Hi Hang,

I don't think binding SID is an option here or future direction. Once the 
traffic lands in the Cloud Backbone, several principals and technologies are in 
place to maintain proper SLA on the traffic. Note that the Cloud Backbone is 
not a public Internet. When the traffic traverses through the Cloud Regions 
Backbone, they get the best services to maintain proper SLA. I don't think 
Cloud Backbone is a concern in maintaining the SLA.

Yes, when the traffic traverses through the public internet, there might be a 
question on the SLA. In this Draft, we used Geneve-based encoding, a popular 
option for carrying Metadata information. From the Draft standpoint, there is 
no such restriction if the Vendors want to use SR-TE on top of a Geneve header 
on the Public internet; that is Vendor's choice.

Though, with the MS Azure hat on, the better option is to use Azure Edge Pops 
from the nearest Branch location. The Express Route connections start from the 
Edge MSEE (Microsoft Enterprise Edge ExpressRoute) Routers. The Branch CPE 
traffic can enter Azure Cloud Backbone through MSEE Routers and use the Cloud 
Backbone for most of the part to maintain certain SLA. That way, most parts of 
the public internet can be avoided. Public documents are available on Azure 
Express Route Connections, please feel free to explore, or I can send you some 
links offline.

Happy to have more discussions in the IETF.

Thanks,
Kausik


From: Shihang(Vincent) <[email protected]>
Sent: Wednesday, July 12, 2023 1:29 AM
To: Kausik Majumdar <[email protected]>; Linda Dunbar 
<[email protected]>; Aseem Choudhary <[email protected]>; 
[email protected]
Cc: [email protected]
Subject: [EXTERNAL] RE: draft-dmk-rtgwg-multisegment-sdwan-00

Hi Kausik,

Got it. I agree the best option is to keep two administrative domains loosely 
coupled. Is binding SID (RFC 9256) such an option?

Thanks,
Hang

From: rtgwg <[email protected]<mailto:[email protected]>> On Behalf 
Of Kausik Majumdar
Sent: Wednesday, July 12, 2023 1:38 PM
To: Shihang(Vincent) 
<[email protected]<mailto:[email protected]>>;
 Linda Dunbar <[email protected]<mailto:[email protected]>>; 
Aseem Choudhary <[email protected]<mailto:[email protected]>>; 
[email protected]<mailto:[email protected]>
Cc: [email protected]<mailto:[email protected]>
Subject: RE: draft-dmk-rtgwg-multisegment-sdwan-00

Hi Hang,

The Cloud Backbone will not expose the detail of TE service to the outside 
world thus end to end SR is not an option. The best option to keep two 
administrative domains loosely coupled. How traffic is steered inside Cloud 
Backbone that is internal to any Cloud providers.

Thanks,
Kausik

From: Shihang(Vincent) 
<[email protected]<mailto:[email protected]>>
Sent: Tuesday, July 11, 2023 9:00 PM
To: Kausik Majumdar <[email protected]<mailto:[email protected]>>; 
Linda Dunbar <[email protected]<mailto:[email protected]>>; 
Aseem Choudhary <[email protected]<mailto:[email protected]>>; 
[email protected]<mailto:[email protected]>
Cc: [email protected]<mailto:[email protected]>
Subject: [EXTERNAL] RE: draft-dmk-rtgwg-multisegment-sdwan-00

Hi Kausik,

You says: In addition to what Linda mentioned, the TE or SRv6 is not an option 
within the Cloud Backbone.

I wonder why SRv6 is not an option within the Cloud Backbone. Do you mean that 
the Cloud Backbone will not use SRv6 for TE at all or it will not expose the 
detail of TE to the outside world thus end to end SR is not an option?

Thanks,
Hang

From: rtgwg <[email protected]<mailto:[email protected]>> On Behalf 
Of Kausik Majumdar
Sent: Wednesday, July 12, 2023 12:48 AM
To: Linda Dunbar 
<[email protected]<mailto:[email protected]>>; Aseem 
Choudhary <[email protected]<mailto:[email protected]>>; 
[email protected]<mailto:[email protected]>
Cc: [email protected]<mailto:[email protected]>
Subject: RE: draft-dmk-rtgwg-multisegment-sdwan-00

Hi Aseem,

Sorry for the late response. In addition to what Linda mentioned, the TE or 
SRv6 is not an option within the Cloud Backbone. The Cloud Backbone is a 
different administrative domain than Branch CPEs, and how the traffic steered 
within the Cloud Backbone that is not exposed to the Branch CPEs is internal to 
the Cloud. Hence, any SR-TE-based end-to-end mechanism doesn't work here. The 
Cloud Backbone might use some Traffic Engineering, but again that is internal 
to the Cloud Backbone on how they steer the traffic within their Backbone.

There is only one Egress GW. It is used as an optional Sub-TLV as a last GW 
within the Cloud Backbone, which is connected to the Destination Branch CPE.

On your 3rd point, the actual GWs within the Cloud Backbone are not exposed to 
the external world. That is solely control of individual Cloud Backbone. The 
best way to influence and encode the Cloud Regions is through Include /Exclude 
Transit Regions. That clearly dictates the intention that Branch CPEs want to 
influence the Transit Regions within the Cloud Backbone. The Cloud GW should be 
able to interpret these Sub-TLVs and can come up with possible paths to honor 
the preferred/de-preferred Regions. We haven't defined the format for these 
Sub-TLVs, but that might come in subsequent versions. In the initial 
implementation, it won't be implemented, most likely. It's the most advanced 
feature, but it gives us a scope to accommodate that in future to 
prefer/de-prefer the Regions within the Cloud Backbone.

Hope it helps. We plan to present this Draft in IETF 117. Please feel free to 
discuss more in the IETF.

Thanks,
Kausik

From: Linda Dunbar 
<[email protected]<mailto:[email protected]>>
Sent: Tuesday, July 11, 2023 7:53 AM
To: Aseem Choudhary <[email protected]<mailto:[email protected]>>; 
[email protected]<mailto:[email protected]>
Cc: [email protected]<mailto:[email protected]>
Subject: [EXTERNAL] RE: draft-dmk-rtgwg-multisegment-sdwan-00

Aseem,

Thanks for reviewing the draft. Answers to your questions are inserted below:

Linda

From: Aseem Choudhary <[email protected]<mailto:[email protected]>>
Sent: Tuesday, July 11, 2023 1:13 AM
To: 
[email protected]<mailto:[email protected]>
Subject: Re: draft-dmk-rtgwg-multisegment-sdwan-00

Fixed a typo.

From: Aseem Choudhary <[email protected]<mailto:[email protected]>>
Date: Sunday, July 9, 2023 at 9:33 PM
To: 
[email protected]<mailto:[email protected]>
 
<[email protected]<mailto:[email protected]>>
Subject: draft-dmk-rtgwg-multisegment-sdwan-00

Hello Authors,

Thanks for the document, Great work!

Having gone through the document, I have some questions/clarifications:


  1.  Section 3.3, it is mentioned SRv6/mpls-te not the best way. If there are 
multiple Cloud GW's and traffic need to be steered through for serviceability 
and performance, why SRv6 not an option?
[Linda] The draft proposes to use GENEVE header simply because wide adoption of 
GENEVE by cloud operators.
In addition, when the traffic from on-prem CPEs to Cloud GWs via the public 
Internet, TE and SRv6 is not supported by the Internet. Internet can only 
forward traffic based on the packets' destination addresses.


  1.  Section 4.5: Can there be multiple Egress GW Sub-TLV (or Next GW Sub-TLV) 
to steer traffic.

[Linda] The Egress GW Sub-TLV carries the information of the SD-WAN end point 
which is used by the egress GW to forward the traffic to.


  1.  Section 4.6/4.7: What is the best way to encode AZ/Regions? Is it 
possible to include/exclude specific Transit GW's?
[Linda] Probably will be "name" for the AZ/Regions, as most cloud operators do 
today, as the actual GW address of different AZ/Regions might be hidden from 
the end users.

I may have further comments.

-thanks,
Aseem


_______________________________________________
rtgwg mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/rtgwg

Reply via email to