Coop Peering Fabric??

David Diaz at
Tue Aug 12 08:58:57 CDT 2008


Love the Borg comment.

Great thread.  Old topic.  It recycles every couple of years.  Not to speak
for telx or Mike L but I do not think anyone was motivated to Borg anything
but to support AIX.  10Gig ports are expensive.

I like the idea of more exchange points in that they usually provide more
recovery pts and redundancy, allow the sharing of skills and knowledge in
the local community, and provide flexibility for growth and change of the
internet. How many COs do we have? There has long been the argument of how
many IXs are needed, would it be 1 per state?  What happens with Voip, IPtv

As for coops I think the argument is would the larger traffic players feel
comfortable connecting and making it a part of their networks?  Who are the
anchors and 1st movers?  What are the guarantees that any investment in
infrastructure needed to get there will be recovered over X years... Will
the coop fold before that pt? Wll it have the resources to upgrade.

I so not think a poison pill is needed. Perhaps just a group or company
championing Coops and giving them booth-space at events, sponsoring
conference travels, providing rack space etc.  But if it's in the BEST
interest of the members to have a larger group come in and take over then
what is the harm? What is the alternative, have members pay membership fees?
Corp Sponsorship?

I agree on much of this. But as with most things it comes down to money. Do
members have a financial incentive to join and what is the financial model
to keep the Coop moving forward as a success.

David D

On Aug 12, 2008, at 8:32 AM, Patrick W. Gilmore wrote:

On Aug 12, 2008, at 3:37 AM, Paul Wall wrote:

If it were as easy as you make it sound, I can assure you people would

be doing it.

People are.  I (and others) mentioned SIX & TorIX, plus I mentioned PaNAP.
 Then there's AtlantaIX, although that recently got slurped by TelX.
 (Hrmmm, could one of the "dangers" of a coop be "borg'ed by for-profit
entity looking to rip out every cent they can"? :)

Tons of others exist, in big and little markets.  There's one in 365 Main
SF, there's KleyReX in the same building as DE-CIX, Big APE in 111 8th, NYCx
there too, ChicagoIX just opened, etc., etc.

Trust me, it _is_ being done.

Also, does your Equinix MSA contain a non-compete clause, which could

be interpreted to mean you can't run a competing IX (metro fabric,

exchange, whatever) out of their facilities?  I hear many do.

So don't run it in an Equinix or S&D cage.




On Mon, Aug 11, 2008 at 11:15 PM, Deepak Jain <deepak at> wrote:

Warning: This may actually be operational too.

Given Cogent (and others) recent pursuit of sub $4/mb/s transit... and the

relatively flat cost of a "paid" peering fabric (even at 10G) and the O(N)

costs for cross-connects, the thought of revisiting the old peering coops

presented itself again.

Assuming 10G PNI model: Assuming even nominal cross-connect fees of

$100-$300/month per fiber pair, plus router port costs for each private peer

(assuming you aren't at >10% utilization on the port) at a commercial

exchange, you are eating a pretty significant cost per megabit you are

actually moving. (plug in your numbers here). Assumption: Above 1Gb/s

utilization, this makes sense or you are counting on growth.

Below 10% you would normally go to a paid peering fabric where you are

paying cross connect + a flat port charge + router port for 1->N peers and

hoping that enough utilization occurs that you get >10% utilization (to

recover capex, opex, etc) and then whatever additional utilization you need

to cover the flat port charge or you are counting on growth.

A "coop", best-effort switch fabric colo'd at a few sites would allow

participants to peer off traffic at a price of the order of a single

cross-connect (~$500/month per 10G port is possible, maybe less),

private-VLANs all-around, or to only-mutually approved peers (e.g. via an

automated web interface, prior art) to avoid many of the /old/ issues. No

requirement for multi-lateral peering. You could peer, sell transit, buy

transit, multicast, etc.

The way I figure it, it removes approximately an order of magnitude from the

operational cost of peering with more than a handful of your largest single

talkers. Especially as 100G LAN Ethernet becomes production before 100G WAN

connections become commonplace. Economic theory (assuming that worked on the

Internet) suggests this would allow for the increase in number of peers by

approximately an order of magnitude (maybe more).

Does this actually improve the present-day "rationale" to peer, or are most

operations' costs so far above (from long haul, etc) or so far below (since

the cost of transit has dropped so much) that this is no longer a relevant

part of the equation?

Warning: This may actually be operational too.

Deepak Jain


More information about the NANOG mailing list