At 04:13 PM 10/8/96 -0400, Richard Stiennon wrote:
At 03:41 PM 10/8/96 -0400, Scott Huddle wrote:
Isn't this the same mission as the vBNS?
University Internet Proposed
No, it sounds like a way for the Universities to upgrade from their T1 connections without making room in their budgets!
Sort of... It's a way for Universities to talk to _each_other_ at higher speeds without having to pay for the *ahem* value "added" by the Internet. Higher-Ed institutions want to talk to each other at higher speeds but don't necessarily want pay for their undergrads' (or professors' for that matter) ability to download nudie-GIFs 20x faster. So while there's no more money in the budget for faster Internet connectivity, many university provosts would be willing to spend more money for bandwidth if it were used for "nobler" pursuits. Basically, they want the Internet they had 6 years ago only faster. Internet II, if it happens, would be Higher-Ed's intranet. --zawada Paul J. Zawada, RCDD | Senior Network Engineer zawada@ncsa.uiuc.edu | National Center for Supercomputing Applications +1 217 244 4728 | http://www.ncsa.uiuc.edu/People/zawada
Sounds great. It all sounded great until it got to the part about "federal funding". I, for one, have a problem with my tax dollars going towards some professor being able to gawk at another professor in a videoconference. -BD On Tue, 8 Oct 1996, Paul J. Zawada wrote:
It's a way for Universities to talk to _each_other_ at higher speeds without having to pay for the *ahem* value "added" by the Internet. Higher-Ed institutions want to talk to each other at higher speeds but don't necessarily want pay for their undergrads' (or professors' for that matter) ability to download nudie-GIFs 20x faster. So while there's no more money in the budget for faster Internet connectivity, many university provosts would be willing to spend more money for bandwidth if it were used for "nobler" pursuits. Basically, they want the Internet they had 6 years ago only faster.
Internet II, if it happens, would be Higher-Ed's intranet.
Recently seen on NANOG:
Internet II, if it happens, would be Higher-Ed's intranet.
Sounds great. It all sounded great until it got to the part about "federal funding". I, for one, have a problem with my tax dollars going towards some professor being able to gawk at another professor in a videoconference.
One could easily name it "NSFNET V2", and it sounds as if they have declared lack of confidence in ISPs to solve their problems. If I were one of them, I might reason, "Why should I pay my lack-of-supprt, peering-point-packet-saturated, overcommitted, BGP-flapping, poor- cooperation, no-bandwidth-reservation ISP, when I can spend my money on this new group focussed on this higher level of service? I don't care if it costs more; it's not like I'm paying for the extra level of service directly. I want to move my research traffic onto this network before Bob's big Internet death happens." If anything, it shows that there's a need for a higher level of service than what's being provided to everyone jumping onto the Internet bandwagon. Who pays for it is arbitrary - whether it's the government or a Fortune 100 company, someone who sees value in it will pay for it. There's no question that service is good enough for the masses - watch them flee online service providers - but now there's a niche market for the quality of service once provided through NSFNET. At a recent Usenix, I attended a tutorial on Win95/NT programming (from a Unix perspective) where the lecturer commented a couple times, "One can complain all they want about how Windows is inferior to Unix, but there are those who see it as an opportunity to write software for Windows that provides the functinality they see lacking and make lots of money in the process." (It's not a direct quote, just my interpretation.) Likewise, people can complain that yet someone else is going to have a federally funded network, or as a hypothetical network service provider or telecommunications company I could fill the need by making a more reliable and feature-rich backbone to connect to and then take the money of the people who were going to build it anyway. I think some of the newer and/or larger ISPs are seeing this need, "Intranet" (a virtual private Internet begging the phrase "X.25 - the next generation"), and you might see one of them get the contract to serve what these universities want. BTW: I noticed after typing that this thread belongs on com-priv, yes? -- Eric Ziegast
It's a way for Universities to talk to _each_other_ at higher speeds without having to pay for the *ahem* value "added" by the Internet. Higher-Ed
Would this be such a Bad Thing(tm)? Many times I have seen research professors and students in the research field getting extremely upset because data that they used to exchange with other schools and countries is now taking 20 times longer because Joe Blow from L.A. is checking out all his favorite sports stories 300 times a day and John Dough is downloading all the porn he can fit on his 5 gig drive from across the country. I wouldn't really see this move as a "Internet Separatist" movement, more as a "return to normalcy" in the true spirit of the Internet. Consider if you were a biochemical research student at biochem.edu and you wanted to transfer a 30Meg molecular model back and forth between biochem.edu and chem.edu but in between both of you were hundreds and thousands of hosts, using Internet resources for commercial and entertainment purposes. Wouldn't you be a little upset when your ftp was finished at .098K/s over a multi-homed DS3? Ben
On Tue, 8 Oct 1996, Tersian wrote:
It's a way for Universities to talk to _each_other_ at higher speeds without having to pay for the *ahem* value "added" by the Internet. Higher-Ed
Would this be such a Bad Thing(tm)? Many times I have seen research professors and students in the research field getting extremely upset because data that they used to exchange with other schools and countries is now taking 20 times longer because Joe Blow from L.A. is checking out all his favorite sports stories 300 times a day and John Dough is downloading all the porn he can fit on his 5 gig drive from across the country.
You neglected to point out that Joe Blow and John Dough are both undergraduate students at the two universities in question and are using up the university's T1 that was virtually empty two years ago.
I wouldn't really see this move as a "Internet Separatist" movement, more as a "return to normalcy" in the true spirit of the Internet. Consider if you were a biochemical research student at biochem.edu and you wanted to transfer a 30Meg molecular model back and forth between biochem.edu and chem.edu but in between both of you were hundreds and thousands of hosts,
This *IS* the NANOG list, my friend. We now know you are a fool or a liar because we all know that TCP/IP is rarely configured to use more than 30 hops between two connections and most .edu sites in North America would have 15 or less hops. Several orders of magnitude less that the hundreds that you claim.
entertainment purposes. Wouldn't you be a little upset when your ftp was finished at .098K/s over a multi-homed DS3?
Sure, if I were a clueless biochem researcher I would be peeved. But if I were a clueful biochem researcher then I would realize that networking is not my specialty and I would be overstepping myself to make claims in the field. Thus I would seek out the networking specialists and ask them to examine the problem and determine why this is so slow. No doubt the problem would be traced to either local Ethernet congestion at the university due to poor network topology or an overloaded T1 line due to administrators who thought that WAN costs would be fixed for the next twenty years. If there are problems, we have the diagnostic tools to trace them down and find the true cause and then fix the true problem. There is no point in guessing because anybody who does this stuff for a living knows that if you have ten problems with virtually identical symptoms you will track them down to ten different root causes. It could be as simple as a faulty Ethernet card on the machine in the office down the hall causing spurious collisions, a broken router in a grossly overheated wiring closet, or water in an sloppily spliced copper cable. Michael Dillon - ISP & Internet Consulting Memra Software Inc. - Fax: +1-604-546-3049 http://www.memra.com - E-mail: michael@memra.com
Hi, I suspect that, as this -is- the NANOG list, there is a tendency to focus on tight, near-term technical issues. Just maybe, the problem is that commodity internet services are being priced for the 28.8 dialup Win95 user and his upstream provider. I would posit that those behind the Inet2 proposals are trying to involk a different price point. In some sense the higher ed. environment is the traditional ISPs worst nightmare. The dialup client is history. All your clients are running with ethernet or faster links to your core. DS3 links are a joke. OC3c is your nearterm (3-6 month) deliverable. You really want 10x OC3c, -clear-channel- pipes from a provider in the next 12 months. This is not your ISPs provider here. Multiple DS3s from anyone or multiple 100Meg ports on a GigaSwitch are not even in the cards. When will MCI/Sprint/WorldCom/ATT provide cost effective, clear channel OC48 pipes? The old NSFnet solved the DS3 issues for y'all. Who is going to take the lead on the next steps? --bill
Dorian writes: So if you follow that train of thought, building a private net for "important/meritorious" traffic makes some amount of sense. Cook: who chooses to define what is meritorious/important?
From what I understand Internet II will be brought to us by IBM using ATT bit pipes. The vBNS is a $50 million NSF investment in MCI. I *DO* wonder how synergy between the efforts is to be achieved?
Ironically I understand that it is MCI measured usage charging on DS-3s that has increased the cost of this bandwidth to universities by 300 to 400 per cent over what they had been paying that is increasing their anguish. Bill manning adds: When will MCI/Sprint/WorldCom/ATT provide cost effective, clear channel OC48 pipes? Cook: when hell freezes over is my guess. ************************************************************************ The COOK Report on Internet For subsc. pricing & more than 431 Greenway Ave, Ewing, NJ 08618 USA ten megabytes of free material (609) 882-2572 (phone & fax) visit http://pobox.com/cook/ Internet: cook@cookreport.com For case study of MercerNet & TIIAP induced harm to local community http://pobox.com/cook/mercernet.html ************************************************************************
On Tue, 8 Oct 1996 bmanning@ISI.EDU wrote:
Bill manning adds: When will MCI/Sprint/WorldCom/ATT provide cost effective, clear channel OC48 pipes?
Cook: when hell freezes over is my guess.
Manning: Thats what I heard a few years ago about DS3s.
Sprint and Nynex already have some ATM OC48 deployed. It's just a matter of time and customers with the money to pay for it. Michael Dillon - ISP & Internet Consulting Memra Software Inc. - Fax: +1-604-546-3049 http://www.memra.com - E-mail: michael@memra.com
On Tue, 8 Oct 1996 bmanning@ISI.EDU wrote:
Bill manning adds: When will MCI/Sprint/WorldCom/ATT provide cost effective, clear channel OC48 pipes?
Cook: when hell freezes over is my guess.
Manning: Thats what I heard a few years ago about DS3s.
Sprint and Nynex already have some ATM OC48 deployed. It's just a matter of time and customers with the money to pay for it.
As do MFS/WorldCom, ATT, TCG, PacBell, GTE, Cable&Wireless, ad nausia. What -is- interesting is that none of them have OC48c available, and none of it priced as such. All of it is priced at unit cost, e.g. a 64k voice channel. Muxing together smaller pipes into bigger ones the the game the telcos want to play. Of course there is the host interface problem... but Peter L. should have more to say there. -- --bill
On Tue, 8 Oct 1996 bmanning@ISI.EDU wrote:
ad nausia. What -is- interesting is that none of them have OC48c available, and none of it priced as such. All of it is priced at unit cost, e.g. a 64k voice channel. Muxing together smaller pipes into bigger ones the the game the telcos want to play.
Is it that bad to mux together 12 OC3's? In particular, since this research network sounds like they want to have guaranteed bandwidth, one way to do it is to manually allocate researchers and labs to different OC3's to some mystical, magical exchange point in a central location like Atlanta. Does anyone have a testbed network with such an exchange point operational? Michael Dillon - ISP & Internet Consulting Memra Software Inc. - Fax: +1-604-546-3049 http://www.memra.com - E-mail: michael@memra.com
On Tue, 8 Oct 1996, Michael Dillon wrote:
You neglected to point out that Joe Blow and John Dough are both undergraduate students at the two universities in question and are using up the university's T1 that was virtually empty two years ago.
Actaully, no, Joe and John were employees at major corporations that have bought T1's for their web servers so that they can sell a product and allow employees to browse the web when they are bored, or use it to find a better job.
entertainment purposes. Wouldn't you be a little upset when your ftp was finished at .098K/s over a multi-homed DS3?
Sure, if I were a clueless biochem researcher I would be peeved. But if I were a clueful biochem researcher then I would realize that networking is not my specialty and I would be overstepping myself to make claims in the
Actually if you were a biochem professor, clueless or not about networking, you would still be pissed off because of A.) all the undergrads downloading porn/warez while you try and get things done B.) The slow speed of the transfer as compared to a year ago. (Yes, the network connections are faster but they have not been upgraded to compensate for the massive influx of people) It's like saying, cars are faster today on the N.Y. freeways, which may be true, but there is still traffic jams and conjestion, so for the most part no one is going any faster. Plus, add 200% more people onto that freeway and you are going nowhere, I don't care if they are all in ferraris or if you widen the road two fold.
you have ten problems with virtually identical symptoms you will track them down to ten different root causes. It could be as simple as a faulty Ethernet card on the machine in the office down the hall causing spurious collisions, a broken router in a grossly overheated wiring closet, or water in an sloppily spliced copper cable.
I belive this goes further than the LAN, or even the local WAN setup. I am sure they could wire the whole school with FDDI and they would get super inter-campus performance, but once they step outside they are back at everyone elses mercy. Ben
On Tue, 8 Oct 1996, Tersian wrote:
Would this be such a Bad Thing(tm)? Many times I have seen research professors and students in the research field getting extremely upset because data that they used to exchange with other schools and countries is now taking 20 times longer because Joe Blow from L.A. is checking out all his favorite sports stories 300 times a day and John Dough is downloading all the porn he can fit on his 5 gig drive from across the country.
I wouldn't really see this move as a "Internet Separatist" movement, more as a "return to normalcy" in the true spirit of the Internet. Consider if you were a biochemical research student at biochem.edu and you wanted to transfer a 30Meg molecular model back and forth between biochem.edu and chem.edu but in between both of you were hundreds and thousands of hosts, using Internet resources for commercial and entertainment purposes. Wouldn't you be a little upset when your ftp was finished at .098K/s over a multi-homed DS3?
This argument is, if you will pardon the phrase, utter and complete bullshit. Before those annoying Joe Schmoe's threw their six-pack-guts up to the rail and started paying for their access, universities usually had what level of access to the net? 56k? Maybe T1? All feeding onto what, a T1 backbone? Commercialization of the network has brought a flood of resources to building out the backbones and making access cheaper for everyone, including the universities. You can't argue that congestion was not a problem before '94, because it was. You can't argue that the transatlantic links weren't always choking before '94, because they were. Commercialization of the 'net has made the vBNS, Internet II, etc., possible at OC-(whatever). Commercialization has pushed development of router, modem, Unix, IP software, etc., technology far faster than the universities ever could have pushed it. This elitish bullshit makes me want to puke. Pull your head out of your ass; it smells much better out here in the real world. __ Todd Graham Lewis Linux! Core Engineering Mindspring Enterprises tlewis@mindspring.com (800) 719 4664, x2804
On Tue, 8 Oct 1996, Todd Graham Lewis wrote:
This argument is, if you will pardon the phrase, utter and complete bullshit. Before those annoying Joe Schmoe's threw their six-pack-guts up to the rail and started paying for their access, universities usually had what level of access to the net? 56k? Maybe T1? All feeding onto what, a T1 backbone?
And I suppose you look down upon the Military for no longer dumping tons of time and resources into the Internet. But I guess thats okay with you because NOW commercial entities _control the net_ *spit*. I can't wait to see when your attitude hits Berkely. Commercial Sendmail should be nice. "What? You want a patch for sendmail? That'll be $200.00 for our service contract, plus .... etc."
Commercialization of the network has brought a flood of resources to building out the backbones and making access cheaper for everyone,
I don't dispute this, commercialization of the Internet was a needed move in the right direction for the Internet. What I am disputing is you blatant disregard for the research and development departments all around the world that developed the Internet. Where the hell would you be without Berkeley Sockets? So before you turn your nose up, look what's under it. You'll find that "your" Internet was built on alot of grants and volunteer work.
Commercialization of the 'net has made the vBNS, Internet II, etc., possible at OC-(whatever). Commercialization has pushed development of
Sure, it's helped development, but don't thumb your nose at who got you to this point. It's not elitism, it's efficiency, they want to go back to sharing resources trans-collegiate without all the garbage inbetween, I really don't see what your problem is.
Pull your head out of your ass; it smells much better out here in the real world.
What smells better? My ass or the real world? bah. Ben
On Wed, 9 Oct 1996, Tersian wrote:
On Tue, 8 Oct 1996, Todd Graham Lewis wrote:
I can't wait to see when your attitude hits Berkely. Commercial Sendmail should be nice. "What? You want a patch for sendmail? That'll be $200.00 for our service contract, plus .... etc."
You obviously missed my .sig
Commercialization of the network has brought a flood of resources to building out the backbones and making access cheaper for everyone,
I don't dispute this, commercialization of the Internet was a needed move in the right direction for the Internet. What I am disputing is you blatant disregard for the research and development departments all around the world that developed the Internet. Where the hell would you be without Berkeley Sockets?
I never contested the contributions that academia has made to networking. I contested the assertion that commercialization is responsible for the horrible state the network is in. 1) The network is not in a horrible state. 2) Whatever problems that are there, namely congestion, were there just as badly before. 3) Commercialization has contributed to the health if the network. To say that the network is worse off than it was before is just plain wrong.
So before you turn your nose up, look what's under it. You'll find that "your" Internet was built on alot of grants and volunteer work.
You mean like my contributions, however small, to Linux? Like Dosemu, GNUS, sendmail, all of which my coworkers have worked on?
Pull your head out of your ass; it smells much better out here in the real world.
What smells better? My ass or the real world?
I overreacted. Apologies. __ Todd Graham Lewis Linux! Core Engineering Mindspring Enterprises tlewis@mindspring.com (800) 719 4664, x2804
On Tue, 8 Oct 1996, Paul J. Zawada wrote:
It's a way for Universities to talk to _each_other_ at higher speeds without having to pay for the *ahem* value "added" by the Internet. Higher-Ed institutions want to talk to each other at higher speeds but don't necessarily want pay for their undergrads' (or professors' for that matter) ability to download nudie-GIFs 20x faster. So while there's no more money in the budget for faster Internet connectivity, many university provosts would be willing to spend more money for bandwidth if it were used for "nobler" pursuits. Basically, they want the Internet they had 6 years ago only faster.
Even with "recreational" use of the Internet, the Internet today is much faster then it was 6 years ago (T1 national backbone (singular)). MAE East and West traffic stats bear this out operating at around %30 capacity. The problem is ISP's who oversell backbones, Universities who overuse T1/T3 links. Given this, it would behove the Universities to form a private "intranet" and become thier own ISP. Of course, this model is an old one allready done by Suranet, etc.. who eventually went commercial for lack of funding when the next wave of technolgy arrived. In the end the benifit went to the equipment manufactures and the likes of AOL. Hopefully students will find new and fun ways to use Internet II - Then we can have Internet III?? Im very excited :) /stb --- Stephen Balbach "Driving the Internet To Work" VP, ClarkNet due to the high volume of mail I receive please quote info@clark.net the full original message in your reply.
participants (9)
-
bmanning＠ISI.EDU -
Bradley Dunn -
Eric Ziegast -
Gordon Cook -
Michael Dillon -
Paul J. Zawada -
Stephen Balbach -
Tersian -
Todd Graham Lewis