Hi all,
I have checked my archives and I see that I have indeed presented my concerns
at the SPRING WG session in
Vancouver<https://datatracker.ietf.org/meeting/120/materials/minutes-120-spring-202407262000-02>
and later clarified them in an email to the authors and the
WG<https://mailarchive.ietf.org/arch/msg/spring/CYZkkuxscmETRkD-4dtc2TbBZeQ/>.
This resulted in several email exchanges with the authors, but we have only
agreed to disagree.
I am still convinced that:
1. The endpoints of the lower layer link discussed in the draft MUST be
represented by IPv6-capable logical interfaces in order to be able to process
the remaining SSRv6 SIDs in the SRH (standard or compressed)
* Step S.14 in Section 3 of the draft explicitly supports the situation
in which the inter-layer SID is not the last SID in the stack
2. An already defined End behavior (that can represent an Adj-SID for an IGP
or a BGP peering segment for BGP) is sufficient for all the purposes of the
draft. Non sunt multiplicanda entia sine necessitate.
Regards,
Sasha
From: Alexander Vainshtein <[email protected]>
Sent: Monday, April 14, 2025 4:37 PM
To: Zafar Ali (zali) <[email protected]>;
[email protected]; Alvaro Retana <[email protected]>; SPRING WG
<[email protected]>
Cc: [email protected];
[email protected]; Zafar Ali (zali) <[email protected]>
Subject: Re: [EXTERNAL] [spring] Re: WG Adoption Call for
draft-dong-spring-srv6-inter-layer-programming
Dear colleagues,
I concur with Zafar.
I have already expressed my position on this draft in various private and
public discussions (including comments at the mike at the SPRING WG session at
one if the recent IETF meetings).
As Zafar has explained, the endpoints of the optical (or any other)
"inter-layer" link MUST be IPv6-capable so that they can handle IPv6 packets
correctly. This strongly suggests to me that a static Adj-SID (associated with
End.X behavior in SRv6) addresses all the needs I can think about.
And it is easy enough to prevent usage of the link in "shortest path" SRv6
paths.
The bottom line: I respectfully object to WG adoption of this draft because
frim my POV it does not meet the first of the two criteria for adoption:
1. Deals with a real problem
2. Represents a reasonable starting point towards solution of this problem.
Regards,
Sasha
Regards,
Sasha
Get Outlook for Android<https://aka.ms/AAb9ysg>
________________________________
From: Zafar Ali (zali)
<[email protected]<mailto:[email protected]>>
Sent: Thursday, April 10, 2025 4:24:46 PM
To: [email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>; Alvaro Retana
<[email protected]<mailto:[email protected]>>; SPRING WG
<[email protected]<mailto:[email protected]>>
Cc:
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>;
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>; Zafar Ali (zali)
<[email protected]<mailto:[email protected]>>
Subject: [EXTERNAL] [spring] Re: WG Adoption Call for
draft-dong-spring-srv6-inter-layer-programming
Hi Minxue
Thanks for your follow-up email.
Re: “The egress (and Ingress) of the underlay connection should also be capable
of L3 processing. It is just the connection between them is not L3.”
Can you please elaborate on if the ingress & egress are capable of processing
L3, why the link does not have L3 or L2 termination?
How do you “directly” take L3 packet (Srv6 encap) over an optical interface
(e.g., lambda)?
Re: “Regarding your suggestion of using BSID, the binding SID (H.Encaps or
End.B6.Encaps in SRv6) was used to instruct a node to encapsulate a new IPv6
header and SRH to the packet”
Of course, in the optical interlayer case, the BSID cannot encapsulate a new
IPv6 header and SRH to the packet.
However, it can hide the optical interface or possible interfaces (optionally
with the lambda value) behind the BSID construct or packet termination.
In your case, you can have an SR policy with single candidate path that
identifies the optical interface identified by “S” in line S15 of your
pseudocode.
Please have a look at an earlier work on this.
https://datatracker.ietf.org/doc/html/draft-anand-spring-poi-sr-08<https://datatracker.ietf.org/doc/html/draft-anand-spring-poi-sr-08>
Re: Your question on IP side debugging is not quite clear to me
Think about how you would debug END.IL where the packet forwarding happens on
the wrong optical interface.
Thanks
Regards … Zafar
From: [email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>
Date: Thursday, April 10, 2025 at 2:16 AM
To: Zafar Ali (zali) <[email protected]<mailto:[email protected]>>, Alvaro Retana
<[email protected]<mailto:[email protected]>>, SPRING WG
<[email protected]<mailto:[email protected]>>
Cc:
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>,
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>, Zafar Ali (zali)
<[email protected]<mailto:[email protected]>>
Subject: Re: Re: [spring] WG Adoption Call for
draft-dong-spring-srv6-inter-layer-programming
Hi Zafar,
Thanks for your interests and comments on this draft.
Regarding your question on whether existing SRv6 behaviors can be used, section
2 of this draft has shown the challenges in establishing L3 adjacency between
the two endpoints of the underlay connection. If it is not an L3 adjacency,
then SRv6 End.X behavior is not applicable, something new is needed for
indicating the forwarding instruction to an non-L3 underlay connection.
Regarding your question on the implementation, section 3 of this draft provides
specifications on how the layer-2 encapsulation information can be obtained.
With that, S15 can be implemented. S14 is executed on the sending side of the
underlay connection, which is capable of processing IPv6 header and SRH. The
egress of the underlay connection should also be capable of L3 processing. It
is just the connection between them is not L3. Actually there are already
implementations which proved the feasibility of this function.
Regarding your suggestion of using BSID, the binding SID (H.Encaps or
End.B6.Encaps in SRv6) was used to instruct a node to encapsulate a new IPv6
header and SRH to the packet, which is quite different from the expected
behavior in this inter-layer case, as no new IPv6 header or SRH should be added.
Your question on IP side debugging is not quite clear to me, you may want to
elaborate on it. To me the OAM of the inter-layer paths can be something
discussed in a separate document.
As a network operator who owns multi-layered networks, this function is needed
for efficient inter-layer path integration, and your contribution is welcome.
Best regards,
Minxue
________________________________
-------------------------------------
王敏学/ Wang Minxue
中国移动通信研究院 基础网络技术研究所 / China Mobile Research Institute
地址: 北京市西城区宣武门西大街32号创新大厦,100053
电话: 010-15801696688-33202
传真：010-63601087
Email: [email protected]<mailto:[email protected]>
-------------------------------------
From: Zafar Ali (zali)<mailto:[email protected]>
Date: 2025-04-09 07:02
To: Alvaro Retana<mailto:[email protected]>; SPRING
WG<mailto:[email protected]>
CC:
[email protected]<mailto:[email protected]>;
[email protected]<mailto:[email protected]>; Zafar Ali
(zali)<mailto:[email protected]>
Subject: Re: [spring] WG Adoption Call for
draft-dong-spring-srv6-inter-layer-programming
Dear author and the WG,
There was a lot of discussion on this draft, especially on the need for
defining "End.IL", which is the basis of the draft.
As far as I know the discussion was not closed and authors have not established
the need for defining "End.IL".
To keep myself honest, I will also respond to one of the original emails in
that thread.
I am happy to be corrected if a closure was obtained.
Comments from that discussion++;
Why a locally instantiated static adjacency SID cannot be used?
The reason given was this is a non-IP link but then the question is how I will
implement the following code in the (IPv6) packet path
S14. Update IPv6 DA with Segment List[Segments Left]
S15. Send the packet through the underlay network connection
identified by S.
S16. }
How would I implement S15.
To implement S15, I need some local construct to forward the digitally encoded
packet on the optical link S.
That local construct can very well be a locally instantiated static adjacency
SID.
It is also not clear how the receiving side processes the “optical signal” to
continue processing of the IPv6 packet (i.e., how to implement the receive side
of S14). Again, you need a packet termination endpoint for it to work.
• There was discussion on the packet termination part does not have IP
address associated with it.
o Use of unnumbered interface was suggested.
If the true need to “hide” optical interfaces behind “S” – use of BSID provides
much better construct for "abstraction" of optical network/ interfaces to
packet network was done here, as suggested in the following draft.
https://datatracker.ietf.org/doc/html/draft-anand-spring-poi-sr-08#section-5<https://datatracker.ietf.org/doc/html/draft-anand-spring-poi-sr-08#section-5>
The way the draft tries to hide optical interface looks like a Layer violation.
• How do I debug IP side if the END.IL is mis-forwarding – assume I can
implement it.
As the authors have not established the need for END.IL and hence the draft, I
respectfully object to the adoption call.
• For the reason mentioned above, I do not know how to implement End.IL
as it is defined or if it is at all needed (see comment above)
• I am happy to participate in the closure of any gap but in its
current state the draft is not ready for adoption.
Thanks
Regards … Zafar
From: Alvaro Retana <[email protected]<mailto:[email protected]>>
Date: Wednesday, April 2, 2025 at 3:06 PM
To: SPRING WG <[email protected]<mailto:[email protected]>>
Cc:
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>,
[email protected]<mailto:[email protected]>
<[email protected]<mailto:[email protected]>>
Subject: [spring] WG Adoption Call for
draft-dong-spring-srv6-inter-layer-programming
Dear WG:
This message starts a two-week adoption call for
draft-dong-spring-srv6-inter-layer-programming, ending on April/16. From the
Abstract:
Following the SRv6 Network Programming concept, this document defines
SRv6 based mechanisms for inter-layer network programming, which can
help to integrate the packet network layer with its underlying layers
efficiently.
https://datatracker.ietf.org/doc/draft-dong-spring-srv6-inter-layer-programming/<https://datatracker.ietf.org/doc/draft-dong-spring-srv6-inter-layer-programming/>
Please review the draft and consider whether you support its adoption by the
WG. Please share any thoughts with the list to indicate support or opposition
-- this is not a vote.
If you are willing to provide a more in-depth review, please state it
explicitly to give the chairs an indication of the energy level in the working
group willing to work on the document.
WG adoption is the start of the process. The fundamental question is whether
you agree the proposal is worth the WG's time to work on and whether this draft
represents a good starting point. The chairs are particularly interested in
hearing the opinions of people who are not authors of the document.
Thanks!
Alvaro (for the Chairs)
Disclaimer
This e-mail together with any attachments may contain information of Ribbon
Communications Inc. and its Affiliates that is confidential and/or proprietary
for the sole use of the intended recipient. Any review, disclosure, reliance or
distribution by others or forwarding without express permission is strictly
prohibited. If you are not the intended recipient, please notify the sender
immediately and then delete all copies, including any attachments.
_______________________________________________
spring mailing list -- [email protected]
To unsubscribe send an email to [email protected]
