Re: [Idr] optional transitive

Hi Ilya,

Do you think that the below draft still does not formally address the 
problem you are describing ?

http://tools.ietf.org/html/draft-ietf-idr-optional-transitive-03

Cheers,
R.

> Hi,
>
> yesterday and few days earlier several people have observed their BGP
> sessions being reset after seeing certain prefix. After some evaluation
> I've found that offending prefix (in global internet table) had optional
> transitive ATTR_SET (code 128) attached. Not having first hand access to
> platforms that exhibited session reset I'm speculating that they've
> recognised attribute as such but found it to be inappropriate for given
> context and so treated update as malformed with all consequences.
>
> BGP specs seem to have only one place that could (with some stretching)
> justify such behaviour - from rfc4271#section-6.3 "...If an optional
> attribute is recognized, then the value of this attribute MUST be
> checked...". Wouldn't it be appropriate to add extra text that would
> clarify that optional transitive attributes out of context should not
> result in session reset?
>
> P.S.: as far as I understood problematic BGP implementation is fixed in
> later software releases, but lack of clarity in BGP specs leaves room
> for similar problems in the future
>
> Cheers,
> iLya
>
>
> _______________________________________________
> Idr mailing list
> Idr@ietf.org
> https://www.ietf.org/mailman/listinfo/idr
>
>

Re: [Idr] optional transitive - recognised but out of context