Net Effects: H&R Block is rid of CompuServe. AOL owns CompuServe's
subscribers and is out of the business of deploying modems and routers.
WorldCom owns the networks built by AOL and CompuServe and sells network
services to AOL. The big get bigger.
Couldn't help noting that UUnet will run modems and routers
for AOL+Compuserve and it continues to do so for their biggest
competitor MSN (at least the last time I checked). Antitrust?
As a Web services provider we are already seeing AS701 as the
biggest consumer of our customer contents. Anyone care to share
their view of how the big boys rank in terms of traffic? IMHO
this does have just a tad bit more relevance to operations that
the original thread 
Regards,
Sanjay.
As a Web services provider we are already seeing AS701 as the
biggest consumer of our customer contents. Anyone care to share
their view of how the big boys rank in terms of traffic? IMHO
this does have just a tad bit more relevance to operations that
the original thread
sure, AS701 may do the most traffic, but when the most traffic is ~7%,
that's hardly an anti-trust suit waiting to happen. Even MCI, UUnet,
Sprint and Cerfnet (our four largest dest. ASes) combined are barely 25%
of all traffic...
Josh Beck - CONNECTnet Network Operations Center - jbeck@connectnet.com
Josh Beck wrote:
> As a Web services provider we are already seeing AS701 as the
> biggest consumer of our customer contents. Anyone care to share
sure, AS701 may do the most traffic, but when the most traffic is ~7%,
We obviously have different applications. More than half our hosting
clients are overseas businesses, most our traffic is outbound, and
almost
40% of content consumers look like they are behind uunet. Our traffic
is perhaps less localized, and more of it exchanged with a fewer number
of
larger, global providers, compared with yours. Naturally it is a cause
for concern when most content consumers are moving behind one provider.
I apologize for perpetuating this increasingly irrelevant thread on
this list. Will stop here. Thanks all for the private and public
feedback. I now realize that many providers are bound by non-disclosures
to not share their traffic distribution. And our traffic distribution
may be more unusual due to the specialized business we are in than
I had thought.
Regards,
Sanjay.
Josh,
Those are really interesting numbers. When you say
"destination AS", are you counting that as "next hop" or
"last hop"? I think the 40% number for 701 is counting
"all that is reached by 701", i.e. "next hop" as opposed
to that which is just within 701 i.e. "last hop"
I'd be interested, though probably not in this forum, to
hear what others have measured using these two different
techniques, as it helps us plan our networks better.
-scott
Those are really interesting numbers. When you say
"destination AS", are you counting that as "next hop" or
"last hop"? I think the 40% number for 701 is counting
"all that is reached by 701", i.e. "next hop" as opposed
to that which is just within 701 i.e. "last hop"
I mean last hop by that. The next hop numbers depend entirely on who else
we are connected to and our routing policy (which, incidentally, is to not
transit anything over uunet that we don't have to because they are _so_
slow these days).
I'd be interested, though probably not in this forum, to
hear what others have measured using these two different
techniques, as it helps us plan our networks better.
check out the cflowd data exporter thing (I can't remember the specifics
of it, because we only ran it a short while do to a bug in the
experimental IOS release)... it breaks down outbound traffic by
destination AS and exports the data, so you can see who you should buy
your next connection to, etc.
Josh Beck - CONNECTnet Network Operations Center - jbeck@connectnet.com