Re: [mpls] [Technical Errata Reported] RFC5960 (2533)

Hi Stewart,
In your earlier email, you said "... but I don't think that is needed because it's an implication of the existing text".

I resonate with Huub's comment: perhaps these kinds of implications are evident to an expert, but for others an explicit statement is generally better than relying on the reader to fish out the implications.

Also, since the statement is implied by the existing text, the statement clearly does no harm.

Given that this was the subject of an ITU-T comment, incorporating it would be responsive to that comment, some readers see this as a helpful (explicit vs implicit), and the statement does no harm, I am not sure that I see the downside of accepting the statement.
Regards,
Steve

-----Original Message-----
From: mpls-bounces@ietf.org [mailto:mpls-bounces@ietf.org] On Behalf Of Stewart Bryant
Sent: Friday, October 08, 2010 11:11 AM
To: hhelvoort@chello.nl
Cc: mpls@ietf.org; BUSI, ITALO (ITALO); danfrost@cisco.com; adrian.farrel@huawei.com; RFC Errata System
Subject: Re: [mpls] [Technical Errata Reported] RFC5960 (2533)

  Hi Huub,

On 08/10/2010 09:42, Huub van Helvoort wrote:
> Hello Stweard,
>
> You replied:
>
>> Speaking as an author, my view is that the original text is correct, 
>> and the correction is not needed.
>
> As an author you are are an expert in this topic.
> However, for the less experienced you knowledge, i.e.:
The preamble was to alert the reader that I was speaking as an individual contributor, and not as an AD. I.e. this is a technical comment and not a management comment.

>
>> It's a fundamental of the IP protocol suite (which includes MPLS)
> > that the parser knows what protocol to expect from the previous
> > header or if there is no previous header from an identifier in
> > the lower layer.
>
> may not be that obvious.
>
> So adding this explanation or the text proposed by Ben:
>
> > If a section layer does not support multiplexing then it has
> > an implicit means of identifying the payload by the interface
> > over which the payload arrived.

Well there are a couple of problems with this. Firstly, other than in 
some implementations, a tunnel (i.e. an LSP) is not really an interface. 
Secondly if the packet arrives on a data-link, we do require the the 
type indicator.

- Stewart

_______________________________________________
mpls mailing list
mpls@ietf.org
https://www.ietf.org/mailman/listinfo/mpls