[Sidrops] Re: Ketan Talaulikar's Yes on charter-ietf-sidrops-01-05: (with COMMENT)

[Job Snijders <[e-mail]>]

On Wed, Jan 07, 2026 at 08:52:35AM -0800, Ketan Talaulikar via Datatracker wrote:
> ----------------------------------------------------------------------
> COMMENT:
> ----------------------------------------------------------------------
> 
> Is RPKI cache to Router protocol is covered under "RPKI Technology
> Stack"? Is there a definition of the components of this stack in some
> RFC? My understand is that RPKI is an infrastructure and includes
> those various components but not sure we can call it a stack.
> 
> I would be more comfortable if the word "stack" were to be replaced
> with the components of the RPKI instead (including the RPKI-RTR).

In my mind the RPKI-To-Router is a distinct component that should be
called out separately. Of note is that the RTR protocol itself does not
involve CMS, X.509, or ASN.1.

OTOH, very recent examples of "Maintenance of the RPKI technology stack"
are the clarifications on how to handle CRL Numbers (RFC 9829), or for
example the updates to Manifest Number handling in RFC-To-Be
draft-ietf-sidrops-manifest-numbers. Less recent examples are the
RFCs in which Manifests & ROAs were revised (RFC 9286 & RFC 9582).

The RPKI is a technology layered on top of other technologies: it makes
use of a restrictive X.509/CRL profile, a number of RPKI-specific ASN.1
profiles have been defined, it uses transport that wasn't invented here
(Rsync) and also transport that the old SIDR group specified (RRDP), and
so on.

I offer these comments as an attempt to clarify some of the thinking
that went into the charter from my side, and I recognize the challenges
in trying to capture a complex technology like the RPKI in a succinct
manner.

Kind regards,

Job