CDNs should pay eyeball networks, too.

Dominik Bay db at rrbone.net
Tue May 1 19:31:52 UTC 2012


On 05/01/2012 09:17 PM, Leo Bicknell wrote:
> In a message written on Tue, May 01, 2012 at 08:23:07PM +0200, Dominik Bay wrote:
>> "Feeding" via some bigger peer networks oder classic transit
> 
> You have made the assumption that their choice is peering with your
> network or sending it out transit.  They may in fact peer with your
> upstream.

No, I also assume 'bigger peer networks', usually a smaller ISPs
upstream with a bigger regional footprint.

> That makes their choice peer with you, or peer with your upstream.
> Peering with your upstream may allow them to reach many people like
> you for cost of managing only a single peering session, as compared
> to maintaining a few dozen.

Agreed, more reach with a single session, but with pros there are cons,
as mentioned in my last mail. If they choose to accept these
limitations, that's fine for me.
But one should be aware, especially when aiming towards low latency,
high bandwidth connections to eyeballs.

> Also, many networks have minimum volume amounts for peering
> relationships.  They may be able to get settlement free peering
> with your upstream by having some minimum traffic level that they
> would not have if they peer with some of the individual customers
> behind that upstream.  Peering with you may drop them below the
> threshold, causing them to pay for transit on 10's of Gigs of
> traffic.

Agreed, but this is more a monetary optimization, not a technical
optimization.
I don't agree on any peering request, especially if this would force
traffic on paths which are preferred but sub-optimal.
A situation which wouldn't be optimal can be this one:

Say a network in Sweden and Estonia reach them via one of their
upstreams which do have a direct connection.
They are both a member of an IXP in, say, Amsterdam.
Both networks exchange their prefixes, and set localpref at this IXP.
This would make the path worse than before, as traffic takes a detour,
and might cost more due to backhauling traffic.




More information about the NANOG mailing list