Re: [spring] Network Programming Interface for Provisioning of Underlay Services to Overlay Networks Using SRv6 (draft-xie-spring-srv6-npi-for-overlay)

[Andrew Alston - IETF <[e-mail]>]

Jingrong,

Limited Domain (in my view) refers to a domain where all points are under single administrative control.  I.E the source, destination and anything in the middle must fall under the same administrative control, otherwise, you are not within the limited domain.  This can be extended using encapsulation (tunneling) where the packet is encapsulated and preferably cryptographically protected, such that any intermediate networks outside of the same administrative control cannot, either by error or deliberate action, act on the information contained within the routing headers.

Anything else would run into issues with section 8.2 of RFC8402.

Thanks

Andrew


From: spring <[e-mail]> On Behalf Of Xiejingrong (Jingrong)
Sent: Monday, March 14, 2022 10:16 AM
To: Gyan Mishra <[e-mail]>; Andrew Alston - IETF <[e-mail]>
Cc: [e-mail]; Tom Hill <[e-mail]>
Subject: Re: [spring] Network Programming Interface for Provisioning of Underlay Services to Overlay Networks Using SRv6 (draft-xie-spring-srv6-npi-for-overlay)

Hi,

I think I now understand better the concern about "the use of SIDs over the public Internet".
In my previous mail, the SID2/SID3 used between CPE2-Internet-CPE3 is to explain the layering model (over vs across), but it is not the proposal of the draft.
The draft is talking about the interface (name NPI) that overlay networks use to access the underlying network in the last mile.
There may be an Access Network (AN) in the last mile, and the AN may also connect to Internet backbone and/or multiple underlying networks, but it is distinct from the "open Internet".
Make more sense if the AN can enhance (if no way to enforce) not to leak the SRv6 BSID to Internet backbone or other TNs ?

For the control plane, it is still not clear if a controller is required, but one thing I considered is to use an IETF standard protocol because the PE need to implement the NPI.

Regards,
Jingrong

From: Gyan Mishra [mailto:[e-mail]]
Sent: Sunday, March 13, 2022 8:26 AM
To: Andrew Alston - IETF <[e-mail]<mailto:[e-mail]>>
Cc: Tom Hill <[e-mail]<mailto:[e-mail]>>; Xiejingrong (Jingrong) <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>
Subject: Re: [spring] Network Programming Interface for Provisioning of Underlay Services to Overlay Networks Using SRv6 (draft-xie-spring-srv6-npi-for-overlay)

Hi Jingrong

I reads the draft and was trying to understand the problem statement as well as the solution.

So I believe the problem statement is how to interconnect desperate sites over the internet using a managed IPSEC VPN or SDWAN solution or managed MPLS and complexity of provisioning CE attachment.

The solution is an automated solution using SRv6 over the internet using BSID.  This involves running SRv6 over the internet, however SRv6 is limited to closed domains.  It appears an E2E pseudowire is used in provisioning the service.  Have you though of using NG L2 VPN EVPN all or single active multi home over SRv6.

Does all the provisioning use a PCE / SDN controller?


Thanks

Gyan

On Thu, Mar 10, 2022 at 9:59 AM Andrew Alston - IETF <andrew-ietf=[e-mail]<mailto:[e-mail]>> wrote:
Hi Jingrong,

I’m struggling to entirely understand this.  I think the question for me is – if you are sending packets with SID’s over the open internet – are you encapsulating those packets and is this encapsulation cryptographically protected – I.E the SID’s are not visible outside of the encapsulation, to preserve the limited domain.

Limited domains are typically extended via tunnel mechanisms, very often with cryptographic protection, hence the question

Thanks

Andrew


From: spring <[e-mail]<mailto:[e-mail]>> On Behalf Of Xiejingrong (Jingrong)
Sent: Thursday, March 10, 2022 9:40 AM
To: Tom Hill <[e-mail]<mailto:[e-mail]>>; [e-mail]<mailto:[e-mail]>
Subject: Re: [spring] Network Programming Interface for Provisioning of Underlay Services to Overlay Networks Using SRv6 (draft-xie-spring-srv6-npi-for-overlay)

Hi Tom,

Thanks for reading the draft and raise discussions.

In the proposal the SRv6 domain is the overlay network, belonging to one administrative domain -- the overlay network operator(say ONO).

For your concern about use of SIDs "across" the public Internet. Let me try to explain using following figure (hope it works):

CPE1 CPE2 CPE3
+ + + +
| +--------+ | | +----------+ |
+---[1] TN1 [1]---+ +---+ Internet |---+
+--------+ +----------+

In the perspective of the ONO, it has the following SIDs:
SID1/2/3: allocated on CPE1/CPE2/CPE3 by the ONO.
SID4/5: allocated by TN operator but serves for the ONO (Tenant-1 of TN, marked [1] in the figure).
The ONO can use these SIDs, and I would think they are all "in the overlay network", and are running "Over the Internet".

You mentioned in the last sentence "the use of SIDs over the public Internet". That is what I am modeling above.

Thanks
Jingrong


-----Original Message-----
From: spring [mailto:[e-mail]] On Behalf Of Tom Hill
Sent: Wednesday, March 9, 2022 10:43 PM
To: [e-mail]<mailto:[e-mail]>
Subject: Re: [spring] Network Programming Interface for Provisioning of Underlay Services to Overlay Networks Using SRv6 (draft-xie-spring-srv6-npi-for-overlay)

Hi Jinrong,

On 08/03/2022 01:58, Xiejingrong (Jingrong) wrote:
> I just posted a draft that specifies a framework and some more detail
> of the idea for provisioning of underlay services
> (Slice/SR-policy/Mcast/etc) to overlay networks(SD-WAN/CDN/etc), using SRv6.
>
> https://datatracker.ietf.org/doc/html/draft-xie-spring-srv6-npi-for-ov<https://datatracker.ietf.org/doc/html/draft-xie-spring-srv6-npi-for-ov>
> erlay
> <https://datatracker.ietf.org/doc/html/draft-xie-spring-srv6-npi-for-o<https://datatracker.ietf.org/doc/html/draft-xie-spring-srv6-npi-for-o>
> verlay>
>
> Please comment and send any feedback.
>
> I would like to discuss this document over e-mail/mail-list.


I'm concerned that this draft is explicitly violating the concept of
SRv6 as a protocol that operates within a Limited Domain.

As per Section 3.2 of this draft, "... the network operator of AN, TN and Internet can be different from each other."

Further, "In some scenarios, the AN can be an Internet exchange provider
(IXP) independent of ISP and NSP. In some other scenarios, the AN can be an ISP that running Internet backbone as well."

This would read to me that the proposal is explicitly intended to be inter-domain, and not at all limited to any one administrative domain.
Additionally, I cannot determine if the draft implicitly requires the use of SIDs across the public Internet?

Could I ask for some clarification on the scope of the draft, with respect to Limited Domains, and also the use of SIDs over the public Internet?

Kind regards,

--
Tom

_______________________________________________
spring mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/spring<https://www.ietf.org/mailman/listinfo/spring>

_______________________________________________
spring mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/spring<https://www.ietf.org/mailman/listinfo/spring>
_______________________________________________
spring mailing list
[e-mail]<mailto:[e-mail]>
https://www.ietf.org/mailman/listinfo/spring<https://www.ietf.org/mailman/listinfo/spring>
--

[图像已被发件人删除。]<http://www.verizon.com/>

Gyan Mishra

Network Solutions Architect

Email [e-mail]<mailto:[e-mail]>

M 301 502-1347