Re: MAE-EAST Moving? from Tysons corner to reston VA.

On Fri, 16 Jun 2000, Mark Tripod wrote:
----- Original Message ----- From: "Richard A. Steenbergen" <ras@e-gerbil.net> To: "ted hardie" <hardie@equinix.com> Cc: "Wayne Bouchard" <web@typo.org>; <nanog@merit.edu> Sent: Friday, June 16, 2000 1:49 AM Subject: Re: MAE-EAST Moving? from Tysons corner to reston VA.
| I don't believe its nearly this bad. It seems to me that the VAST majority | of backbone private peering is still happening at the OC3 level. In a | public exchange point where you're talking to multiple networks across a | shared media it makes sense to do GigE, multiple GigE, 10GigE, etc, with | OC48 or > feeding it, but due to the nature of the beast (the fact that | you're exchanging traffic in multiple locations with multiple networks), | private peering is in no such immediate danger. | | Infact a lot of "backbone to backbone" peering is still done with DS3s | (think of the content providers with little need to talk amongst eachother | in bulk, GlobalCenter, Exodus, AboveNet, etc). The interesting traffic is | where the data providers (like those mentioned above) meet the data | suckers (@Home, AOL, etc), especially as streaming media takes off.
You'd be surprised by how much bandwidth is required between hosting companies. Companies that host at multiple locations and on different providers networks require that the connectivity between those networks be very robust so that they can transfer large amounts of data between their server farms.
http://stats.sjc.above.net/traffic/nyc/nyc-exodus.html http://stats.sjc.above.net/traffic/iad/iad-exodus.html http://stats.sjc.above.net/traffic/sjc2/sjc2-exodus.html http://stats.sjc.above.net/traffic/sjc/sjc-exodus.html http://stats.sjc.above.net/traffic/iad/iad-globalcenter.html Not nearly as much as one would thing, and if nothing else at least not mainly for that reason. Most of this bandwidth could probably be accounted for by the occational customer that doesn't fit the usual webhosting role and is pulling content from the "other" hosting companies. Though I'll be the first to agree that if they were serious about rhobust reliable service this would be more of the case. :P At any rate this is primarily burst traffic, and in the grand scheme of things it probably wouldn't account for a majority of traffic even if many more people were doing this. -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/humble PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)

I'm not sure that I'd use AboveNet as the benchmark for the hosting industry. ----- Original Message ----- From: "Richard A. Steenbergen" <ras@e-gerbil.net> To: "Mark Tripod" <mark@exodus.net> Cc: <nanog@merit.edu> Sent: Friday, June 16, 2000 7:32 PM Subject: Re: MAE-EAST Moving? from Tysons corner to reston VA. | http://stats.sjc.above.net/traffic/nyc/nyc-exodus.html | http://stats.sjc.above.net/traffic/iad/iad-exodus.html | http://stats.sjc.above.net/traffic/sjc2/sjc2-exodus.html | http://stats.sjc.above.net/traffic/sjc/sjc-exodus.html | http://stats.sjc.above.net/traffic/iad/iad-globalcenter.html | | Not nearly as much as one would thing, and if nothing else at least not | mainly for that reason. Most of this bandwidth could probably be accounted | for by the occational customer that doesn't fit the usual webhosting role | and is pulling content from the "other" hosting companies. | | Though I'll be the first to agree that if they were serious about rhobust | reliable service this would be more of the case. :P At any rate this is | primarily burst traffic, and in the grand scheme of things it probably | wouldn't account for a majority of traffic even if many more people were | doing this. | | -- | Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/humble | PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6) |

On Fri, 16 Jun 2000, Richard A. Steenbergen wrote:
this one isnt even in use. and remember, some agreements may not allow links to be posted. and from what you posted, 2 were OC3 and 3 were DS3s. It would seem to me that the move is towards OC3 and not DS3. Christian

On Fri, 16 Jun 2000, Christian Nielsen wrote:
On Fri, 16 Jun 2000, Richard A. Steenbergen wrote:
this one isnt even in use. and remember, some agreements may not allow links to be posted.
and from what you posted, 2 were OC3 and 3 were DS3s. It would seem to me that the move is towards OC3 and not DS3.
Obviously the move is towards OC3, and obviously some agreements do not allow traffic stats to be posted, but I think that most of AboveNet's are open (even if their http://stats.sjc.above.net/traffic page is woefully out of date). My point is that even on the OC3s, the traffic is rather low, and on that DS3 which is no longer in use the traffic was not always excessively high for a good portion of its lifetime. AboveNet doesn't even want to deal with the hastle of supporting DS3s, infact they will not do new DS3 private peers I believe. But my point is, in many cases the backbone to backbone (or at least hosting company to hosting company) peering traffic is hardly in the realm of exceeding our current circuit capacity. -- Richard A Steenbergen <ras@e-gerbil.net> http://www.e-gerbil.net/humble PGP Key ID: 0x138EA177 (67 29 D7 BC E8 18 3E DA B2 46 B3 D8 14 36 FE B6)
participants (3)
-
Christian Nielsen
-
mark@exodus.net
-
Richard A. Steenbergen