Hi Jeffrey,
I've already provided one of the key motivations in my previous email. There is
an operational challenge in the current mobility networks: there are no means
to have a fine-granular SLA across both the mobile and the transport network.
SRv6 is bridging this gap by integrating the underlay and the overlay control
from the source node (either the gNB or the UPF).
I am not aware of any standardized solution that solves this problem.
Therefore, I disagree on your statement "the only benefit of SRv6-replacing-GTP-U is
the MTU saving between SRv6 tunnel endpoints".
Cheers,
Pablo.
-----Original Message-----
From: Jeffrey (Zhaohui) Zhang <[email protected]>
Sent: lunes, 14 de junio de 2021 22:19
To: Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
As you have alluded to, we're not converging but that's ok.
My point is not to write a draft about SRv6-transported-GTP-U (there is no need for such
draft, just like we don't need another draft about "transporting foo" with
SRv6). My point is that the only benefit of SRv6-replacing-GTP-U is the MTU saving
between SRv6 tunnel endpoints.
There are other details that need to be taken care of in the draft, which we
have been discussing.
Jeffrey
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Monday, June 14, 2021 3:47 PM
To: Jeffrey (Zhaohui) Zhang <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Hi Jeffrey,
Inline with [PC6]. Many thanks for your explicit answers.
Cheers,
Pablo.
-----Original Message-----
From: Jeffrey (Zhaohui) Zhang <[email protected]>
Sent: lunes, 14 de junio de 2021 21:01
To: Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
Let me also copy (zzh2>, zzh3>) and explicitly add (zzh4>) my answers below.
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Monday, June 14, 2021 2:22 PM
To: Jeffrey (Zhaohui) Zhang <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Jeffrey,
Inline with [PC5].
Cheers,
Pablo.
-----Original Message-----
From: Jeffrey (Zhaohui) Zhang <[email protected]>
Sent: lunes, 14 de junio de 2021 19:32
To: Pablo Camarillo (pcamaril) <[email protected]>;
Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
If we look at the my original points:
[PC5] Jeffrey, I believe both of your points have been answered. I've copied
the relevant answers for your reference here below with PC, PC2 and PC4. Answer
does not differ. If any clarification is needed on my previous answers, please
let me know. Thanks.
-------
1. Since GTP-U can be transported over SRv6, which can also make use of TE capability and
used for service programing (NFS chaining), the only real difference between SRv6 tunnel
and GTP-U tunnel is that the UDP/GTP-U header is no longer needed in the SRv6 tunnel case
(in particular, the GTP-U TEID becomes part of the IPv6 address). With that, most of
" 3. Motivation" are not really applicable.
[PC] The integration of the overlay with the underlay SLA and service chaining
cannot be achieved with GTP-U.
[PC2] As documented in 5.2, the SID list imposed by the source node includes
segments for traffic engineering, NFV, and the overlay creation -all within the
context of one network slice-. All the intermediate nodes in the fabric are
stateless, hence you achieve highly scalable SLA.
[PC2] With an SRv6 transported GTP-U, the gNB only inserts the GTP-U header.
Then the cell site router needs to do a mapping of GTP-U sessions to SLA
policies. This implies that you need some mapping mechanism at the Cell Site
Router, that needs to have state. The gNB will not have any control of the
underlay path, ...
Zzh2> For SRv6-transported GTP-U, why can't the gNB itself put on the
appropriate SRH like in the SRv6-replacing-GTP-U case?
Zzh2> As quoted/discussed later, even in SRv6-replacing-GTU-U case, the gNB
does not get the SRv6 information from the AMF:
The gNB MAY resolve the IP address received via the control plane
into a SID list using a mechanism like PCEP, DNS-lookup, LISP
control-plane or others. The resolution mechanism is out of the
scope of this document.
Zzh2> With that, even if you have a separate device to do the mapping, why
can't that device do the same?
Zzh2> In other words, I still don’t see difference wrt SLA between
SRv6-transported-GTP-U and SRv6-replacing-GTP-U.
[PC3] If the gNB also adds the SRH as you suggest, then you are changing the
N3/N9 interface as well.
Zzh3> Wouldn't it be the other way around - SRv6-replacing-GTP-U is "changing the
N3/N9 interface"? SRv6-transported-GTU-U is still the same N3/N9 interface, just that
now they have SRH for traffic steering purpose.
[PC4] SRv6-replacing-GTP-U is a change in the N3/N9 interface.
Zzh4> I said in zzh2> that with SRv6-transported-GTP-U, gNB could also add traffic
steering via SRH - no different from SRv6-replacing-GTP-U. You said that's changing N3/N9,
and I replied in zzh3> SRv6-replacing-GTP-U is changing the N3/N9, which you confirmed
in [PC4]. Therefore, you have not really countered zzh2> above.
[PC6] There are no means to have fine-grained mapping of transport SLA with the
mobility SLA.
[PC6] Any transport protocol that encapsulates the gNB packet in the cell site
router cannot achieve fine-grained SLA due to scaling limitations and control
plane complexity.
[PC6] One option to solve this problem is to use SRv6 as a replacement to
GTP-U, whereas the gNB encodes the SLA in the packet. This is what this draft
is about.
[PC6] Certainly you write a new I-D to define a new N3/N9 interfaces that
bundle together the overlay and underlay protocols (e.g. the gNB pushes the
GTP-U header together with the MPLS label stack of the transport). I don't see
the benefits of such approach, and this draft is not proposing anything close
to it.
-------
Aggregation can be done for traditional GTP-U, SRv6 transported GTP-U, or SRv6
replacing GTP-U.
[PC4] I'm not aware that 29.281 allows aggregating N3/N9 traffic from different
TEIDs into the same TEID -despite from any local policy-. That said, this I-D
is not proposing a change in GTPU. Hence, any discussion on GTPU aggregation or
change to allow it in GTP-U is out of the scope of this draft.
-------
Zzh4> It's not clear why you would refer to 29.281 - since it certainly does
not talk about SRv6 either. Local policy means it's local behavior based on
unchanged N2/N4 - if you play SRv6 tricks for aggregation one can certainly play
other tricks for GTP-U based on local policy. I understand that aggregation in
GTP-U (via local policy) is outside the scope of this draft and I am not
suggesting to add that - my point is only that the aggregation benefit is not only
enabled by or exclusive to SRv6.
[PC6] I completely agree with you. Aggregation can be achieved with GTP-U. It
needs to be modified first though, as 29.281 -or any other 3GPP spec- does not
seem to allow it; while in the case of SRv6 we are writing an I-D that
explicitly allows it.
[PC6] Please let me know if there is any action item with respect to this
comment on this I-D. As said, GTP-U aggregation is out of the scope of the
document.
Zzh4> In summary, the benefit of SRv6-replacing-GTP-U is really just the MTU
saving between the SRv6 endpoints, and I have not seen a real counter argument
against that.
[PC6] I disagree with this comment for the reasons exposed above.
Zzh4> Jeffrey
[PC6] Please let me know whether there is any extra clarification needed.
Our exchanges had some clarifications on details, but I don't think the above
points have been invalidated.
[PC5] Please let me know whether there is any clarification needed on top of
the previous answers.
Jeffrey
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Monday, June 14, 2021 1:16 PM
To: Jeffrey (Zhaohui) Zhang <[email protected]>; Pablo Camarillo
(pcamaril) <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Hi Jeffrey,
Inline with PC4.
Pablo.
-----Original Message-----
From: dmm <[email protected]> On Behalf Of Jeffrey (Zhaohui) Zhang
Sent: lunes, 14 de junio de 2021 18:00
To: Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: Re: [DMM] Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
Please see zzh3> below. I don't agree with the two [PC3] points.
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Friday, June 11, 2021 10:17 AM
To: Jeffrey (Zhaohui) Zhang <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Hi Jeffrey,
Inline with [PC3].
Thanks,
Pablo.
-----Original Message-----
From: dmm <[email protected]> On Behalf Of Jeffrey (Zhaohui) Zhang
Sent: lunes, 7 de junio de 2021 15:58
To: Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: Re: [DMM] Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
Please see zzh2> below.
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Friday, June 4, 2021 12:53 PM
To: Jeffrey (Zhaohui) Zhang <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Inline with PC2.
Many thanks and have a good weekend.
-----Original Message-----
From: Jeffrey (Zhaohui) Zhang <[email protected]>
Sent: martes, 1 de junio de 2021 16:46
To: Pablo Camarillo (pcamaril) <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
Hi Pablo,
Please see zzh> below.
-----Original Message-----
From: Pablo Camarillo (pcamaril) <[email protected]>
Sent: Tuesday, June 1, 2021 7:23 AM
To: Jeffrey (Zhaohui) Zhang <[email protected]>
Cc: [email protected]
Subject: RE: Comments on draft-ietf-dmm-srv6-mobile-uplane-13
[External Email. Be cautious of content]
Hi Jeffrey,
Thanks for the reviews. Answers inline with [PC].
Thanks,
Pablo.
-----Original Message-----
From: dmm <[email protected]> On Behalf Of Jeffrey (Zhaohui) Zhang
Sent: martes, 1 de junio de 2021 4:33
To: Pablo Camarillo (pcamaril) <[email protected]>;
[email protected]
Subject: [DMM] Comments on draft-ietf-dmm-srv6-mobile-uplane-13
After many email exchanges and seeing two revisions (-12, -13), I now have a
much better understanding of the draft. Here my further
observations/understanding/comments.
Observations/Understanding:
A. This does *not* require *any* changes in the N2/N4 signaling.
AMF/SMF/gNB/UPF will still use existing signaling based on GTP-U, but gNB/UPF
will use either SRv6 or GTP-U tunnels based on local policy.
B. If both ends (gNB/UPF) of a tunnel use SRv6, no interworking is needed, and
two modes (traditional/enhanced) are defined. This is in section 5.1 and 5.2
respectively.
C. If gNB does not use SRv6 but UPF does, an SRGW is placed next to the gNB.
This is covered in 5.3 using enhanced mode example, but can be applied to
traditional mode as well.
D. If neither gNB nor UPF uses SRv6, GTP-U tunnels could still be changed to
SRv6 between two GWs - another SRGW is placed next to the UPF as well. This is
covered in 5.4 and referred to as drop-in mode.
[PC] All the observations look good, with a side note on A: Future documents
may define a new signaling.
Comments:
1. Since GTP-U can be transported over SRv6, which can also make use of TE capability and
used for service programing (NFS chaining), the only real difference between SRv6 tunnel
and GTP-U tunnel is that the UDP/GTP-U header is no longer needed in the SRv6 tunnel case
(in particular, the GTP-U TEID becomes part of the IPv6 address). With that, most of
" 3. Motivation" are not really applicable.
[PC] The integration of the overlay with the underlay SLA and service chaining
cannot be achieved with GTP-U.
Zzh> Can you elaborate how they're achieved w/ SRv6 replacing GTP-U and not
achieved with SRv6 transported GTP-U?
[PC2] As documented in 5.2, the SID list imposed by the source node includes
segments for traffic engineering, NFV, and the overlay creation -all within the
context of one network slice-. All the intermediate nodes in the fabric are
stateless, hence you achieve highly scalable SLA.
[PC2] With an SRv6 transported GTP-U, the gNB only inserts the GTP-U header.
Then the cell site router needs to do a mapping of GTP-U sessions to SLA
policies. This implies that you need some mapping mechanism at the Cell Site
Router, that needs to have state. The gNB will not have any control of the
underlay path, ...
Zzh2> For SRv6-transported GTP-U, why can't the gNB itself put on the
appropriate SRH like in the SRv6-replacing-GTP-U case?
Zzh2> As quoted/discussed later, even in SRv6-replacing-GTU-U case, the gNB
does not get the SRv6 information from the AMF:
The gNB MAY resolve the IP address received via the control plane
into a SID list using a mechanism like PCEP, DNS-lookup, LISP
control-plane or others. The resolution mechanism is out of the
scope of this document.
Zzh2> With that, even if you have a separate device to do the mapping, why
can't that device do the same?
Zzh2> In other words, I still don’t see difference wrt SLA between
SRv6-transported-GTP-U and SRv6-replacing-GTP-U.
[PC3] If the gNB also adds the SRH as you suggest, then you are changing the
N3/N9 interface as well.
Zzh3> Wouldn't it be the other way around - SRv6-replacing-GTP-U is "changing the
N3/N9 interface"? SRv6-transported-GTU-U is still the same N3/N9 interface, just that
now they have SRH for traffic steering purpose.
[PC4] SRv6-replacing-GTP-U is a change in the N3/N9 interface.
[PC4] SRv6-transported-GTP-U, if SRv6 is starting at the gNB and terminating at
the UPF, to me it is also a change in N3/N9.
[PC4] SRv6-transported-GTP-U, if SRv6 is the transport starting at the cell
site router, to me this is NOT a change in N3/N9.
[PC2] By the way, the limitations of SRv6 transported GTP-U are also present in
the Drop-in mode of this document.
[PC] Actually you have a lot more of motivation in this document:
draft-kohno-dmm-srv6mob-arch-04
Zzh> I had similar comments in that thread.
2. Besides the TEID, there are other parameters in the GTP-U header. How those
are represented in the SRv6 header needs to be defined.
[PC] draft-murakami-dmm-user-plane-message-encoding-03
Zzh> That should be a normative reference then.
[PC2] I don't think any of those is required to implement the I-D. I could
foresee an operator that is interested on this ID to also be interested on that
one, but I don’t think this is a reason to add the normative reference..
Zzh2> This document assumes that the 5G NFs will encode/decode SRv6 headers
based on signaled GTP-U parameters, or GWs will convert between GTP-U and SRv6
headers. If that does not warrant a normative reference to draft-murakami, I am
not sure what normative references are for. I'll defer this to chairs/ADs/IESG.
3. I still think there is no need to define the traditional/enhanced mode (see
my reasoning below).
4. Now it's not clear if the interworking mode (including the drop-in mode) is
worth the trouble (see my reasoning at the end).
[PC] Answers to 3 and 4 below.
Let me expand on #3 above.
"5.1. Traditional mode" focuses on the one-to-one mapping among (PDU session, GTP-U
tunnel, SRv6 tunnel) and casually mentions "SID list only contains a single SID".
" 5.2. Enhanced Mode" talks about two things: SID list for TE and service
programing/chaining, and scalability improvement via aggregation.
5.2 says:
The gNB MAY resolve the IP address received via the control plane
into a SID list using a mechanism like PCEP, DNS-lookup, LISP
control-plane or others. The resolution mechanism is out of the
scope of this document.
That means the use of SID list for TE and service programming is not per the
mobile architecture, but purely per operator's choice [PC] Indeed. The operator
is the one that decides whether traffic needs a particular SLA. As in
wireline...
, and it can also be used for both traditional mode and SRv6-transported GTP-U
tunnels - really nothing special to be limited to enhanced mode only.
It is true that some gNB may not be able to put on an SRH, but that equipment limitation should not
become a criteria - just like that for general SRv6 (outside this mobile user plane context) we
don't have "basic" vs. "advanced/enhanced" modes just because some devices
cannot insert SRH.
[PC] The motivation behind traditional mode is:
" The traditional mode minimizes the changes required to the mobile
system; hence it is a good starting point for forming a common
ground."
Zzh> Why did we not have "basic" vs. "advanced/enhanced" modes for SRv6 in
wireline networks 😊 There is no change to mobile system/architecture, except the transport network,
which is no different from wireline.
[PC2] Fair point. 😉 I understand your opinion, but after three years the
working group seemed to prefer the traditional mode. I note your preference.
Defining traditional/enhanced mode based on aggregation is more reasonable.
However, consider the following aspects:
- Currently only up link traffic can benefit from aggregation, and
that's only when the AMF provides the same <UPF address, TEID> for
multiple PDU sessions
- The same can be done for traditional GTP-U, SRv6 transported GTP-U, or SRv6
replacing GTP-U.
[PC] GTP-U does not allow that. Certainly you could go to 3GPP and change the
specs to allow it, but as per today it is not allowed.
Zzh> We established that there is no change to N2/N4 or mobile architecture: AMF/SMF
continue to signal GTP-U parameters like <tunnel endpoint, TEID> but gNB/UPF may
use SRv6 replacing GTP-U per local policy. With that, how is uplink aggregation achieved?
[PC2] The local policy indicates to aggregate several bearers into the same
SRv6 SID list.
Zzh2> For GTP-U, whether transported by SRv6 or not, traffic is routed based on
the UPF/gNB address - that is aggregation already. If finer granularity is needed,
similar local policy (to the ones used in SRv6-replacing-GTP-U case) can aggregate
smaller subsets of UEs, right?
Zzh2> Basically, I really don't see any difference.
[PC3] As stated above, certainly you change the current GTP-U spec to allow
aggregation as proposed in this I-D. However, the current GTP-U standard does
not allow it.
Zzh3> There is no change to GTP-U spec or N2/N4 interface. As stated above, it
is the local policy (when finer granularity is needed) that does the aggregation -
no different from SRv6-replacing-GTP-U.
[PC4] I'm not aware that 29.281 allows aggregating N3/N9 traffic from different
TEIDs into the same TEID -despite from any local policy-. That said, this I-D
is not proposing a change in GTPU. Hence, any discussion on GTPU aggregation or
change to allow it in GTP-U is out of the scope of this draft.
Zzh3> Jeffrey
Zzh2> Jeffrey
Zzh> Thanks.
Zzh> Jeffrey
The reason the aggregation is not applicable to downlink traffic is because the
gNB does not do IP lookup based on inner header.
[PC] Indeed that is one option. That is new compared to today's mobile
architecture. I don’t think its complex to implement though. End of the day
we've done it for quite some time in wireline. Another option is to forward to
a group of UEs and let the UE drop the packet based on the IPV6 DA.
Rather, downlink traffic forwarding on a gNB is purely based on TEID (whether
it is in the GTP-U header or in the SRv6 SID). Therefore, the uplink
aggregation or downlink aggregation (if gNB starts doing IP lookup based on
inner header, which would be a big departure from existing architecture) is
really controlled by the mobile architecture, not by the use of SRv6 tunnel. To
achieve aggregation, the AMF/SMF will need to signal different GTP-U
parameters, even though the signaling format does not need to change.
As a result, defining traditional/enhanced mode for SRv6 user plane really is
not necessary.
[PC] There is a motivation behind each of the modes. Operators find it easier
to deploy this way.
Now let me expand on #4 above.
There is no real difference between SRv6-transported GTP-U and SRv6 replacing
GTP-U (other than that the latter does not have UDP/GTP-U headers). If both
tunnel ends can support SRv6 natively, it's reasonable to use SRv6 tunnels
(replacing GTP-U) right at the tunnel ends. But if a gNB has to start/end with
GTP-U (with the UDP/GTP-U headers), what is the benefit of converting to/from
SRv6 by a GW, which means additional capex/opex? It's an additional failure
point - the implementation could have bugs and it could fail for various
reasons. It may be better off to only do SRv6 tunneling when both ends can
support it. It's not clear to me that the bandwidth saving between the GW and
UPF is worth the trouble.
[PC] Well... there are two differences between SRv6-transported GTPU and SRv6
replacing GTP-U. The first and obvious is the removal of the UDP/GTP-U header
-which already has huge benefits as the IPv6 Flow Label for entropy 😉-. The
second and most interesting difference is that the SRv6 replacement of GTP-U
allows the integration of the overlay with the underlay SLA and service
chaining. More in draft-kohno-dmm-srv6mob-arch-04.
These are high level comments. I may have more to come, and I definitely have
more text/wording comments to share afterwards.
Thanks.
Jeffrey
Juniper Business Use Only
_______________________________________________
dmm mailing list
[email protected]
https://urldefense.com/v3/__https://www.ietf.org/mailman/listinfo/dmm_
_;!!NEt6yMaO-gk!TuSYNLYkiLkieDF_Rvz9641_tQoKTcUOG-4qEG4eLshWb-mDfdVjr_
YwD9DLfzGU$
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
_______________________________________________
dmm mailing list
[email protected]
https://urldefense.com/v3/__https://www.ietf.org/mailman/listinfo/dmm_
_;!!NEt6yMaO-gk!Sx5QlZ5tYiUVPbgIXEQfo3TCPMtohG0NFBC42ie9MbEDb5emPpkAeW
rOJgJ1JgQ3$
Juniper Business Use Only
Juniper Business Use Only
_______________________________________________
dmm mailing list
[email protected]
https://urldefense.com/v3/__https://www.ietf.org/mailman/listinfo/dmm_
_;!!NEt6yMaO-gk!Sc4_uTC6g9WGXUpKJ2nlwgUzssL4oYfk82fb2KzDDkczbl6d2OPlrP
g8r6Qn-opA$
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
Juniper Business Use Only
_______________________________________________
dmm mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/dmm
