Re: [mpls] Opinion sort: [Technical Errata Reported] RFC3107 (2319)

[<[e-mail]>]

Eric,

Thanks for reviewing the comments.
More inlined.

> 
> Adrian> Can I have the authors' and WG's opinion on Bruno's comments,
> Adrian> please.
> 
> It's possible to go through any document and find something that could
have
> been stated more precisely, but when a document has been around for
ten
> years without misleading anyone, it's usually better just to let it
be.
>
> The suggested fixes are no more correct than the original text.

So we both agree that the original text is not correct.

And I also agree with you that the first proposed text is not
significantly better.

> - The EBGP/IBGP distinction is not relevant here, 

Agreed: eBGP/iBGP distinction is not enough.

> as this does not properly
>   capture the notion of whether a BGP speaker is in the data path.


What we need is a Label Switched Path (LSP) toward the BGP _Next Hop_ of
the received routes.
The BGP speaker by itself is barely relevant.
So I see no basis for prohibiting the use of RFC 3107 based on the BGP
speaker (address).


> - The ability to send an MPLS packet from one router to another does
not
>   necessarily depend on there being either a sequence of MPLS routers
>   between them or on there being an L2 connection between them.  There
might
>   be an L3 tunnel, for example.
> 
> If you are asking whether the errata should be "accepted" (whatever
that
> means exactly), I'd have to say no.  That does not seem like a proper
way of
> making a technical correction, the correction isn't really correct
anyway,

So let's fall back to my second proposition, which was to remove the
sentence.
That would address your 2 points. 

> and it's hard to imagine anyone being seriously misled by this bit of
text.

What do you mean exactly?
- existing implementation do not comply with this sentence and hence
with the RFC?
- existing implementation comply with this sentence and hence restrict
the use of this RFC for no reason?


Regards,
Bruno