Currently there are at least 60 ISPs serving the San Diego county
area. There are LOTS of packets from "home" to "office" that make a
round-trip via MAE-West. Some people have decided that this is
silly. Even if it is "cost-effective", it *squanders* bandwidth at
MAE-West that could best be used for other traffic.
This is true in almost every major metro area. I help found an
exchange point in St. Louis. Not MAE-East levels of traffic, the
exchange peak at about 2.4 Mbps of traffic (more than a T1's worth). http://www.stlouix.net/ But that's 2.4Mbps of traffic that didn't
travel 2,000 miles to get across town.
Of course, not every local ISP participates. The state subsidized
education network doesn't connect, nor do some the dialup ISPs. But
it gets a reasonable level of support from several of the larger
area providers.
But exchange points are one of those weird creatures. If I'm paying
a big expensive backbone, why would I get anything from a local exchange
point? And of course, the ever popular "What's the catch?" Since
local exchange points are generally run on a non-profit basis, that
means there isn't a large marketing organization, or a huge gaggle of
salespeople trying to sell it. If you like, we can call it a "managed
connection" and charge you $1,000/month. But that seems steep for
essentially a port on a catalyst switch.
But we've found once an ISP connects, they generally keep it.
Of course, not every local ISP participates. The state subsidized
education network doesn't connect, nor do some the dialup ISPs. But
it gets a reasonable level of support from several of the larger
area providers.
The same type of project was attemted in Toronto. CANIX was essentially set
upto cross connect traffic rather than having to traverse the entire US
network to get
to the other side of Toronto. The problem was, it became an exclusive
bilateral peering
arrangemt with 6 players. That was 1 1/2 years ago. Currently only 2 are
peered. What in fact was the point. UUnet and Sprint were the big players up
here and nobody appears to want to cooperate.
But exchange points are one of those weird creatures. If I'm paying
a big expensive backbone, why would I get anything from a local exchange
point? And of course, the ever popular "What's the catch?" Since
local exchange points are generally run on a non-profit basis, that
means there isn't a large marketing organization, or a huge gaggle of
salespeople trying to sell it. If you like, we can call it a "managed
connection" and charge you $1,000/month. But that seems steep for
essentially a port on a catalyst switch.
Well, the problem is of course that local exchanges concentrate power
and make backbone packet transport into a commoditiy with machines
(routers) making buying decisions in realtime.
[ On Tue, May 26, 1998 at 15:38:24 (-0400), Damian O'Gorman wrote: ]
Subject: Re: different thinking on exchanging traffic
The same type of project was attemted in Toronto. CANIX was essentially set
upto cross connect traffic rather than having to traverse the entire US
network to get
to the other side of Toronto. The problem was, it became an exclusive
bilateral peering
arrangemt with 6 players. That was 1 1/2 years ago. Currently only 2 are
peered. What in fact was the point. UUnet and Sprint were the big players up
here and nobody appears to want to cooperate.
They're basically supplying rack space and an ethernet switch port in
Toronto's premier downtown communications interchange location (151
Front St.). All routing, peering, etc. is completely separate and
totally up to those who drag their own connectivity down to this
location.
I'm not involved in this -- just an interested bystander for now, but it
looks like it's unique in its conception to anything else I've seen
anywhere else to date.