subsections, e.g., paging, 
mobile HO, etc., are sort of 3GPP RAN & CN-CP specific. 
Add & Revise: 
o   A good scenario is added in the section#3 to justify the feasibility of 
ANUP: the solution proposals of the 3GPP Rel-19 SID (Satellite Phase-3) suggest 
the integrations of (a) CP registration/session establishment; and (b) UP 
gNB/UPF (eNB/MME).

§  Revise 6.11 (satellite), move & merge to the section#3. 

o   Move the LIPA section (was sec#5) to be a sub-section under the section#3, 
to demonstrate it’s being an ANUP-like feature already in 4G.
Also modified the subject lines of some sections to make the draft more 
readable and easier to follow: Sections#3, #4, including sub-sections.
 
(From adoption-call & IETf-120) Comment#3: implementation & realization on 
‘consensus’ claim
When it comes to rough consensus, should it be proved by the running code? – 
The DMM WG Chair has clearly stated that ‘the claim of reaching consensus among 
WG participants’ means, in the context of IETF, implementation must have been 
successfully coded.
o   Thanks to this ‘consensus’ claim, similar comment can be seen: "the focus 
of the draft should be changed to address details of realizing "ANUP""

[Reply]:

·         We recognize that the usage of ‘consensus’ might be a 
misrepresentation of our intention. Thus, we will correct the 'consensus' claim 
from now on.

·         Regardless, we have made it clear that this MUP/ANUP draft represents 
some thoughts from the IETF/wireline background, though the actual 
specification work needs to happen in 3GPP. That is, IETF does not do 3GPP 
standardization.

·         Further, in our common understanding,  ‘rough consensus’ does not 
necessarily require running code. As we know, many great solutions started 
without running code and not every WG requires running code before progressing 
documents. Running code does not mean rough consensus either - only multiple 
reputable interoperable running codes are meaningful.

 
(From IETF-119/-120) Comment#4: Regarding the subtle demarcation btwn IETF and 
3GPP
·         What kinds of advisory references could this IETF draft provide to 
3GPP? E.g., how this might be useful to the on-going 6G evolution (CP & UP)?

[Reply]:

·         This I.D. targets at offering to the imminent 6G evolution a possible 
reference design, out of numerous proposals, focusing on integrating the 
transport functionalities of both AN and CN. 

·         As we know, there are lots of research contributions revolving around 
the 6G network architecture. For example, an 6G paper published in IEEE 
Communication Magazine in July 2023, 6G Architecture Design: from Overall, 
Logical and Networking Perspective, analyzed various new scenarios, 
requirements as well as experience learned from the 5G deployment & operations. 
It proposed the holistic SBA and DAN (Distributed Autonomous Architecture) 
architectural principles for 6G, both of which talk about self-contained 
building blocks for AN and CN. Evidently, the integrated ANUP bears the 
potentials to offer transport advisory reference to 3GPP/6G evolution.

 

 
In conclusion:
This draft targets at 6G evolution: 
·         The 3GPP Rel-20 (6G roadmap) will start from the Q1 of Y-2025, a 
perfect timing for the exploration of the ANUP-like work, e.g., to investigate 
a possibly distributed AN, potentially integrated with (some components) of a 
(holistic) SBA CN, etc.

·         Further, many contributions are currently discussing the potential 
network architecture evolution for 6G, technical directions spanning both RAN 
and CN, new infrastructure impacts with TN-NTN integration, etc.

·         The draft mainly proposes enhancements and potential extensions for 
6G planning. In our thoughts, the ANUP work might result in some good advisory 
implications to 6G evolution, like: a flattened, simplified network 
architecture with the ANUP taking on both wireless (via air interface) and 
wireline (via IP network) connections.
Current status & proposed scope of the I.D. work: 
·         This draft represents work/thoughts from the IETF/wireline 
background. It has attracted significant interests and/or supports from many 
mobile operators & vendors.

·         While the contribution from the draft strives to offer advisory 
reference to 3GPP, the actual specification work needs to happen in 3GPP, which 
is out of the scope of this document.

 

 

Thank you and BR,

 

-Tianji

On behalf of all co-authors of the I.D.

 

 

 

 

 

 

_______________________________________________
dmm mailing list -- [email protected]
To unsubscribe send an email to [email protected]

_______________________________________________
dmm mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to