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

[Alexander Azimov <[e-mail]>]

Hi Yunan,

Please see my comments below.

ср, 6 мая 2020 г. в 10:12, Guyunan (Yunan Gu, IP Technology Research Dept.
NW) <[e-mail]>:

> Hi Alex,
>
>
>
> Thanks for the elaboration, which totally makes sense to me. A suggestion
> here, the description of how to deal with the P2P relation using ASPA (just
> as what you described below) can be added to the draft for clarification.
>
>
>
> Further, Section 5.2 says:
>
> When route is received from provider or RS it may have both Upstream
>
>    and Downstream paths, where a Downstream follows an Upstream path.
>
>
>
> The “downstream” refers to both P2C path and P2P path, right? If so, I
> suggest to change the usage of “downstream” to “non-upstream” or similar
> wording to avoid confusion.
>
I do agree, this part isn't clear. I'll try to rephrase it.


> Another point to confirm with you about Section 5.2:
>
> Additional caution should be exercised when processing prefixes that
>
>    are received from transparent IXes, as they don't add their ASN in
>
>    the ASPATH.
>
>
>
> To my understanding, that in this draft, you assume by default that RS
> prepends its own ASN to the AS_Path, right?
>
No, there is no such assumption.


> And a final question about Section 7 Siblings (Complex Relations),
> regardless of the current description of what sibling relation is, do we
> think that sibling ASes are ASes that belong to the same operator? So any
> type of transition is free of charge?
>
Speaking about ASPA, we are speaking about peering relations, without any
guess about the business between parties.
And from what I know - siblings are also spread among small networks, so
this term is not strictly bound to the same administrative domain.