I will also add...
With the exception of one of Matt's earliest comments in this thread, some
thing similar to "Cogent is Poor Quality", I agree with just about
everything that he said.
(He uses Cogent more than he admits, but he's one of the guys that know when
and where to use and not to use them.)
I will even admit, that I just recently made the move to have Matt do
Transit for a large segment of our network. (1 of 2 providers)
I made that decission because, after much debate with him privately, I've
come to realize he's is probably the smartest person that I know on the deep
topics of BGP Peering and transit, and I'm confident that his network will
perform for us, as we need it to, to compliment our Cogent offering.
I will share why Cogent enabled us to be successful, and why we see Cogent
limiting us from being successful.
Cogent was an Excellent place for us to start out. They saved us loads of
money, becaues we could get the performance of being in a major carrier
hotel, but not actually have to be in the major carrier hotel, avoiding
those costly carrier hotel fees. They also had a PtMP fiber solution that
allowed a provider using lower bandwdith at their remote regions to share
capacity of a 100Mbps head end. I've never second guessed my decssion to
use Cogent, and I'd argue that that decession was one of the best decission
that I had made, leading to our competitive advantage. But times change, as
we grow....
This year we identified three flaws with Cogents model....
1) They are very rigid on everything being 100mbps increments. This would
translate to much higher bandwdith costs, as we upgraded our network to
speeds higher than 100mbps.
We are installing many 300mbps and GB links, although the average capacity
is rarely low still. The jump to 200mbps is to limiting compared to
burstable GB, and makes the transit to expensive when you are pushing only
110mbps and paying for 200mb. This becomes more of an issue when the intent
is to have two transit connection providing greater than 200mbps, when the
average use is low.
2) Everywhere we wanted to use different technical designs (Layer2,
Bursting) they were unable to acommodate us at the locations where we needed
it accommodated. Meaning they did not offer all the services remotely that
they offered in a carrier hotel.
3) We no longer needed cogent's PtMP benefits locally, because we could
already do it better and faster wirelessly, (Cogent Rings tend to funnel all
traffic to one point before re-routing, adding more points of possoble
failures apposed to direct wireless shots to where we wanted to go)
So our reasons, for migrating services off of Cogent were not performance,
they were stategic to find offerings that better matched our Wireless value
proposition, and that would allow us to save money, offering an overall
grander solution.
I believe anyone doing route optimization, and not having Cogent as one of
their provider, is undersighted. Because our experience has shown that
Cogent has had some of teh BEST performance backhauling data across the
country. But that doesn't mean that when they connect to other parties,
those parties have the best paths to the customer from that point. (all
sorts of Net Netrality issues involved). Where things change is when one
evolves to the stage where peering is necessary to support one's client base
adequately. Once you ahve a reason to be in the Carrier Hotel, there are
unlimited options, and many choices of near equivellent value. When one
starts out in peering, they tend to spread traffic out and no longer need as
large of pipes per provider, at least not for a while. The advantage of
being a small provider is that you are under the radar of teh big boys, and
can make more specific routing decssions, apposed to goliaths like Cogent
that have to make broad routing decissions, and constantly fighting with the
other Big Boys.
The one thing I ahve learned in this industry, without second guess is never
pay one penny more for something, longer than you have to, beyond a period
where you receive value for having paid for those services. Pay for it when
you need it. In the early stages, it made sense to simplify life, and have
a company like Cogent manage our IPs, and our Transit/BGP, they did it
better, when we were to small to justify the time and effort required to
manage it. But once one grows past that point, one develops the flexibilty
and scale to justify using each provider for the application that they are
best capable to deliver.
Tom DeReggi
RapidDSL & Wireless, Inc
IntAirNet- Fixed Wireless Broadband
----- Original Message -----
From: "Matt Liotta" <[EMAIL PROTECTED]>
To: "WISPA General List" <wireless@wispa.org>
Sent: Thursday, September 13, 2007 11:22 AM
Subject: Re: [WISPA] Alternate transport providers, Cogent
Tom DeReggi wrote:
Cogent's performance is great 95% of the time. Regardless of what anyone
says, when ever we see congestion problems, which does happen
occasionally, the congestion usually happens several hops after leaving
Cogent's network. I'd call this a problem with the other provider. It a
complicated game, stategizing who's network will carries the majority
burden and cost to backhaul the traffic the majority path. All I can say
is, when Cogent carries the data the majority path, the
performance/latency is better. We've had Cogent for 7 years and thats why
we've kept them. Many providers meet their price, They are not just the
lowest cost option anymore. Cogent tech support is also very good.
I think it is important to point out that Tom is connected to Cogent in
DC, which is the one place in the entire world where Cogent is fully
peered. The performance of Cogent in other cities will never be as good as
it is in DC until the Cogent peering situation changes.
-Matt
--------------------------------------------------------------------------------
** Join us at the WISPA Reception at 6:30 PM on October the 16th 2007 at
ISPCON **
** ISPCON Fall 2007 - October 16-18 - San Jose, CA www.ispcon.com **
** THE INTERNET INDUSTRY EVENT **
** FREE Exhibits and Events Pass available until August 31 **
** Use Customer Code WSEMF7 when you register online at
http://www.ispcon.com/register.php **
--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
WISPA Wireless List: wireless@wispa.org
Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless
Archives: http://lists.wispa.org/pipermail/wireless/
--
No virus found in this incoming message.
Checked by AVG Free Edition. Version: 7.5.485 / Virus Database:
269.13.16/1005 - Release Date: 9/13/2007 11:45 AM
--------------------------------------------------------------------------------
** Join us at the WISPA Reception at 6:30 PM on October the 16th 2007 at ISPCON
**
** ISPCON Fall 2007 - October 16-18 - San Jose, CA www.ispcon.com **
** THE INTERNET INDUSTRY EVENT **
** FREE Exhibits and Events Pass available until August 31 **
** Use Customer Code WSEMF7 when you register online at
http://www.ispcon.com/register.php **
--------------------------------------------------------------------------------
WISPA Wants You! Join today!
http://signup.wispa.org/
--------------------------------------------------------------------------------
WISPA Wireless List: wireless@wispa.org
Subscribe/Unsubscribe:
http://lists.wispa.org/mailman/listinfo/wireless
Archives: http://lists.wispa.org/pipermail/wireless/
