On Mon, Mar 7, 2011 at 8:58 AM, John G. Scudder <[email protected]> wrote:
> On Mar 4, 2011, at 5:39 AM, Christopher Morrow wrote:
> ...
>> A few folks noted that perhaps 'route' was not the right word here,
>> perhaps NLRI is. Using a wikipedia definition:
>
> I love Wikipedia, but the quoted definition is wrong, at least with respect 
> to the definitions given in RFC 4271:

I looked for the best rfc definition, missed 4271 :(  In the end I was
looking for the set of terms closest to what would be expected in the
IDR/SIDR/routing-protocols-implementors world.

> 1.1.  Definition of Commonly Used Terms
>
> ...
>
>   NLRI
>      Network Layer Reachability Information.
>
>   Route
>      A unit of information that pairs a set of destinations with the
>      attributes of a path to those destinations.  The set of
>      destinations are systems whose IP addresses are contained in one
>      IP address prefix carried in the Network Layer Reachability
>      Information (NLRI) field of an UPDATE message.  The path is the
>      information reported in the path attributes field of the same
>      UPDATE message.
>
> Given that what we're talking about securing is BGP, it would seem sensible 
> to use BGP's terminology.  Given that, the text as written seems right to me:
>
> ----------------------------------------------------------
> * Is the AS-Path represented in the route the same as the path
>      through which the route update traveled
> ----------------------------------------------------------
>
> or if you insist perhaps
>
> ----------------------------------------------------------
> * Is the AS-Path represented in the route the same as the path through
> which the NLRI traveled
> ----------------------------------------------------------

The original poster's note was an attempt, I think, to convey the fact
that a 'route' (prefix + path + attributes) was not actually whole
across all of the path. each bgp speaker it hopped through made some
judgment and added or subtracted bits to the 'route'.

NLRI seemed to capture just the prefix+path part of the 'route', which
ideally should not change at each hop (aside from having an AS added
to the path). The second form you have seems to keep that idea in
tact.

> But if this is the biggest nit we have to pick then I'll agree with "ship it".

yes... going to the AD's + list + housely now.

thanks!
-chris

> By the way, which Wikipedia article were you quoting, so I can think about 
> fixing it in my CFT?
>
> ...
>>   but still the material relevant to IDR should go there for 
>> checkbox/changes/etc.
>
> Yes.  Speaking for myself I don't have a problem with progressing this work 
> in SIDR for the time being given that there is energy and expertise to do it 
> here.  We can reshuffle deck chairs in the future as seems warranted, and in 
> any case should have review by both groups no matter which WG the work is 
> done in.
>
> --John
_______________________________________________
sidr mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/sidr

Reply via email to