Re: [Sidrops] question on draft-ietf-sidrops-aspa-verification-04

[Alexander Azimov <[e-mail]>]

Hi Yunan,

Let say that that AS1 is the originator of the prefix, and has created ASPA
records. Since AS2 is a peer it will not be included in the list of
providers. The AS2 isn't creating any records + making the leak and AS3 is
performing ASPA verification procedure.
The 5.1 says:

   When a route is received from a customer, a literal peer, or by a RS
   at an IX, each pair (AS(I-1), AS(I)) MUST belong to customer-provider
   or sibling relationship


In our case, AS3 needs to check that AS1 -> AS2 belongs to c2p pairs. It
will check the existence of ASPA record for AS1 - it is present, then it
will check the presence of AS2 in the list - it will not be there, thus
making the path invalid.

Let me know if you have further questions.

ср, 29 апр. 2020 г. в 05:21, Guyunan (Yunan Gu, IP Technology Research
Dept. NW) <[e-mail]>:

> Hi Alex and Ruediger,
>
>
>
> To continue our discussion at the meeting, let’s use the example you
> provided yesterday. Both lateral peering between AS1—AS2, and AS2—AS3.
>
>
>
> First question, does AS1 or AS2 sign any ASPA object for this pair and
> how?  I see description for Sibling relation representation in Section 7,
> but haven’t been able to find any for P2P.
>
>
>
> Second question, if yes to the above question, how does AS 3 use any of
> the ASPA object(s) to detect the leak? And if no, how to detect the leak
> anyway?
>
>
>
> Thanks.
>
>
>
> Yunan
>
>
>
> [image: 10.1.1.0/24]
> [image: AS 1] [image: AS 2,AS 3]
>
>
>

-- 
Best regards,
Alexander Azimov