Re: [Idr] draft-walton-bgp-hostname-capability-00

["Thomas Mangin" <[e-mail]>]

> Some of the vendors already have problems with the BGP code bloat.

Agree, let’s close the mailing list, we are all guilty here :-)

> And BTW - last time I checked, BGP was still limited to 4kB segment 
> and this space will all extended communities and so on is getting 
> crowded very fast.

This is the first argument I fully agree with .. The draft should indeed 
have included some wording that the capability not not be added should 
the OPEN message be larger than the allowed message size with it.

> If the improvement is for operations and not for core of BGP, let's 
> deal with this at operational/NMS level (and possibly protocols).

I am not fully un-sympathetic to that argument but why not let customers 
vote with their money by asking their vendors what to implement or not. 
Not all implementations include all the BGP RFC !

My issue is that without a Capability number, any implementation of the 
draft could later lead to a conflict and DO cause real interoperability 
issue.

I already pointed to Cumulus that the code they have in their Quagga 
implementation (131) was used by Cisco for Multi-Session before it 
became an RFC ! Mine (72 - which I intend to change too) is perhaps used 
elsewhere too. _That_ could cause a problem (more than anything 
discussed in this thread IMO) !

Thomas