Re: [Sidrops] WGLC = draft-ietf-sidrops-aspa-verification - ENDS 02/17/2023 (Feb 17 2023)

["Sriram, Kotikalapudi (Fed)" <[e-mail]>]

Hi, Claudio and all,

Yes, please wait a few weeks and then read the upcoming version 12 (ASPA verification) draft for commenting.

I was puzzled as did some of my co-authors seeing the WGLC posted yesterday.

I have closely interacted with Claudio who has a nearly complete implementation of ASPA verification. Much of the proposed revisions for v-12 are motivated by his feedback.   

Claudio wrote: 
>I sent a mail on Nov 2nd with a large amount of feedback to this draft.
>None of it has been incorporated so a last-call for this draft is very
>premature.

Your Nov 2nd post: https://mailarchive.ietf.org/arch/msg/sidrops/KlILGgdItF-0O-xUsJde1DqGCoE/ 
My Nov 4th response: https://mailarchive.ietf.org/arch/msg/sidrops/cWcF2Y0msvko73zY912Y3cQj7SM/

Yes, your excellent and detailed feedback will be incorporated in v-12.

Claudio wrote:
>There are additional concerns I have with this draft after working on the
>only BGP implementation over the last weeks. I will send a follow up
>regarding my findings.

Yes, please send. In part, you may be referring to the IXP RS AS (transp./non-transp.) and RS-client discussion we had (via video chat). I posted about our discussion and conclusions here:
Jan 13, 2023 post: https://mailarchive.ietf.org/arch/msg/sidrops/JgxHTtynu-mihYnPxw1xgVGopPg/

So, yes those proposed changes about the part of the verification algorithm that deals with IXP RS & RS-client will also be included in v-12. 

Claudio wrote:
>1. Introduction 
>
>    When the CAS has multiple Providers, all Provider ASes are listed in the
>    ASPA including any internet exchange point (IXP) route server (RS) AS that
>    serves the CAS.

>This requirement is wrong and I will not approve any version of this draft
>which requires inclusion of "any internet exchange point (IXP) route server
>(RS) AS" in the SPAS.

Yes, as you and I discussed, that requirement will only apply to non-transparent IXP RS AS. We'll make that clear in both ASPA drafts.
This was also stated in my post of Jan 13: https://mailarchive.ietf.org/arch/msg/sidrops/JgxHTtynu-mihYnPxw1xgVGopPg/ 

Also, I plan to incorporate comments from Rich Compton:
   https://mailarchive.ietf.org/arch/msg/sidrops/zPXDR5iL52REtpIDoL6qdvJ2IsI/ 

Job asked:
>What is the status of ASPA in NIST BGP-SRx? 

I had posted about implementation of ASPA path verification in BGP-SRx in Dec. 2021:
https://mailarchive.ietf.org/arch/msg/sidrops/C6Ethp2aC9sJsxDtBrLQDjmCqk4/ 
BGP-SRx ASPA unit test cases/examples: 
https://github.com/usnistgov/NIST-BGP-SRx/blob/master/examples/example-demo-aspa-new/README 
BGP-SRx project Github page: https://github.com/usnistgov/NIST-BGP-SRx

We have not yet included the IXP RS portion of the ASPA verification algorithm in BGP-SRx implementation.

Thank you.

Sriram