Erik, I couldn't agree more. BBN (GTEI) is only doing what Sprint started doing about 3 years ago, when some associates of mine tried to turn up peering with them. All these arguments about traffic flow were used then, by so-called "engineers" no less! One place I worked I had to answer trouble calls from people who couldn't get to our customers, and customers who couldn't get to specific sites, to a particular four-letter network I won't mention. Their customers were NOT glad to hear that the reason they couldn't get to our customers was because the four-letter network had refused to peer. But the status-quo has changed. Remember the charter of the FCC, to regulate "scarce resource" communications. What happens when some brain-dead congress-crawler starts getting calls from their constituents that "we bought internet service but couldn't get to xyz!" Some staffer takes a 30 second look into the problem, and decides that peering is *so* overwhemed that in order for everyone to reach everyone else the Government should step in and regulate it. The 'Net gives some measure of universal connectivity because it was driven at its root by engineering. Who on NANOG, with any real world experience in networking, denies that maximum open peering benefits everyone? This decision, and the arguments in favor of pay-per-peer, has nothing to do with engineering. It has to do with paper-pushers who wouldn't know the difference between peering and transit if it were on 500 power-point slides, because they look at MONEY. They see "bits per second", and who they came from, and conclude that those packets are not sourced on their network and therefore someone should be paying for them specifically. And, when the engineering breaks because a political answer is trying to be imposed on a technical problem, and Little Jonhy can't get to www.Bob.com for his school assignment, the cry will go up to "protect the children" and regulation will fall faster than Madame LaFarge ever thought possible. Then those same "pay-per-peer" morons will get appointed to the governing body of Internet Communications, just like the Railroad Barrons of the late 19th Century, and history will endlessly repeat, again. Thanks, Erik, for the pointer on the AT&T monopoly. Do you have a source? Curt- ----- Begin Included Message -----
From: "Erik E. Fair" <fair@clock.org> Subject: Re: BBN Peering issues Sender: owner-nanog@merit.edu Content-Length: 1119
There is a customer perception, dating from the earliest days of the Internet that when you connect to the Internet, you will be able to reach all sites that are up, everywhere. That this is still mostly true is a tribute to the hard work of a lot people on this list, and elsewhere. So far, the cases for which this is not true are small in both number and relative importance. If this perception breaks down, watch out. Theodore Vail was allowed to create the regulated monopoly AT&T in the early part of this century on the promise of Universal Service, which meant not only that everyone had a telephone, but that *all* telephones could call *all* other telephones - one big, happy, PSTN. The Internet presents this kind of universality today without the regulation, but don't doubt for a second that if the ISPs (of whatever size) begin destructive pissing matches of the form "I'm bigger than you, pay me or we disconnect" that the FCC will be pressured to regulate the ISPs in such a way to guarantee the universal connectivity aspect of the Internet. Your customers will demand it. Erik <fair@clock.org> ----- End Included Message -----
Some staffer takes a 30 second look into the problem, and decides that peering is *so* overwhemed that in order for everyone to reach everyone else the Government should step in and regulate it.
The 'Net gives some measure of universal connectivity because it was driven at its root by engineering. Who on NANOG, with any real world experience in networking, denies that maximum open peering benefits everyone?
You are missing the point entirely. We are not in some idyllic r&d environment where j.random hacker gets to play with his new ip router and maybe even connect two computers together. We are motivated by profit. If a certain position can maximize the profits of a company, it will rationally move towards that position.
This decision, and the arguments in favor of pay-per-peer, has nothing to do with engineering. It has to do with paper-pushers who wouldn't know the difference between peering and transit if it were on 500 power-point slides, because they look at MONEY. They see "bits per second", and who they came from, and conclude that those packets are not sourced on their network and therefore someone should be paying for them specifically.
You play engineering and revenue as a dichotomy when in fact it is a duality. In the "real world" that you like to refer to, it is more of a symbiotic relationship. Though it would be nice to be aloof in a lab and be lavishly compensated, I would venture to say that such a situation is not the norm. Let us get one point understood. You deserve nothing more than you pay for. You were not, by merit of your diction, infrastructure, prowess, or mandate from heaven, conferred a right to the resources of another individual or conglomeration of individuals. If you abhor these paper-pushers so, you are more than able to find the capitalisation to finance a global altruistic network where anyone can peer with anyone and we will all get along. It will be akin to the communes of the 60s. We can just hallucinate until the reality evaporates. BR
Good greaf, Bradley, who pissed in your cheerios?
The 'Net gives some measure of universal connectivity because it was driven at its root by engineering. Who on NANOG, with any real world experience in networking, denies that maximum open peering benefits everyone?
You are missing the point entirely. We are not in some idyllic r&d environment where j.random hacker gets to play with his new ip router and maybe even connect two computers together. We are motivated by profit. If a certain position can maximize the profits of a company, it will rationally move towards that position.
It will be interesting to watch what happens. Yes, charging for peering, except of course with real "Peers" such as the top 5 other networks period, will ensure that either there are only 5 networks after a while, or that the customers of those that try to charge for peering go elsewhere. Yes, it will be "interesting."
You play engineering and revenue as a dichotomy when in fact it is a duality.
Nay, good sir, I specifically stated that open peering benefits everyone. I mean benefit in that everyone gets to talk to everyone else, and those customers for whom bandwidth or latency are major factors will pay premium networks for that extra umph. ...just like now. I don't see CNN homing their web site on GeoCities, even though it would cost less...
In the "real world" that you like to refer to, it is more of a symbiotic relationship. Though it would be nice to be aloof in a lab and be lavishly compensated, I would venture to say that such a situation is not the norm.
Again, it is exactly that symbiotic relationship which I refer to. It is the symbiotic relationship which will fail when one component becomes so self-important that it defends itself from illusional dangers to the exclusion of all else. Profit, for profits sake, does not engender quality.
Let us get one point understood. You deserve nothing more than you pay for. You were not, by merit of your diction, infrastructure, prowess, or mandate from heaven, conferred a right to the resources of another individual or conglomeration of individuals.
It was my impression that connectivity customers paid by the bandwidth they wanted, or used, and that web hosting customers pay by the bandwidth they wanted, or used. If one or the other cannot get to the market they wish to, only one power can keep them from going to another provider who can. Regulation.
If you abhor these paper-pushers so, you are more than able to find the capitalisation to finance a global altruistic network where anyone can peer with anyone and we will all get along.
Too late. What, did you think I hadn't already thought of that?
It will be akin to the communes of the 60s. We can just hallucinate until the reality evaporates.
My oh my, it must have been an *interesting* bowl of cheerios....
BR
Curt-
participants (2)
-
Bradley Reynolds
-
howland@Priss.com