Re: [Idr] RFC-compliant use of Extended Length for new BGPattributes

["Samita Chakrabarti" <samitac at ipinfusion.com>]

>> I agree to all you say. However I would say all implementations
>> following RFC1771 strictly would break.
>
>Implementations that strictly followed RFC 1771 wouldn't have
>interoperated well with much of the Internet.  Consider the fact that
>AS_PATH length is not one of the inputs to route selection.
>
>Also keep in mind that RFC 2119 wouldn't be published for almost 2 more
>years after 1771.
>
>It's probably best to keep in mind that 1771 was known to have bugs -
>enough bugs for 26 consequent drafts to fix them.  There are still bugs
>in 4271 although their magnitude is probably significantly less on an
>interoperability basis.  Conformance to the spec to the exclusion of
>interoperating with other vendors may earn you friends with the test
>tool manufacturers but it certainly won't make your customers happy.

[SC>] Agree. Is there any interop event or any consortium that arranges for
RFC4271 or other BGP related interop event ?

Thanks,
-Samita


_______________________________________________
Idr mailing list
Idr at ietf.org
https://www.ietf.org/mailman/listinfo/idr