UUNET Press Release on Peering

Tue May 13 01:28:43 UTC 1997

On Mon, 12 May 1997, Vadim Antonov wrote:

> 
> >I know of at least one.  I'm sure there are more.  It seems,
> >that equity in traffic exchange may be an overriding
> >principle, although running many web-farms and small
> >backbone can certainly give you that.
> 
> This is a serious misunderstanding of the economics of Internet.

No it's not.  It's a statement highlighting that UUNet's 
peering policy is not completely objective.  It also says
that if you run a web-farm it is easy to make outflow >>
inflow, and therefore any type of model-based just on 
traffic exchange differentials is also flawed.  I'm not
asserting that content should be an "overriding principal", 
but it certainly seems to be with UUNET.  Why?

Most networks send traffic over someone else's backbone 
and receive it on their own.  NSPs who focus on geographically
concentrated content providers can win over those that focus 
on a necessarily more dispersed consumer base. If the entire 
Internet were polarized into one of these two networks then it 
would obviously be a bad business decision on the part of either 
not to peer, at least in the short term.

>What is important is equal share in carrying packet flows, which
>can be approximately measured in miles*bps.  *That* costs money.

No, this is what is *fair* based on the economics of building 
networks with comparable user-bases.  What is *important* is not
to make a decision which will ultimately lose you money.

Mark