Re: [Idr] IETF 73

On Nov 15, 2008, at 6:07 AM, Paul Francis wrote:
> The sense in which they are not orthogonal is this.  IACs are only  
> useful
> insofar that upstream ASes have a choice of path.  Of course, with  
> BGP it
> often so happens that upstream ASes have choices, but once any AS  
> makes a
> choice, it limits the number of choices it presents to its  
> upstream.  If you
> have tunnels, then more choices are presented upstream, and there  
> should
> therefore me more opportunities to make selections.  The mapped-bgp  
> spec
> offers even more choices than the tunnel-endpoints draft.
>
> So basically my point is that if you are prepared to modify BGP to  
> improve
> load balancing, you ought to consider the additional improvements  
> the tunnels
> can afford you.
>
> As for the tunnel-endpoint draft being more appropriate for RRG,  
> this is
> feedback I'd like to get.  Narrowly viewed, the tunnel-endpoint  
> draft is
> something that improves the performance of (intra-domain) VA, which is
> already work that is being done in IDR (and was explicitly turned  
> away by
> RRG).

Just my own 2 cents (with all hats off): before Dublin, the VA looked  
to me as something that, if adopted, could be deployed *right away*,  
so IDR seemed to me a better venue to push it--this is not so much of  
turning VA away per se.  As I mentioned to you privately, I always  
viewed VA as "poor man's map" in map-n-encap, which can serve as step  
stone towards incremental deployment (of solutions falling under map-n- 
encap).

>  Broadly viewed, the tunnel-endpoint draft has wider scope, which
> Raszuk pointed out, and this needs to be discussed in IDR.
>
> PF
>
>
>> -----Original Message-----
>> From: Iljitsch van Beijnum [mailto:iljitsch@muada.com]
>> Sent: Thursday, November 13, 2008 4:41 AM
>> To: Paul Francis
>> Cc: Inter-Domain Routing List
>> Subject: Re: [Idr] IETF 73
>>
>> On 12 nov 2008, at 13:16, Paul Francis wrote:
>>
>>> It's worth pointing out that if you convey inter-AS tunnel
>> endponts in
>>> BGP along the lines of
>>> http://tools.ietf.org/id/draft-xu-idr-tunnel-00.txt
>>> , then different ASes (or even individual routers) can select
>>> different paths, and so an IAC scheme can become more
>> fine-grained.
>>> Personally I think if we are going to look at something
>> like IAC, we
>>> should consider tunneling at the same time.
>>>
>>
>> Both the problems that these drafts address and the solutions
>> that they propose look completely orthogonal to me, so I
>> don't see how that would be helpful.
>>
>> About the tunnel draft: isn't this something that could
>> easily be part of the RRG efforts? Standardizing this now
>> while overlapping working is going on in RRG may be premature.
>>
>> Iljitsch
>>
> _______________________________________________
> Idr mailing list
> Idr@ietf.org
> https://www.ietf.org/mailman/listinfo/idr

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