Re: additional requirement: cooperation

[Joe Abley <jabley@nlri.org>]

On Mon, Nov 12, 2001 at 08:05:19AM -0800, Michel Py wrote:
> I think this one should be changed as follows:
> 
> "A multihoming strategy MAY require cooperation between a site
> and its transit providers, but MUST NOT require cooperation
> directly between the transit providers, EXCEPT when the transit
> providers are directly connected to one another".
> 
> "Directly connected" means that the traffic from/to the two
> concerned transit providers does not have to cross any AS other
> than one owned by either provider.
> 
> We don't want to require providers to peer or something with
> every other provider in the world, but some cooperation with
> other providers connected directly, which typically means a
> peering agreement or being in the same NAP is acceptable.
> 
> In other words, for the same reason some cooperation might be
> needed between a site and its transit providers, some might
> also be needed between a transit providers and its neighbors.
> 
> If unchanged, this sentence is very restrictive and would make
> impossible multihoming as it is done today (not that I recommend
> it..) because the necessary BGP peering between transit providers
> would be considered cooperation.

I think your modified version is more restrictive than the original.
Multihoming today does not require cooperation at all between
transit providers in the general case (although it can be
necessary to achieve certain goals, e.g. load sharing between
providers without PI address space).

Having said that, I think your changes are reasonable, although I
would substitute "site" for AS, since we have site defined and AS
suggests a routing solution.


Joe