Re: Persistent loops when mixing rtgwg-enterprise-pa-multihoming and rtgwg-dst-src-routing

[David Lamparter <[e-mail]>]

On Fri, Jul 28, 2017 at 02:42:59PM +0200, Matthieu Boutier wrote:
[snip]
> Perhaps having something like the following(?):
> 
>     3.  Forwarding tables representations
>     3.1.  Source Address Dependant Forwarding tables
>       -> this is just a dump of the announces
>     3.2.  Source-Prefix-Scoped Forwarding Tables
>       -> using "Generating Source-Prefix-Scoped Forwarding Tables v2"
>     3.3.  Examples

It's already there: if you look at -04 of the draft, your 3.1 is 3.1,
and your 3.2 is 3.4 ;)
In the -05 version I've moved this to the new appendix A.

Note as mentioned in my reply to Fred, we actually have 3 cases here:
- "d-s-ft" destination-first with fallthrough / backtracking
- "d-s" destination-first
- "s-d" source-first

The way the draft is outlined uses "d-s-ft" to specify expected
behaviour and has A.1 / A.2 describing translations to "d-s" & "s-d".


-David
-