a question about the economics of peering

Chrisy Luke chrisy at flix.net
Sat Dec 1 00:43:45 UTC 2001

up at 3.am wrote (on Nov 30):
> I thought the whole idea of "transit" was that you got connectivity to
> networks that you can't peer with.  Obviously. the packets can't tell if
> money changes hands or not.

Well, it's hardly "the point" of it. Transit is the purchase of all
the prefixes a network sees and the service of having your prefixes
advertised to all their peers. This much egg-sucking everyone knows.

Whether peering is the mechanism by which you

- Reduce transit costs/requirements
- Increase network performance
- Increase control of your network (exit point for traffic, choice beterrn
  use of peering/transit, etc)
- Increase the "value" of your sales proposition (ergo, by implication, the
  fact you have all of the above)
- boast to your mates down the pub how big your network is

is down to your personal/company view. 

To me, connecting (directly) to those networks which operate within your
market (eg, the UK, the EU, etc) is essential to be considered a viable 
player in said market.

To connect (directly) to those outside involves other factors, mostly to do
with either cost or marketability - if it's outside the place where your
customers are, the relative "feeling" of where the peers are is often
overshadowed by the cost/distance/whatever of getting traffic to/from that
place anyhow.

Within my markets, I'd pay (something) to peer with people. Outside, I
need to show I'm saving cash by doing so.


More information about the NANOG mailing list