The pressure is on to use co-location service only from Big Players. Indeed, remember the big fight over Exodus peering arrangements? Someone (GTE?) decided that Exodus should pay them for transit and pulled peering. since no other large network pulled such stunt the result was that GTE customers were inconvenienced more than Exodus customers. The message is loud and clear. If you want your server farm to have good access, put it in a good co-location facility in the US run by (or connect your co-located equipment to) a very large provider who has good redundancy not only of their network as a whole but of their colo facility (a co-lo facility with only one WAN circuit does not have good redundancy even if the LAN is exceedingly good and fault-tolerant etc.). Outside the US, the filtering of long prefixes doesn't seem to be as much of a problem -- but nobody wants to see an announcement from their peer of a /16 broken into individual /24 all with the same next hop. Even when router memory and CPU capacity are not near limits, it is annoying. (sorry that the previous version went out in MIME format, I had replied to someone else's MIME formatted message). Dana
On Sun, 5 Dec 1999, Dana Hudes wrote:
The pressure is on to use co-location service only from Big Players. Indeed, remember the big fight over Exodus peering arrangements? Someone (GTE?) decided that Exodus should pay them for transit and pulled peering. since no other large network pulled such stunt the result was that GTE customers were inconvenienced more than Exodus customers. The message is loud and clear. If you want your server farm to have good access, put it in a good co-location facility in the US run by (or connect your co-located equipment to) a very large provider who has good redundancy not only of their network as a whole but of their colo facility (a co-lo facility with only one WAN circuit does not have good redundancy even if the LAN is exceedingly good and fault-tolerant etc.).
I'd disagree whole-heartedly (partly because I am not a huge, national tier-1). Wouldn't you rather connect your equipment to a smaller company, that is potentially more flexible, has more clueful people, has better pricing, and is multihomed to maybe 3 or 6 or 9 backbones?
No, but it depends on the capacity requirements. We looked into self-homed vs. colo. Given that; 1) Most eCommerce projects need to be completed inside of six months. 2) Connectivity needs to happen in the first 3 weeks of project kick-off. 3) Telco WAN circuit delivery, for large capacity, takes anywhere from 6 to 18 weeks per circuit (depending on RBOC ... could be MUCH longer). 4) Facility build-out takes even longer (3 to 6 months). For large capacity sites, colo is the only way, with potential self-homing within two years. It just can't happen faster than that. Also, smaller providers are out, because of public peering point congestion and that is usually their only avenue. Large providers, with their own private dark-fiber network, leaving only last-mile traffic to the public Internet, appears to be the only way to go <sigh>.
-----Original Message----- From: owner-nanog@merit.edu [mailto:owner-nanog@merit.edu]On Behalf Of Alex Rubenstein Sent: Sunday, December 05, 1999 9:54 AM To: Dana Hudes Cc: nanog@merit.edu Subject: Re: multi-homing
On Sun, 5 Dec 1999, Dana Hudes wrote:
The pressure is on to use co-location service only from Big Players. Indeed, remember the big fight over Exodus peering arrangements? Someone (GTE?) decided that Exodus should pay them for transit and pulled peering. since no other large network pulled such stunt the result was that GTE customers were inconvenienced more than Exodus customers. The message is loud and clear. If you want your server farm to have good access, put it in a good co-location
facility in the
US run by (or connect your co-located equipment to) a very large provider who has good redundancy not only of their network as a whole but of their colo facility (a co-lo facility with only one WAN circuit does not have good redundancy even if the LAN is exceedingly good and fault-tolerant etc.).
I'd disagree whole-heartedly (partly because I am not a huge, national tier-1).
Wouldn't you rather connect your equipment to a smaller company, that is potentially more flexible, has more clueful people, has better pricing, and is multihomed to maybe 3 or 6 or 9 backbones?
PS. The ONLY provider I can get to my Colorado Springs facility (Black Forest) is USWest/Qwest.
-----Original Message----- From: owner-nanog@merit.edu [mailto:owner-nanog@merit.edu]On Behalf Of Roeland M.J. Meyer Sent: Sunday, December 05, 1999 11:01 AM To: 'Alex Rubenstein'; 'Dana Hudes' Cc: nanog@merit.edu Subject: RE: multi-homing
No, but it depends on the capacity requirements. We looked into self-homed vs. colo. Given that; 1) Most eCommerce projects need to be completed inside of six months. 2) Connectivity needs to happen in the first 3 weeks of project kick-off. 3) Telco WAN circuit delivery, for large capacity, takes anywhere from 6 to 18 weeks per circuit (depending on RBOC ... could be MUCH longer). 4) Facility build-out takes even longer (3 to 6 months).
For large capacity sites, colo is the only way, with potential self-homing within two years. It just can't happen faster than that. Also, smaller providers are out, because of public peering point congestion and that is usually their only avenue. Large providers, with their own private dark-fiber network, leaving only last-mile traffic to the public Internet, appears to be the only way to go <sigh>.
-----Original Message----- From: owner-nanog@merit.edu [mailto:owner-nanog@merit.edu]On Behalf Of Alex Rubenstein Sent: Sunday, December 05, 1999 9:54 AM To: Dana Hudes Cc: nanog@merit.edu Subject: Re: multi-homing
On Sun, 5 Dec 1999, Dana Hudes wrote:
The pressure is on to use co-location service only from
Big Players.
Indeed, remember the big fight over Exodus peering arrangements? Someone (GTE?) decided that Exodus should pay them for transit and pulled peering. since no other large network pulled such stunt the result was that GTE customers were inconvenienced more than Exodus customers. The message is loud and clear. If you want your server farm to have good access, put it in a good co-location facility in the US run by (or connect your co-located equipment to) a very large provider who has good redundancy not only of their network as a whole but of their colo facility (a co-lo facility with only one WAN circuit does not have good redundancy even if the LAN is exceedingly good and fault-tolerant etc.).
I'd disagree whole-heartedly (partly because I am not a huge, national tier-1).
Wouldn't you rather connect your equipment to a smaller company, that is potentially more flexible, has more clueful people, has better pricing, and is multihomed to maybe 3 or 6 or 9 backbones?
[ i am most definitely not speaking for my daytime job. and note change of subject: ]
I'd disagree whole-heartedly (partly because I am not a huge, national tier-1).
<g>
Wouldn't you rather connect your equipment to a smaller company, that is potentially more flexible, has more clueful people, has better pricing, and is multihomed to maybe 3 or 6 or 9 backbones?
flexibility granted, and with it a more personal touch etc. this is good and can be a real benefit, especially to the smaller customer who needs the high-touch support. this discussion certainly has not demonstrated that smaller isps have more clueful people, and is not something i would recommend a small isp show to a prospective customer. while i often feel under-clued, i continue to be impressed at the level of clue at many of the biggest players. most have become soooo clueful that they no longer participate in discussions such as this <g>. given that tier-N, where N>=2, providers have to pay more for routing (be it transit or 'paid peering', which is just limited transit), the assertion that they have a price advantage would seem to assume that overall per-unit costs increase with scale, as opposed to decrease. as this goes against the models of most industries, it warrants explanation. the big providers connect to all other significant providers. and with very hefty bandwidth at many points. it's not a conspiracy, it's what we have to do to move the traffic. in the long run, this industry will be owned by a extremely few very large international players, all or almost all telephants. not that i like this, even though i helped build one of the eight tier one isps. it goes against my leftist upbringing. but i fear it is reality. i suspect that, in the long run, the only difference between the internet and the railroads, telegraphs, telephants, aerospace, etc. is that many more of the startup working folk have made real money from it. so someone goofed, but not so significantly that it will be noticed in the long run. randy
On Sun, 5 Dec 1999, Randy Bush wrote:
this discussion certainly has not demonstrated that smaller isps have more clueful people, and is not something i would recommend a small isp show to a prospective customer. while i often feel under-clued, i continue to be
Perhaps not...but in general I suspect the smaller the company, the easier it is to access the clueful people (assuming they have some of course). ---------------------------------------------------------------------- Jon Lewis *jlewis@lewis.org*| Spammers will be winnuked or System Administrator | nestea'd...whatever it takes Atlantic Net | to get the job done. _________http://www.lewis.org/~jlewis/pgp for PGP public key__________
Behalf Of Randy Bush, Sent: Sunday, December 05, 1999 11:26 AM
in the long run, this industry will be owned by a extremely few very large international players, all or almost all telephants.
I agree. The are many reasons for this, mostly dealing with sunk-cost infrastructure and capitalization, awa grandfathered rights of way. MHSC.NET lost its last residential customer, to the cable-modem guys, earlier this year. We just couldn't touch their price/bandwidth price-point without becoming a CLEC. I don't want to be a CLEC. MHSC remains an ISP, leaving the IAP business to someone else.
I know we've talked about this on the list before. Anyone know of neutral colos in either of these 2 cities ala Equinix type. It's interesting but most carriers and major ISPs wont go into anothers colo anymore. They are looking for something neutral even if it is a small colo. Any leads in these two cities would be appreciated. Happy Holidays, Dave Thank you, David Diaz Chief Technical Officer Netrail, Inc email: davediaz@netrail.net, davediaz@fla.net, cougar@mail.rockstar.org pager: 888-576-1018 NOC: 404-522-1234 Fax: 404 522-2191 ----------------------------------- Build 1: 46 cities nationwide -- COMPLETE Build 2: 80 OC48s Nationwide [no typo] ++ FAILURE IS NOT AN OPTION! ++ ------------------------------------
participants (6)
-
Alex Rubenstein -
Dana Hudes -
David Diaz -
jlewis＠lewis.org -
Randy Bush -
Roeland M.J. Meyer