At 05:17 PM 5/20/98 -0400, you wrote:
I do so wish we could get over the "tier" fixation.
If I start the Tampa Bay Internet Exchange, let's say, and I haul in OC-3 links from the 5 top backbones, and DS-3's to the 4 NAP's, I can then (very likely) a) resell bandwidth to local ISP's for quite a bit less than the backbones could sell them a local drop, which would b) be quintuply redundant in cast of feed failure, and c) unload all the cross provider traffic from the NAP's, and indeed, the backbone itself.
That happened here in Columbus with a provider whom I won't name since they are very imfamous for threatening legal action at the drop of a hat. The fact of the matter is that they claim to be carrying some ungodly portion of the internet traffic on their "private network" thereby providing faster connections for their customers. Well, this is America folks. Lets face it. If there is opportunity for someone to make a quick buck, they will. These guys ended up with DS1's to the big-4, put in an OC3 circuit provisioned for a single DS3 to a local ISP that was a partner in the project and at the time had a 10MB partial DS3 circuit to one of the big-4 and advertised their network as "OC3 Fiber Optics based" and "Faster than a Teier One". Lets see how they are performing today, shall we? [somebox]:~# traceroute customer.of.blah-12345-we-are-a-nap.net traceroute: Warning: Multiple interfaces found; using x.x.x.x @ eth0 traceroute to customer.of.blah-12345-we-are-a-nap.net (y.y.y.y), 30 hops max, 40 byte packets 1 (default gateway) (x.x.x.1) 1.069 ms 0.454 ms 0.434 ms 2 (Tier One provider) (x.x.x.21) 1.456 ms 1.295 ms 1.268 ms 3 aads-nap.att.net (198.32.130.60) 42.083 ms 55.715 ms 51.440 ms 4 br2-a3110s2.cgcil.ip.att.net (192.205.31.122) 50.639 ms 54.973 ms 51.153 ms 5 br1-h30.sl9mo.ip.att.net (12.127.15.210) 59.831 ms 51.588 ms 44.678 ms 6 ar1-a300s1.sl9mo.ip.att.net (12.127.4.5) 42.652 ms 33.012 ms 35.015 ms 7 (y.y.y.y) 50.081 ms 42.371 ms 43.084 ms 8 customer.demarc.blah-12345-we-are-a-nap.net (y.y.y.y) 69.753 ms 74.542 ms OSPF? (y.y.y.z) 89.058 ms 9 router.customer.of.blah-12345-we-are-a-nap.net (y.y.y.y) 92.875 ms 87.486 ms 95.569 ms 10 www.customer.of.blah-12345-we-are-a-nap.net (y.y.y.y) 86.939 ms 58.300 ms 64.930 ms Wow. Now, that's some lightning speed. <sic> And NOW, lets look at someone with a real Tier One provider: [somebox]:~# traceroute www.mcs.net traceroute: Warning: Multiple interfaces found; using x.x.x.x @ eth0 traceroute to www.mcs.net (205.164.8.12), 30 hops max, 40 byte packets 1 (Default Gateway) (x.x.x.1) 0.495 ms 0.415 ms 0.385 ms 2 (Tier One provider) (x.x.x.21) 2.424 ms 1.630 ms 1.529 ms 3 aads.good.net (198.32.130.49) 15.131 ms 13.261 ms 13.930 ms 4 mcsnet.chicago.good.net (207.98.189.130) 14.775 ms 15.296 ms 16.193 ms 5 www.mcs.net (205.164.8.12) 14.592 ms 14.750 ms 14.504 ms [somebox]:~# The Tier one's who want the business will maintain adequate bandwidth for their downstreams. Time and time again, people who haven't been interested in investing in the facilities to provide Tier One access have cut corners to line their pockets at the expence of their downstreams who far too often have no clue and accept the poor performance of their circuit as "normal." BTW: Karl, lightning fast as always to your site. ------- John Fraizer (tvo) | __ _ | The System Administrator | / / (_)__ __ ____ __ | The choice mailto:tvo@EnterZone.Net | / /__/ / _ \/ // /\ \/ / | of a GNU http://www.EnterZone.Net/ | /____/_/_//_/\_,_/ /_/\_\ | Generation A 486 is a terrible thing to waste...
On Thu, May 21, 1998 at 12:44:11AM -0400, John Fraizer - Administrator wrote:
The Tier one's who want the business will maintain adequate bandwidth for their downstreams. Time and time again, people who haven't been interested in investing in the facilities to provide Tier One access have cut corners to line their pockets at the expence of their downstreams who far too often have no clue and accept the poor performance of their circuit as "normal."
Yes, John, I've been reading NANOG long enough to be aware that there are hundreds of ways to do it wrong. There are about a dozen ways to do it right, too. There are only about 40 ISPs in this area, OC-3 is probably overkill... not to mention whether it's actually _available_ reasonably, but you'll notice that's where I started. Cheers, -- jra -- Jay R. Ashworth jra@baylink.com Member of the Technical Staff Unsolicited Commercial Emailers Sued The Suncoast Freenet "Two words: Darth Doogie." -- Jason Colby, Tampa Bay, Florida on alt.fan.heinlein +1 813 790 7592 Managing Editor, Top Of The Key sports e-zine ------------ http://www.totk.com
On Thu, May 21, 1998 at 12:44:11AM -0400, John Fraizer - Administrator wrote:
At 05:17 PM 5/20/98 -0400, you wrote:
I do so wish we could get over the "tier" fixation.
If I start the Tampa Bay Internet Exchange, let's say, and I haul in OC-3 links from the 5 top backbones, and DS-3's to the 4 NAP's, I can then (very likely) a) resell bandwidth to local ISP's for quite a bit less than the backbones could sell them a local drop, which would b) be quintuply redundant in cast of feed failure, and c) unload all the cross provider traffic from the NAP's, and indeed, the backbone itself.
That happened here in Columbus with a provider whom I won't name since they are very imfamous for threatening legal action at the drop of a hat. The fact of the matter is that they claim to be carrying some ungodly portion of the internet traffic on their "private network" thereby providing faster connections for their customers. Well, this is America folks. Lets face it. If there is opportunity for someone to make a quick buck, they will. These guys ended up with DS1's to the big-4, put in an OC3 circuit provisioned for a single DS3 to a local ISP that was a partner in the project and at the time had a 10MB partial DS3 circuit to one of the big-4 and advertised their network as "OC3 Fiber Optics based" and "Faster than a Teier One".
Lets see how they are performing today, shall we?
[somebox]:~# traceroute customer.of.blah-12345-we-are-a-nap.net traceroute: Warning: Multiple interfaces found; using x.x.x.x @ eth0 traceroute to customer.of.blah-12345-we-are-a-nap.net (y.y.y.y), 30 hops max, 40 byte packets 1 (default gateway) (x.x.x.1) 1.069 ms 0.454 ms 0.434 ms 2 (Tier One provider) (x.x.x.21) 1.456 ms 1.295 ms 1.268 ms 3 aads-nap.att.net (198.32.130.60) 42.083 ms 55.715 ms 51.440 ms 4 br2-a3110s2.cgcil.ip.att.net (192.205.31.122) 50.639 ms 54.973 ms 51.153 ms 5 br1-h30.sl9mo.ip.att.net (12.127.15.210) 59.831 ms 51.588 ms 44.678 ms 6 ar1-a300s1.sl9mo.ip.att.net (12.127.4.5) 42.652 ms 33.012 ms 35.015 ms 7 (y.y.y.y) 50.081 ms 42.371 ms 43.084 ms 8 customer.demarc.blah-12345-we-are-a-nap.net (y.y.y.y) 69.753 ms 74.542 ms OSPF? (y.y.y.z) 89.058 ms 9 router.customer.of.blah-12345-we-are-a-nap.net (y.y.y.y) 92.875 ms 87.486 ms 95.569 ms 10 www.customer.of.blah-12345-we-are-a-nap.net (y.y.y.y) 86.939 ms 58.300 ms 64.930 ms
Wow. Now, that's some lightning speed. <sic>
And NOW, lets look at someone with a real Tier One provider:
[somebox]:~# traceroute www.mcs.net traceroute: Warning: Multiple interfaces found; using x.x.x.x @ eth0 traceroute to www.mcs.net (205.164.8.12), 30 hops max, 40 byte packets 1 (Default Gateway) (x.x.x.1) 0.495 ms 0.415 ms 0.385 ms 2 (Tier One provider) (x.x.x.21) 2.424 ms 1.630 ms 1.529 ms 3 aads.good.net (198.32.130.49) 15.131 ms 13.261 ms 13.930 ms 4 mcsnet.chicago.good.net (207.98.189.130) 14.775 ms 15.296 ms 16.193 ms 5 www.mcs.net (205.164.8.12) 14.592 ms 14.750 ms 14.504 ms [somebox]:~#
The Tier one's who want the business will maintain adequate bandwidth for their downstreams. Time and time again, people who haven't been interested in investing in the facilities to provide Tier One access have cut corners to line their pockets at the expence of their downstreams who far too often have no clue and accept the poor performance of their circuit as "normal."
BTW: Karl, lightning fast as always to your site.
Lots of this has to do with your constituents. If you're a dial-only provider, you can afford not to care (much) about latency. You ALWAYS care about packet loss, but latency can go to the dogs and your customers won't know much about it. If you serve DS1 customers, its a different ballgame. Now you HAVE TO provide good network infrastructure, or people consider your plant to "suck". Then again, my background and engineering efforts have been geared towards the Fortune-50 marketplace and OLTP environments where ANY packet loss or significant hysteresis in the transport times is considered unacceptable. We build 'em "a bit different" around here. -- -- Karl Denninger (karl@MCS.Net)| MCSNet - Serving Chicagoland and Wisconsin http://www.mcs.net/ | T1's from $600 monthly / All Lines K56Flex/DOV | NEW! Corporate ISDN Prices dropped by up to 50%! Voice: [+1 312 803-MCS1 x219]| EXCLUSIVE NEW FEATURE ON ALL PERSONAL ACCOUNTS Fax: [+1 312 803-4929] | *SPAMBLOCK* Technology now included at no cost
participants (3)
-
Jay R. Ashworth -
John Fraizer - Administrator -
Karl Denninger