Re: [auth48] AUTH48: RFC-to-be 9262 <draft-ietf-bier-te-arch-13> for your review

[Alvaro Retana <[e-mail]>]

On August 30, 2022 at 2:35:55 PM, Lynne Bartholomew wrote:
> > On Aug 26, 2022, at 11:23 AM, Toerless Eckert wrote:


Lynne/Toerless:

Hi!


I've been thinking about this point:

> > (64) [rfced] Please review the "Inclusive Language" portion of the
> > online Style Guide at ,
> > and let us know if any changes are needed. Aside from "native", our script
> > did not identify problematic terms.
> >
> > I could not find on those URLs a place where "native" was listed as
> > problematic, where/what is the script you are using ?

I looked at several external references and found some discussions
mostly about using "native" to refer to specific groups of people, and
how to correctly do that.  But that doesn't apply here since the text
is not referring to people.


There are two places where "native" is used in the draft:

   If the BIER-TE topology represents (a subset of) the underlying
   (Layer 2) topology of the network as shown in the first example, this
   may be called a "native" BIER-TE topology.  A topology consisting
   only of "forward_routed()" adjacencies as shown in the second example
   may be called an "overlay" BIER-TE topology.  A BIER-TE topology with
   both forward_connected() and forward_routed() adjacencies may be
   called a "hybrid" BIER-TE topology.

   ...

   1.  Determine the desired BIER-TE topology for BIER-TE subdomains:
       the native and/or overlay adjacencies that are assigned to
       BPs.  Topology discovery is discussed in Section 3.2.1.1, and
       the various aspects of the BIER-TE controller's determinations
       regarding the topology are discussed throughout Section 5.



I agree with Toerless that there's no common well-known one-word
replacement for "native".  The computing related term that is often
suggested is "built-in", but that doesn't fit here.


OTOH, I wonder if we can come up with an alternative based on the
context, especially because there are only two occurrences and the
second one refers to the definition of the first.

The differentiation between "native", "overlay", and "hybrid"
topologies is the type of adjacencies: a "native topology" only has
forward_connected() adjacencies.  I think we can call this type of
topology a "connected topology" without any loss of meaning while
reinforcing the adjacency concepts.

Please consider this replacement.

Thanks!

Alvaro.