by publius
I'm pretty sure I have never -- before today -- yelled at NPR's Robert Siegel in my car. But his intro to the segment on the FCC ruling left me no choice (audio):
The Internet has gone largely unregulated since it came into being more than 30 years ago[.] But today the Chairman of the [FCC] announced his intention to change all that. . . .
No no no no. The FCC is changing nothing. It is preventing change.
As I've argued before, the Internet succeeded because of regulation. It's true, of course, that higher-layer content services (email, videos, browsing, etc.) have always been unregulated. But these services have traditionally relied upon an underlying regulated open access regime that allowed them to flourish.
To be grossly general, just imagine roads. The world today would be much different if the Ford Motor Company owned the interstate highways and could block Hondas from using it. In short, you could imagine a "closed" interstate highway system. But that's not how the interstates work. They're "open." Anyone can use them. Any "device" with wheels will work on them.
The Internet could have very easily evolved into a "closed" network, but it didn't. Instead, the Internet is open because we adopted policies in the 1960s and 70s that required it to be open (namely, we prevented AT&T from strangling it). The FCC today is merely protecting what has always been.
The point of confusion is that the Bush Administration -- an ongoing gift to the world -- adopted new legal labels for "the roads." In short, the Bush FCC started deregulating the physical transmission facilities by legally equating them with content services like Google.
In sum -- the Bush FCC expanded the definition of "the Internet" to include things that have always been regulated. The FCC today merely reaffirmed these traditional protections. And so it's only an "unprecedented" regulation if you ignore the entire history of the Internet. Come on NPR -- you're NPR.
Interstates are not really open: try using a moped, or bicycle, or tricycle. They have wheels, but you are not allowed to use them. Try hitchhiking or even walking...same deal.
Interstates are fairly well restricted to a class of transportation: vehicles with high rates of speed.
Posted by: jrudkis | September 22, 2009 at 01:22 AM
Ok - fair enough, but hopefully the larger point stands
Posted by: publius | September 22, 2009 at 01:43 AM
I don't know: going forward with your analogy, we have some vehicles that cause greater congestion because they are large or slow (video packets, for example, the internet equivalent of a Wide load double wide mover). Should they be banned from the highway, relegated to the backroads, or should the high speed traffic subsidize larger roads to accommodate the guy with the house moving business who wants to move during rush hour? How many roads say "no trucks" or briges have weight limits?
Any system will always have capacity limits, especially at peaks and at the fringes of service.
I think it is okay for public highways to restrict house moving to periods of low traffic. It might be okay for the internet overlords to restrict high volume services to off peak times too...or at least make them pay for the inconvenience caused to others.
Posted by: jrudkis | September 22, 2009 at 02:22 AM
[...]
In recent years, self-styled "network neutrality" activists have pushed for legislation to prevent network owners from undermining the end-to end principle. Although the concern is understandable, such legislation would be premature. Physical ownership of internet infrastructure does not translate into a practical ability to control its use. Regulations are unnecessary because even in the absence of robust broadband competition, network owners are likely to find deviations from the end-to-end principle unprofitable.
New regulations inevitably come with unintended consequences. Indeed, today's network neutrality debate is strikingly similar to the debate that produced the first modern regulatory agency, the Interstate Commerce Commission. Unfortunately, rather than protecting consumers from the railroads, the ICC protected the railroads from competition by erecting new barriers to entry in the surface transportation marketplace. Other 20th-century regulatory agencies also limited competition in the industries they regulated. Like these older regulatory regimes, network neutrality regulations are likely not to achieve their intended aims. Given the need for more competition in the broadband marketplace, policymakers should be especially wary of enacting regulations that could become a barrier to entry for new broadband firms.
[...]
The Durable Internet: Preserving Network Neutrality without Regulation
Posted by: CharlesWT | September 22, 2009 at 03:03 AM
Charles WT: Regulations are unnecessary because even in the absence of robust broadband competition, network owners are likely to find deviations from the end-to-end principle unprofitable.
Bwah-ha-ha-ha.
Thanks, it does you good to laugh.
Now, returning to serious discussion...
Posted by: Jesurgislac | September 22, 2009 at 04:06 AM
jrudkis: Video packets are exactly the same size as any other kind of packets. There's just a larger total number of them.
Posted by: Hob | September 22, 2009 at 04:21 AM
I'm having a difficult time understanding why you've never before yelled at Robert Siegel, in your car or anywhere else. You must not have been listening to NPR during the runup to the blitzkrieg against Iraq, when Siegel was allowing any yahoo who wanted to, to say the most outrageous things about the "need" for the war and about opponents of the war.
Siegel is one of the poorest, most incompetent talking heads on NPR. [Second place is held by a female who asks run-on questions. She was soundly put down one time by Zbiegniew Brzezinski after she interrupted his response to one of her questions (a response which was obviously not what she had hoped to hear) to ask him another question.]
Please, shout at Siegel more often.
As for this comment by jrudkis--"Interstates are not really open: try using a moped, or bicycle, or tricycle. They have wheels, but you are not allowed to use them. Try hitchhiking or even walking...same deal."--I see mopeds, Harleys, bicyclists and walkers on IH-10 frequently. I suspect a tricycle would not be banished if it were one of the adult size. I've not seen hitchhikers on IH-10, but that could be because I travel it mostly through West Texas, where the traffic is sparse and traveling by hitchhiking could require years to get anywhere.
Posted by: Doran | September 22, 2009 at 05:36 AM
New regulations inevitably come with unintended consequences. . . . Like these older regulatory regimes, network neutrality regulations are likely not to achieve their intended aims
I have never understood this viewpoint--as if there is some law of physics that prevents human ingenuity from crafting regulations that work . . .
Posted by: rea | September 22, 2009 at 07:16 AM
I'm pretty sure I have never -- before today -- yelled at NPR's Robert Siegel in my car.
You must have a fairly high tolerance for the phrases "enhanced interrogation techniques" and "which administration critics call 'torture.'"
Come on NPR -- you're NPR.
"Nice Polite Republicans"
Posted by: Ben Alpers | September 22, 2009 at 07:55 AM
Physical ownership of internet infrastructure does not translate into a practical ability to control its use.
This is not correct. Network providers can easily alter their routing tables to block or preferentially degrade service to companies that compete with them. For example, Verizon or Comcast can easily block or degrade their customers' access to VOIP companies that compete with telephony services offered by Verizon or Comcast.
CharlesWT, do you actually know anything about network engineering? I mean, have you actually read anything about net neutrality before today?
Posted by: Turbulence | September 22, 2009 at 08:40 AM
I'm having a difficult time understanding why you've never before yelled at Robert Siegel, in your car or anywhere else.
No kidding! I have even developed a Robert Siegel impersonation over the years: 'I'm Robert Siegel, and I'm schmarter than YOO'. NPR is just the more smug version of the CW you get everywhere else. Sorry, NPR's crap.
It's funny (not 'ha ha') that in the cartoon partisan GOP worldview, NPR and the NYT are supposedly the liberal analogue to Fox News - propaganda arms of liberalism. Would that that were so!
Come on NPR -- you're NPR.
"Nice Polite Republicans"
Exactly, Ben A.!
It's a good thing the open internet exists...
Posted by: jonnybutter | September 22, 2009 at 08:54 AM
"Coming up on 'All Things Considered', Danner Albright Vanderbilt McWaspy files the first installment of her 10 part report on the efforts of a group of college students in Anchorage to help inner city Inuit youth rediscover their cultural touchstones via an innovative theater project called, 'Words For Snow.'"
Posted by: jonnybutter | September 22, 2009 at 09:11 AM
publius:
I know you're busy and all, but I'd appreciate you expanding on the Bush FCC started deregulating the physical transmission facilities by legally equating them with content services like Google. In my naive worldview, regulation and deregulation are, literally, application of the law.
So, it'd be interesting for me to see some mention of which laws changed, and in what way they changed. And why the FCC is at fault for applying those laws.
Or, alternatively, correction of this (highly possibly) erroneous idea of how things work. If you have time, I mean.
Posted by: Slartibartfast | September 22, 2009 at 09:28 AM
It's funny (not 'ha ha') that in the cartoon partisan GOP worldview, NPR and the NYT are supposedly the liberal analogue to Fox News - propaganda arms of liberalism. Would that that were so!
Quelle coïncidence!
Posted by: Uncle Kvetch | September 22, 2009 at 09:55 AM
Such NPR hostility -- I clearly hit a nerve. :)
Slart -- I can do that. If you want a background, check out the 2005 Supreme Court opinion called Brand X -- it lays a lot of it out.
The upshot is that the FCC traditionally divided the world into "telecommunications services" and "information services" (used to be basic/enhanced). The former referred to the transmission services that telephone carriers provided. These were classified as common carrier services, and thus open.
Internet services (higher layers services) were called "information services" and thus largely unregulated. What the FCC technically did is reclassified the physical transmission services as "information services." So now everything Internet-related (transmission, content) is an "information service."
This is what the Comcast case is about (which I'm assisting with in an adverse way to Comcast). The question is -- what residual or "ancillary" jurisdiction does the FCC retain over information services. If there's none, that's very bad news b/c Comcast can do things like block BitTorrent, etc. And that's exactly what they did -- and it's what makes what the FCC did so necessary
Posted by: publius | September 22, 2009 at 10:06 AM
" If there's none, that's very bad news b/c Comcast can do things like block BitTorrent, etc."
I follow all of the discussion except why this is bad. It is a really bad example of a problem because it is information, that should be able to be blocked for commercial or legal reasons.
Is there another example?
Posted by: Marty | September 22, 2009 at 10:13 AM
Marty, why should BitTorrent be blocked? Are you under the impression that it is used only for illegal purposes? It is used to distribute large files that are perfectly legal to distribute. Should phone service be banned because people use phones in committing crimes?
Posted by: KCinDC | September 22, 2009 at 10:28 AM
Put me in the camp that doesn't get how you haven't yelled at NPR on the radio before. I've finally stopped listening after 20-some years.
Posted by: Irene Miles | September 22, 2009 at 10:43 AM
It's bad because Comcast is arbitrarily choosing a sort of network traffic and blocking it indiscriminately.
BitTorrent may seem like a bad example because a large portion of its usage is to violate copyright laws, but 1) by no means all of it is (many companies use P2P software downloads to reduce their business' bandwidth requirements when providing large chunks of data), and 2) it is a corporation (not a law enforcement entity or suchlike) taking it upon itself to decide what (legal) applications can and cannot use its transmission services.
Posted by: Nombrilisme Vide | September 22, 2009 at 10:46 AM
To follow up along the lines of KCinDC, FTP has long been used to pirate copyrighted material. Should Comcast be allowed to block FTP traffic?
Posted by: Nombrilisme Vide | September 22, 2009 at 10:49 AM
"BitTorrent may seem like a bad example because a large portion of its usage is to violate copyright laws, but 1) by no means all of it is (many companies use P2P software downloads to reduce their business' bandwidth requirements when providing large chunks of data), and 2) it is a corporation (not a law enforcement entity or suchlike) taking it upon itself to decide what (legal) applications can and cannot use its transmission services."
I am not sure I understand why Comcast (as just a corporation, not a government entity)shouldn't be able to protect its bandwidth capacity to provide the services it has contracted to provide(my internet access for example) and its business of providing, uh, movies. Thats why its a bad example.
If you move up a tier I can understand real pros and cons where I have no choices for my internet access, much more comparable to phone transmission services.
Posted by: Marty | September 22, 2009 at 11:18 AM
I follow all of the discussion except why this is bad. It is a really bad example of a problem because it is information, that should be able to be blocked for commercial or legal reasons.
Then feel free to replace "using bittorrent to watch movies" with "using iTunes to rent movies on demand".
I am not sure I understand why Comcast (as just a corporation, not a government entity)shouldn't be able to protect its bandwidth capacity to provide the services it has contracted to provide(my internet access for example) and its business of providing, uh, movies. Thats why its a bad example.
I get my internet service from Verizon. Verizon also offers phone service. I don't like Verizon. I don't want to buy their phone service. So I buy VOIP service from another company that provides phone service. But that service depends on Verizon faithfully moving packets from me. In a world where companies didn't respect network neutrality at all, Verizon would happily block or degrade service to VOIP companies that compete with it.
This would be bad for me: it would cause me to spend a lot more money than I wish to spend buying an inferior product. It would be bad for the VOIP companies because they were providing a valuable service but now there is literally nothing they can do to compete with Verizon: Verizon's ability to crush them comes not because consumers prefer Verizon (if they did, there would be no reason to block or degrade service) but because Verizon has a monopoly on internet service provision in some areas. It would also be bad for Verizon since Verizon would no longer feel pressure to satisfy customers or improve their service.
Can you explain to me why this world is preferable? It harms everyone except market incumbents, and even incumbents are harmed in that it causes them to become sclerotic and insensitive to their customers' needs. Free market competition is the basis of our economy, yes? Or is that only true when competition doesn't threaten the interests of incumbents?
Posted by: Turbulence | September 22, 2009 at 11:35 AM
If you move up a tier I can understand real pros and cons where I have no choices for my internet access, much more comparable to phone transmission services.
Note that many many people do not have a choice regarding internet service providers. Comcast is the only game in town.
Beyond that, having a choice of multiple providers does not mean much if all providers act according to the same incentives. If I can choose between Verizon and Comcast but both those companies sell phone service in addition to data service, then they're both going to block or degrade service to VOIP competitors in a world without net neutrality.
Posted by: Turbulence | September 22, 2009 at 11:39 AM
"Beyond that, having a choice of multiple providers does not mean much if all providers act according to the same incentives. If I can choose between Verizon and Comcast but both those companies sell phone service in addition to data service, then they're both going to block or degrade service to VOIP competitors in a world without net neutrality."
VOIP is a unique business that has, through technology, managed to disintermediate a market by using the costs of its competitors to provide a really cheap service. I think in a regulated world they should pay for the bandwidth they use on their competitors networks, phone companies actually have to do this. It is a classic example of an appropriate place to have to pay or get blocked.
Posted by: Marty | September 22, 2009 at 11:45 AM
Thanks for the response, publius. I don't understand it yet, but at least I have some reading to do now. Again, to me, the sticking point is to what extent federal regulatory agencies can decide what to regulate independently of, well, regulations. As in laws.
Maybe I'm inappropriately conflating regulation and law, but to me one ought to at least be driven by the other.
Comcast, as far as I can tell, still exists only because it has monopoly power in some markets for various reasons like abysmal customer service and frequent outages. That may have changed by now, but I don't know any Comcast customers that wouldn't pick something better if only they could.
I'd guess Comcast is liable for copyright violations via its networks only if the law says they are. And if they were responsible for others using their network to break the law in that way, that'd open up all kinds of other liability that would, I think, make it nearly impossible for them to operate.
Posted by: Slartibartfast | September 22, 2009 at 11:46 AM
I am not sure I understand why Comcast (as just a corporation, not a government entity)shouldn't be able to protect its bandwidth capacity to provide the services it has contracted to provide
Because 'protecting bandwidth capacity' is often a red herring. Comcast can *say* that that's what they're doing when it has other motives (e.g., they want you to buy content from them; they simply want to be paid fees for large network activity). Although, as i understand it, coax cable actually has a ton of throughput, relatively speaking, if bandwidth *is* an issue, it ought to be dealt with globally. If you think of the internet as a public utility (which I do), you don't want individual companies deciding arbitrarily what apps or content to block. Deal with bandwidth issues as such.
Others on this thread (and publius of course) have far more expertise on this than I do, but my understanding it that bandwidth scarcity is a bit of an artificial problem anyway. It's in the interests of telcos and cable companies for their commodity to be scarce.
Posted by: jonnybutter | September 22, 2009 at 11:48 AM
VOIP is a unique business that has, through technology, managed to disintermediate a market by using the costs of its competitors to provide a really cheap service. I think in a regulated world they should pay for the bandwidth they use on their competitors networks, phone companies actually have to do this. It is a classic example of an appropriate place to have to pay or get blocked.
This is simply wrong. Every VOIP provider in the world pays for network access to the internet. They pay for every byte they send or receive. Every VOIP provider in the world pays for access to telephony exchanges.
Look, moving data across the internet is not free. It costs money. Those costs are well known. And VOIP providers pay for the costs they incur. If they didn't, their network providers would cut them off.
Perhaps I'm misunderstanding you though: what specific bandwidth do VOIP companies use and not pay for?
Is iTunes selling and renting movies also "a unique business that has managed to disintermediate a market"? Because everything that I've said above about VOIP services applies equally well to iTunes and other video on demand services.
Posted by: Turbulence | September 22, 2009 at 11:51 AM
VOIP is a unique business that has, through technology, managed to disintermediate a market by using the costs of its competitors to provide a really cheap service. I think in a regulated world they should pay for the bandwidth they use on their competitors networks, phone companies actually have to do this. It is a classic example of an appropriate place to have to pay or get blocked.
Marty,
If I want VOIP service, I have to pay someone for sufficient bandwidth to support it. Verizon shouldn't care what kind of packets I'm bringing in so long as I'm paying for the level of service I'm receiving. If they weren't also a phone company, they wouldn't give a rat's behind about VOIP, just like they don't give a rat's behind about all the other data that doesn't compete with their other services.
I'm the customer with the internet connection and I'm ALREADY PAYING FOR IT. People who use more bandwidth pay for it. What they choose to do with that bandwidth should be their business within certain legal constraints. How is VOIP different from any other beneficial use of the internet?
Posted by: hairshirthedonist | September 22, 2009 at 11:57 AM
data that doesn't
or: data that don't
Posted by: hairshirthedonist | September 22, 2009 at 11:59 AM
My understanding is that some (many) ISPs block port 80 so that customers can't run websites from their home computers. I'm not seeing how throttling or blocking bittorrent would be different as long as it was applied equally to all source providers?
Posted by: 243 | September 22, 2009 at 12:07 PM
I am not sure I understand why Comcast (as just a corporation, not a government entity)shouldn't be able to protect its bandwidth capacity to provide the services it has contracted to provide(my internet access for example)...
They sell the same service to both you and me: typically 7 Mbps downstream, 2 Mbps upstream, a limit on monthly data transfers of 250 GB. If they want to impose further limits on usage -- say, daily transfer limits -- then let them make it part of the generic terms of service. Or if they want to impose traffic shaping in a generic sense, as in the number of bits/minute in addition to the number per second during peak hours, make it part of the generic terms of service.
I was doing research work and systems analysis for using IP networks for real-time communications for a Baby Bell back in 1993, and one of the truly valuable aspects of IP as a foundation for a network is that it is content-neutral. Since we don't know what the next big thing will be, users should be extremely reluctant to let their ISP do anything that takes away that aspect of the service.
Posted by: Michael Cain | September 22, 2009 at 12:11 PM
New regulations inevitably come with unintended consequences
Having no regulations, or removing existing regulations, inevitably likewise.
Posted by: russell | September 22, 2009 at 12:50 PM
I think it is okay for public highways to restrict house moving to periods of low traffic. It might be okay for the internet overlords to restrict high volume services to off peak times too...or at least make them pay for the inconvenience caused to others.
And
My understanding is that some (many) ISPs block port 80 so that customers can't run websites from their home computers. I'm not seeing how throttling or blocking bittorrent would be different as long as it was applied equally to all source providers?
This would be all well and good, if we could trust the telcos. But once we've given them the power, it'll be easy for them to eg discriminate against VOIP traffic (even when their networks aren't congested) and claim that it's part of a general, content-neutral throttling based on protocol.
Marty, the truck analogy isn't apt- I think a better one would be a highway full of trucks, but the highway provider also owns a fruit distribution business, so they won't allow competitor fruits trucks on the highway, or delay them to the point that the fruit rots (ie VOIP is useless if the packets are dropped or delayed often enough).
Much better IMO what Michael C suggests- if there are congestion issues, providers should change overall bandwidth, eg offering cheaper bandwidth at night, which torrent users will then preferentially use. If there were solely concerned about bandwidth, this would be a cheaper solution for them anyway- inspecting every truck on the highway looking for fruit is expensive.
Posted by: Carleton Wu | September 22, 2009 at 12:50 PM
BTW, what Turbulence said about hypothetical VOIP companies being throttled to death is exactly what happened to independent DSL companies. The phone companies were supposed to offer open access to their networks (built while a sanctioned monopoly and/or with copious public subsidies) to anyone who wanted to run a DSL network. The DSL companies would rent the space on the wires, and provide internet connections to customers.
They all died off several years ago, because the local phone monopolies did EXACTLY what Turbulence is talking about, they degraded the independent DSL service, they were slow at fixing outages, yet somehow the phone company's own DSL services always worked fine. And now, because of this, the independent DSL companies are dead, and you have a choice of either the local phone company or local cable company for broadband, generally. There's wireless possibilities, but those depend on either cellular or satellite reception, and are often slower, laggier, drop packets, etc. Lower Quality of Service.
This is not a hypothetical question here, about things that might happen in the future. This is about things that have happened, and are happening.
Posted by: Nate | September 22, 2009 at 12:55 PM
What strikes me as particularly insane is that it's entirely likely the telcos internally send your voice calls on your phone service using Voice-Over-IP...
(I worked on a project in the UK to convert one of the major mobtelcos here to using VOIP a few years ago because it was much cheaper to deliver over IP than the way they were doing it before - I'd be surprised if all the major networks hadn't switched over, at least for communications between exchanges. Obviously the last mile will always be PSTN.)
Posted by: sanbikinoraion | September 22, 2009 at 12:59 PM
What strikes me as particularly insane is that it's entirely likely the telcos internally send your voice calls on your phone service using Voice-Over-IP...
From Wikipedia on IP telephony:
PSTN and mobile network providers
It is becoming increasingly common for telecommunications providers to use VoIP telephony over dedicated and public IP networks to connect switching stations and to interconnect with other telephony network providers; this is often referred to as "IP backhaul".[14][15]
Posted by: hairshirthedonist | September 22, 2009 at 01:12 PM
"Perhaps I'm misunderstanding you though: what specific bandwidth do VOIP companies use and not pay for?"
The final delivery leg. They pay to get on the backbone, but not off. They don't own or pay for any of the local connect bandwidth. Verizon is a harder example because it plays on both tiers. Comcast doesn't. It provides local connect, pays for access and the VOIP vendiors use their investment to provide local connect.
Posted by: Marty | September 22, 2009 at 01:30 PM
Marty: *I* pay for the last mile of bandwith for VoIP, from the ISP to me. It's not something Comcast/Verizon/Bellsouth/etc isn't getting paid for.
And 90% of the last mile networks are run by the same companies who also run backbones, that the VoIP companies pay.
I'm not sure where you think somebody's not getting paid. It looks like you want the ISPs to get paid twice?
Posted by: Nate | September 22, 2009 at 01:36 PM
The final delivery leg.
Isn't that what Comcast's and Verizon's customers pay for? Again, why is VOIP any different from anything else a customer might use their internet connection for?
Posted by: hairshirthedonist | September 22, 2009 at 01:39 PM
The final delivery leg. They pay to get on the backbone, but not off. They don't own or pay for any of the local connect bandwidth. Verizon is a harder example because it plays on both tiers. Comcast doesn't. It provides local connect, pays for access and the VOIP vendiors use their investment to provide local connect.
This is very very confused. Individual customers pay for network providers for access over the providers' networks. If Verizon wants to charge me much less for internet access than it actually costs them, well, there's an easy solution to that: charge me more. Moreover, when a VOIP provider pays for bandwidth from another network provider, part of what they're paying for is transmission to the end user, not just over the internet. The VOIP provider's ISP has cut a deal with Verizon to move packets; the terms of that deal require that Verizon actually move packets to the end user in exchange for letting Verizon hook up to a backbone.
Marty, with respect, I don't think you actually understand how data networks operate well enough to support your arguments.
Posted by: Turbulence | September 22, 2009 at 01:53 PM
Again, to me, the sticking point is to what extent federal regulatory agencies can decide what to regulate independently of, well, regulations. As in laws.
Regulations are not the same as laws. In some domains, Congress gives Executive branch regulatory agencies broad latitude to enact very specific regulations that implement much vaguer goal-oriented legislation that Congress has passed. The regulatory powers of the FCC stem directly from laws that Congress has passed but the specific regulations enacted by the FCC
I'd guess Comcast is liable for copyright violations via its networks only if the law says they are. And if they were responsible for others using their network to break the law in that way, that'd open up all kinds of other liability that would, I think, make it nearly impossible for them to operate.
The DMCA contains safe harbor provisions that protect Comcast provided that it takes certain minimal efforts to respond to complaints filed by copyright owners. In practice, Comcast faces no real liability when its customers use bittorrent to exchange copyrighted material.
Posted by: Turbulence | September 22, 2009 at 01:55 PM
"Marty, with respect, I don't think you actually understand how data networks operate well enough to support your arguments."
Actually Turb, I think I do. VOIP has always been an odd child. It competes directly with the vendors on their own backbone. While it pays for the backbone bandwidth it uses, it doesn't pay Comcast to deliver those packets. If Comcast chooses to not deliver their competitors packets on their network it doesn't seem to be an unreasonable business decision. There really isn't an analogy to it anywhere(iTunes comparisons notwithstanding).
Keep in mind that I am talking about Comcast in particular, again, Verizon ends up in a more tenuous position as it operates on multiple tiers of the network.
Posted by: Marty | September 22, 2009 at 02:10 PM
If Comcast chooses to not deliver their competitors packets on their network it doesn't seem to be an unreasonable business decision.
But, Marty, that's kind of the point. It's picking and choosing based on content to stifle competition. And that's bad for the consumer. Do you also think Standard Oil should have been allowed to do whatever it wanted as well?
Maybe it's not an unreasonable business decision (i.e. rational). But it's also not unreasonable to prevent such reasonable business decisions through regulation if such regulation is better for the consumer and the marketplace in general (i.e. good).
Posted by: hairshirthedonist | September 22, 2009 at 02:44 PM
While it pays for the backbone bandwidth it uses, it doesn't pay Comcast to deliver those packets.
Again, the end-user is paying for the delivery of those packets. No provider of any service over the internet pays for the delivery to the end-user.
VOIP is not odd in this regard, it is exactly like everything else.
There really isn't an analogy to it anywhere(iTunes comparisons notwithstanding).
Only until Comcast parters with Nexflix and blocks other video-on-demand traffic. Again,m VOIP is not special.
If Comcast chooses to not deliver their competitors packets on their network it doesn't seem to be an unreasonable business decision.
Sure, it makes sense for Comcast. It just doesn't make sense for anyone else, including the public. Unregulated monopolies, pricing collusion, and regulatory capture are also good for specific businesses, but we discourage them because they produce bad outcomes.
I don't even know why you'd present that as an argument- "it's good for Comcast, so we should all just live with it."
Posted by: Carleton Wu | September 22, 2009 at 02:44 PM
Marty, I don't follow that logic. Verizion and Comcast both offer telephone service, so why would they be in a more tenuous position?
And since Comcast offers cable, should they be able to throttle Hulu, YouTube, iTunes and Netflix streaming movies? Or YouTube videos of cats eating broccoli?
Comcast offers email, so should they be able to 'lose' GMail's packets? What if Comcast decides to get into the plumbing business, can they make all other plumbers websites go 404 for their internet customers? If Comcast makes a GPS division, can they alter all Google Maps results to make the wrong turn in Albequrque?
Hell, should Comcast be able to block customers from viewing Verizion's internet setup page? Are these all reasonable business decisions?
VoIP companies pay for the bandwith they use. Comcast's internet customers pay for the bandwith THEY use. Comcast's local and backbone usage are paid for by the people using them. Just because some of that usage belongs to applications or companies that offer things that Comcast ALSO offers gives Comcast no right to degrade that usage, no matter how reasonable or effective a business practice that may be for them in their local near-monopoly situation.
Posted by: Nate | September 22, 2009 at 02:49 PM
One more thing to note. Compliance with this regulation requires no effort at all. It is NON-compliance that would require effort.
By default, TCP/IP networking is completely indifferent to what the actual traffic looks like. You have to work hard to treat different kinds of traffic in different ways.
The only compliance cost of this is that in order to deliver satisfactory performance to customers, ISPs might have to install additional network capacity. All the problems with excess traffic load that ISPs claim to want to solve by violating net neutrality can be solved just as well by adding bandwidth - which they really ought to be doing anyway since the US has very slow broadband compared to Europe, Japan, or Korea.
(As I posted on the first of the two previous FCC posts, the situation is a bit more complex for wireless, although the eventual goal should be the same.)
Posted by: Jacob Davies | September 22, 2009 at 03:06 PM
Are you thinking of Terry Gross? Though she mostly interviews entertainment personalities with questions like, "Let's listen to a sample of your work. I have an insight about this which I'm sure will demonstrate how smart I am. Would you care to comment on this insight and agree that this shows that I am an erudite consumer of your oeuvre?"
Posted by: JustMe | September 22, 2009 at 03:09 PM
@JustMe: Oh come on. For all of NPR's faults, Terry Gross isn't one of them. Just because she actually familiarizes herself with a significant portion of the work of the person she is interviewing isn't a reason to hate her - it's called being a good interviewer. She's regularly one of the more interesting personalities on public radio.
NPR's big problem is in its news coverage. It's been completely cowed by the liberal bias accusation and that fear permeates all their new coverage.
Posted by: kyllaros | September 22, 2009 at 03:21 PM
kyllaros, they're symptoms of the same thing: an interest in self-promotion and a desire to be liked, rather than an interest in doing good journalism.
I remember reading a comment somewhere in the blogosphere that NPR always seems focused in doing good *radio*, not doing good *journalism*. This misplaced focus causes them to falter.
Posted by: JustMe | September 22, 2009 at 03:25 PM
This is silly quibbling that entirely misses the point. The purpose of the highways is high-volume transit of motor vehicles. To the extent that states and localities have laws governing what can and cannot travel on them, these laws typically concern proper licensing, the requirement of powered motors and their ability to maintain a safe minimum speed.
In other words, nothing even remotely analogous to net neutrality.
In order to properly appreciate the issues here, you need to understand how the internet functions, and it really helps to have a grasp of the OSI model. For those who already know what I'm talking about, please forgive the inevitable simplification that will occur.
The data which travels over the internet is stored in IP packets. Every time you send information of any kind, it is broken up and encoded into these packets. There is nothing structurally that makes one different from another (aside from size, which has to do mainly with the network hardware it passes through and is unrelated to the content). All of this happens at the Transport and Network layers.
BitTorrent, by contrast, is an application (or rather, a protocol used by many applications). It is considered part of the Application layer, along with FTP, DNS, HTTP, and many others.
Network congestion occurs when the total size of all packets traveling through a point exceeds available bandwidth. This is related to the Physical, Data Link and Network layers, and to some extent on how the packets are handled by the Transport layer.
There is no such thing as a BitTorrent packet, or an FTP packet, or anything of the like. Packets may or may not contain data that is part of one of these protocols, and the protocol used has no relationship to the size of the packets--it is simply a language for communicating between two or more points, and has no functional relationship to network congestion. The problem with BitTorrent-related network congestion is one of usage, not function.
What Comcast was doing was scanning for the use of specific protocols, and sending back a forged packet that fooled the user's application into thinking the connection was reset--effectively interrupting the conversation between the user and the remote machine.
They were using a Layer 7 solution to deal with a Layer 8 problem.
Posted by: Catsy | September 22, 2009 at 03:31 PM
VOIP has always been an odd child. It competes directly with the vendors on their own backbone. While it pays for the backbone bandwidth it uses, it doesn't pay Comcast to deliver those packets. If Comcast chooses to not deliver their competitors packets on their network it doesn't seem to be an unreasonable business decision.
I think you're missing something critical here. Comcast can only function because it has cut peering and transit deals with tier 1 network providers. Without those deals, Comcast's customers wouldn't be able to access the internet. Those deals require the tier 1 providers to move Comcast's packets (i.e., packets originated by Comcast's customers) but also require that Comcast move packets from the tier 1 providers (i.e., their customers) to Comcast's customers. Every VOIP company is a customer (perhaps indirectly) of a tier 1 provider. To the extent that Comcast refuses to pass on a VOIP company's packets to its customers, it is violating the deal it has with the VOIP company's network provider.
It is insane to talk about VOIP companies paying Comcast extra. The cost of transporting the VOIP companies' data is incorporated into the peering and transit agreements that Comcast has struck with its upstream providers and that each VOIP company has struck with its network providers. The whole point of the internet is that individual services (like VOIP companies or blogs or Google) don't have to negotiate separate financial arrangements with every single network provider on Earth.
There really isn't an analogy to it anywhere(iTunes comparisons notwithstanding).
It seems like a straightforward case of service bundling or tying. I imagine an attorney with experience in antitrust law should have many useful analogies.
You still haven't dealt with the iTunes case. If you feel that VOIP is "an odd child" (whatever the frak that means), then please explain whether iTunes video sale/rental is also "an odd child". Should Comcast be able to block or degrade service to iTunes since it competes with Comcast's video services?
Keep in mind that I am talking about Comcast in particular, again, Verizon ends up in a more tenuous position as it operates on multiple tiers of the network.
Feel free to replace Verizon with Comcast in my comments. I talk about Verizon because I have their service now, but all of my arguments applied just as well back when I had Comcast service.
Posted by: Turbulence | September 22, 2009 at 03:53 PM
FCC regulations had very little to do with the evolution of the internet. They did keep common carriers from tariffing their data lines in such a manner that ISPs would have been strangled, so their role wasn't zero, but it was rather small.
Internet standards are defined (but not regulated or even enforced) by the Internet Engineering Task Force. An IETF meeting is about 75% abject geekery and 25% the most vicious politics that you've ever seen. (How standards get adopted can afford huge advantages to one competitor over another, just as regulations can.) But nobody has the authority to impose conformance to the standards--except the ISPs themselves. They tend to filter out non-conforming traffic because it can make their networks collapse or seriously misbehave if they don't.
This is where things start to get tricky. The line between a legitimate application and one that is misbehaving is often in the eye of the beholder, especially since some legitimate applications use non-standard protocols, or leverage the standards in odd ways (BitTorrent is a good example of this). Network managers simply cannot treat all traffic equally. There are too many ways to crash big hunks of the network if you don't pay attention to what's being injected into it.
Which brings us to network neutrality. The term itself isn't adequately defined, but usually has one of two definitions. It can either mean:
1) To accept, route, and deliver content in such a way that no traffic policy is applied based on the origin or destination of the content, or
2) To treat all content equally.
Definition #1 is perfectly fine. Definition #2 will cause the internet either to become wildly more expensive or to become useless for all but the most trivial tasks.
The problem has to do with how real-time traffic (VoIP, live video, or, to some extent, streaming video) gets routed. Ordinary internet packets get routed best-effort, which means that packets that get sent on a regular, periodic basis (like VoIP or video) can sometimes arrive in big bursts, separated by long gaps of time. This means that the voice will either break up or it will have to be delayed for so long that conversation is very difficult.
ISPs deal with this problem today by over-designing their network routing fabric, supplying lots of spare capacity. However, as real time HD video becomes more pervasive, that capacity is being eaten up pretty quickly.
At some point, an ISP either has to add more routers and network connections or run the risk of providing lousy real time service. It's expensive to upgrade your network. Those costs get passed along to consumers in higher subscription rates.
However, if you can discriminate between real time traffic and ordinary web browsing, email, IM, and other non-real time stuff, you can manage your router capacity much more closely, and you can charge the consumers and producers of real time stuff more, keeping the bills of ordinary internet users lower.
But this packet discrimination (it's called differentiated services, or diffserv) would be made illegal by definition #2 above. The end result will either be much higher subscription rates for internet service, or the death of the various real time services.
Note that the FCC's recent fifth principle (http://www.openinternet.gov/read-speech.html#book5) doesn't really distinguish between the two different neutrality definitions above, and the accompanying explanation is more than a little bit slippery. There is a real possibility for the FCC to screw the pooch on this issue. Net neutrality sounds great on paper, but it's not so great out in the real world.
Posted by: TheRadicalModerate | September 22, 2009 at 04:00 PM
Which brings us to network neutrality. The term itself isn't adequately defined, but usually has one of two definitions.
This statement is ridiculous. You know what terms are not "adequately defined"? Due process. Torture. Obscenity. Life, liberty and the pursuit of happiness. There are no interesting policies that are "adequately defined" by their two word title. That's life. The two word title is a just a reference to dig up a larger body of thought.
The problem has to do with how real-time traffic (VoIP, live video, or, to some extent, streaming video) gets routed. Ordinary internet packets get routed best-effort, which means that packets that get sent on a regular, periodic basis (like VoIP or video) can sometimes arrive in big bursts, separated by long gaps of time. This means that the voice will either break up or it will have to be delayed for so long that conversation is very difficult.
Have you used a VOIP service recently? In practice, VOIP generally works fine without providers having to kill themselves to provide expensive QoS. The issue is preferential degradation or blocking, not inherent limitations of the technology in general. If providers gave customers a crummy VOIP experience because they didn't want to upgrade their networks, that would be fine with net neutrality advocates as long as the providers' VOIP services suffered the same crummy performance.
ISPs deal with this problem today by over-designing their network routing fabric, supplying lots of spare capacity. However, as real time HD video becomes more pervasive, that capacity is being eaten up pretty quickly.
Cite? Note that this has precisely zero to do with VOIP, which is very low bandwidth.
At some point, an ISP either has to add more routers and network connections or run the risk of providing lousy real time service. It's expensive to upgrade your network. Those costs get passed along to consumers in higher subscription rates.
From what I understand, the primary issue in net neutrality is last mile bandwidth, not bandwidth in the provider's core network or the internet core at large.
However, if you can discriminate between real time traffic and ordinary web browsing, email, IM, and other non-real time stuff, you can manage your router capacity much more closely, and you can charge the consumers and producers of real time stuff more, keeping the bills of ordinary internet users lower.
This is wrong. VOIP traffic is so low bandwidth that you could never hope to generate significant revenue streams if you priced it fairly. There isn't much else that actually requires real time bounds. Video conferencing requires low latency and low jitter for audio, but I'm skeptical that such properties are needed for the video portion.
In any event, the bill that most internet users pay has absolutely nothing to do with what bandwidth costs the provider. If people payed per byte based on fractional usage, the average customer would see a large drop in price. But we're not going to do that.
But this packet discrimination (it's called differentiated services, or diffserv) would be made illegal by definition #2 above. The end result will either be much higher subscription rates for internet service, or the death of the various real time services.
I've never seen any net neutrality standard used by net neutrality advocates that would prevent providers from offering preferential service to traffic on a user defined basis. I.e., there are ZERO net neutrality problems involved in having providers respect the IP TOS fields as long as users use them responsibly (i.e., don't designate all their traffic as delay sensitive) because the user gets to decide what packets are selected for low delay versus bulk transmission. Can you point me to any net neutrality advocates who would object to this network policy?
Posted by: Turbulence | September 22, 2009 at 04:40 PM
@JustMe: Not every NPR show is a news show. Fresh air (Terry Gross's show) is mostly interviews of culture makers (I struggled with an appropriate term here) - that is, novelists, musician, and artists of various types, though she also interviews politically oriented movers and shakers too. She always seems to me to put a lot of effort into knowing the artist or issue very well before conducting her interview.
If you've heard her interview about the estate tax with a republican strategist, you know that she insists on facts being respected in her show. If you don't like her style, that's your business, but I don't think that she symptomatic of NPR's problems in general. I mean, you wouldn't criticize "Car Talk" because it's not hard hitting news radio, right? Some shows can just be good radio.
Apparently you hit a nerve... I'll quit with the off topic talk now.
Posted by: kyllaros | September 22, 2009 at 04:51 PM
Turb--
In practice, VOIP generally works fine without providers having to kill themselves to provide expensive QoS.
VoIP is not as much of a problem (although wideband VoIP likely will be). Real-time video, however, is going to be extremely difficult. (Note the difference between streaming stored video, where you can buffer it as much as you need to overcome jitter, and conversational video, where you can't.)
If providers gave customers a crummy VOIP experience because they didn't want to upgrade their networks, that would be fine with net neutrality advocates as long as the providers' VOIP services suffered the same crummy performance.
How 'bout if you made all service more expensive because the provider had to upgrade his network to remain competitive? Look, I'm not arguing that you can't solve the problem by throwing (low-latency) bandwidth at it. I'm arguing that that's a really stupid way to engineer an internet.
From what I understand, the primary issue in net neutrality is last mile bandwidth, not bandwidth in the provider's core network or the internet core at large.
That's the political issue, but it's not the technical one. End-to-end services need end-to-end QoS, not just last-mile QoS.
This is wrong. VOIP traffic is so low bandwidth that you could never hope to generate significant revenue streams if you priced it fairly. There isn't much else that actually requires real time bounds. Video conferencing requires low latency and low jitter for audio, but I'm skeptical that such properties are needed for the video portion.
Ever hear of lip sync? You have to delay the audio to keep it sync'd with the video. If you delay the audio too much, the end-to-end latency makes conversation difficult-to-impossible.
Also, you're falling into a common fallacy here: The issue is not "bandwidth." The issue is "latency" and "jitter". You can have plenty of aggregate bandwidth and still be in trouble because you didn't move a bunch of 20 ms voice packets ahead of lower-priority traffic.
In any event, the bill that most internet users pay has absolutely nothing to do with what bandwidth costs the provider. If people payed per byte based on fractional usage, the average customer would see a large drop in price. But we're not going to do that.
What customers pay is a function (among other things) of the capital and operational costs of the network, which are directly related to capacity. ISPs charge a flat rate not because bandwidth is free, but because it's easier to average the cost of their entire subscriber base. The subscribers like it that way because they have a predictable bill. But it's still all about total capacity.
...there are ZERO net neutrality problems involved in having providers respect the IP TOS fields as long as users use them responsibly (i.e., don't designate all their traffic as delay sensitive)...
Well, of course everybody sets their TOS (actually DSCP) to a higher-than-necessary priority, if they can get away with it. The only way to prevent them from doing it is to charge them for the privilege--which is kinda the point.
Can you point me to any net neutrality advocates who would object to this network policy?
Here's Wu and Lessig's letter to the FCC, from 2003. In it: "A network that is as neutral as possible is predictable: all applications are treated alike." This is not cherry-picked; they really want application-agnostic (and hence protocol-agnostic) treatment.
You appear to have a completely reasonable attitude towards this issue. In fact, you're so reasonable that you're having trouble understanding that there are lots of unreasonable people out there. The problem is that you can use that "net neutrality" buzzword to institute some truly regressive policies. (Hence, my definition of the two basic meanings, the ones that you found so ridiculous.)
Posted by: TheRadicalModerate | September 22, 2009 at 05:54 PM
Wait, how exactly is having the Internet remain application agnostic "regressive"? It's conservative, if you want to put it that way, which makes the alliances over it even funnier, but in what way is "saying run the Internet the same way it has successfully been run for decades" regressive?
Posted by: Nate | September 22, 2009 at 06:02 PM
Wait, how exactly is having the Internet remain application agnostic "regressive"?
It's the same way how insisting on open standards is "stifling innovation."
Verizon or Comcast are free to sell me some kind of newfangled data service that streams real-time video. Once they claim to be selling me internet service, though, I expect it to conform to the IP standard. Furthermore, Verizon and Comcast have *openly* stated that their goal is to try to extract rents from Yahoo and Google by threatening to degrade packet delivery from their servers that users ask for.
Posted by: JustMe | September 22, 2009 at 06:17 PM
But this packet discrimination (it's called differentiated services, or diffserv) would be made illegal by definition #2 above.
I don't know about others, but I'd be fine with quality-of-service pricing (ie a cheaper plan that's all low-priority bandwidth and a more expensive one with allowance for high-priority traffic). Or off-peak and on-peak pricing for bandwidth.
But I would want to avoid letting providers down to the protocol-level for discrimination, because that gives providers too much leeway to make their choices based on anti-competitive goals rather than actual quality of services.
In it: "A network that is as neutral as possible is predictable: all applications are treated alike." This is not cherry-picked; they really want application-agnostic (and hence protocol-agnostic) treatment.
They're writing at a pretty high level; I don't think you should be interpreting this at a technical level to mean that they rule out QoS or even pricing based on QoS. And, based on the rest of their letter, their concerns are with anticompetitive activity, not QoS.
Posted by: Carleton Wu | September 22, 2009 at 06:21 PM
but in what way is "saying run the Internet the same way it has successfully been run for decades" regressive?
You can't run the internet the same way it's been run for decades, because the application mix in today's internet is different from the non-realtime mix that used to exist. If you'd like to regress back to Web 1.0 and a bit of email, then we'll have no problem at all.
Posted by: TheRadicalModerate | September 22, 2009 at 07:01 PM
But I would want to avoid letting providers down to the protocol-level for discrimination, because that gives providers too much leeway to make their choices based on anti-competitive goals rather than actual quality of services.
I'm guessing that this is not the forum for a tutorial on diffserv code points and RSVP, but there are extremely robust standards in this area. I'd hate to see the FCC mandating adherence to IETF standards--as would the IETF--but this isn't rocket science.
Also, please note that ISPs make filtering decisions all the time based on protocol information, for a variety of reasons, security and congestion control being high on the list. (You just wouldn't believe the stuff that gets done to TCP packet streams under even moderate load.) The genie's pretty much out of the bottle on this one.
[Wu and Lessig are] writing at a pretty high level; I don't think you should be interpreting this at a technical level to mean that they rule out QoS or even pricing based on QoS. And, based on the rest of their letter, their concerns are with anticompetitive activity, not QoS.
Oddly, no. I was a bit surprised when I plowed through this letter. One of their principal arguments is that content providers need to be able to count on a uniform environment in which to deploy their applications, and that they can't do that if the ISPs can deploy differentiated services. They make an analogy to the electric grid to argue that the one-size-fits-all service provision enables the market. (I guess they don't know about three-phase.)
That's kinda thin. If an ISP provides a differentiated service, then it ought to be forced to treat all traffic marked with the appropriate diffserv codepoints the same, so there's no advantage that can accrue to one provider over another (unless the provider is stupid enough not to mark its packets correctly).
Again, I think that regulation to enforce source- and destination-neutrality is an excellent idea, but there are unobvious technical issues associated with application neutrality that simply aren't going to go away. You legislate or regulate in this area at your peril.
Posted by: TheRadicalModerate | September 22, 2009 at 07:27 PM
TheRadicalModerate: Here's Wu and Lessig's letter to the FCC, from 2003. In it: "A network that is as neutral as possible is predictable: all applications are treated alike." This is not cherry-picked; they really want application-agnostic (and hence protocol-agnostic) treatment.
But you're reading things into what Lessig said that aren't there. They, and all other net neutrality advocates I know of, want the network to remain application-agnostic, so providers can't distinguish applications that potentially complete with their products and deliberately provide poorer QoS for those applications.
But this is tangential to what you're discussing, which is differential pricing for better QoS for real-time applications. I don't think the FCC or anyone else would have an issue with this, provided real-time QoS was available for all applications, regardless of the vendor. Its never been economic in the past for anyone to even try to do this (and believe me they have), because the costs of billing for local-loop internet traffic basically exceed anything you could possibly charge for it, but I don't see a problem in principle.
Posted by: Simon K | September 22, 2009 at 07:27 PM
They, and all other net neutrality advocates I know of, want the network to remain application-agnostic, so providers can't distinguish applications that potentially complete with their products and deliberately provide poorer QoS for those applications.
You solve this problem by mandating neutrality with respect to source and destination addresses, not by application neutrality (although what we're really talking about here is protocol neutrality). I'm all for this. But there's real stuff that has to be done to certain protocols to make them work right, irrespective of whether they have extra QoS policies applied to them. You simply can't restrict an ISP's ability to manipulate traffic for those legitimate purposes and expect still to the have the internet that you probably want.
Posted by: TheRadicalModerate | September 22, 2009 at 07:42 PM
That's the political issue, but it's not the technical one. End-to-end services need end-to-end QoS, not just last-mile QoS.
No, that is not true in practice. If it were, then VOIP would be unusable today. In practice, VOIP works fine provided the last mile isn't congested.
VoIP is not as much of a problem (although wideband VoIP likely will be). Real-time video, however, is going to be extremely difficult. (Note the difference between streaming stored video, where you can buffer it as much as you need to overcome jitter, and conversational video, where you can't.)
You've got to be kidding me. Wideband VOIP will be a problem? That might be true...if the human auditory system suddenly made a quantum leap in evolutionary development in the next year. In the real world, wideband VOIP is going to consume a factor of two or maybe three more bandwidth above regular VOIP. With compression, we're talking say, 5 kilobytes per second. This is nothing.
I just don't see a huge demand for real time video. People can watch youtube videos for hours but most people just can't tolerate videoconferencing for very long. Videoconferences are either small intimate affairs, which are difficult to sustain much psychologically or they're large broadcast style events where most people never talk and for which most people don't need real time requirements. I mean, if you're watching a lecture, do you really care if it is delayed by 2 seconds? Either way, we're not talking about huge numbers.
How 'bout if you made all service more expensive because the provider had to upgrade his network to remain competitive? Look, I'm not arguing that you can't solve the problem by throwing (low-latency) bandwidth at it. I'm arguing that that's a really stupid way to engineer an internet.
You haven't demonstrated that lack of bandwidth in the core is a problem.
Ever hear of lip sync? You have to delay the audio to keep it sync'd with the video. If you delay the audio too much, the end-to-end latency makes conversation difficult-to-impossible.
How much bandwidth do you need to track the 1% of the image that covers the speaker's lips? Videoconferencing involves video that compresses well.
ISPs charge a flat rate not because bandwidth is free, but because it's easier to average the cost of their entire subscriber base. The subscribers like it that way because they have a predictable bill. But it's still all about total capacity.
ISPs charge a flat fee because human beings prefer to buy the freedom of not having to worry about overage charges and are thus willing to pay more than they would if they bought metered bandwidth. There's no reason that ISPs can't charge people for metered service with an extra percentage to cover network investments and maintenance. This is what every tier 1 ISP does.
Here's Wu and Lessig's letter to the FCC, from 2003. In it: "A network that is as neutral as possible is predictable: all applications are treated alike." This is not cherry-picked; they really want application-agnostic (and hence protocol-agnostic) treatment.
You're reading stuff that isn't there. That sentence does not in any way conflict with my example about respecting TOS bits.
In fact, you're so reasonable that you're having trouble understanding that there are lots of unreasonable people out there.
Of course there are some unreasonable people in the world and some of those people have crazy notions about Net neutrality. But those people aren't exactly dominating the conversation. I mean, Google is strongly advocating for net neutrality: do you think Google's corporate views on the subject are crazy? Do you think Google has less influence on FCC policy than some random crazy guy you can dig up screaming in a comment thread?
Posted by: Turbulence | September 22, 2009 at 08:58 PM
But there's real stuff that has to be done to certain protocols to make them work right, irrespective of whether they have extra QoS policies applied to them. You simply can't restrict an ISP's ability to manipulate traffic for those legitimate purposes and expect still to the have the internet that you probably want.
Which protocols? I'm having trouble understanding what you mean. Routers generally don't have lots of cycles left over for protocol specific processing which is why it is not generally done.
Posted by: Turbulence | September 22, 2009 at 09:56 PM
@Catsy -- You had a very good point. If I could simplify a bit. What Comcast did was basically the same as the Post Office opening envelopes you send and throwing out things that they unilaterally and secretly don't like.
@RadModerate -- On the point about "filtering already exists"... this is where the difference b/w openness and neutrality is most clear. Yes, some types of network management (DOS attacks, viruses, etc.) are necessary. The point so much isnt the management, but CHANNELING MANAGEMENT IN A CONSTRUCTIVE, OPEN, AND TRANSPARENT DIRECTION.
I don't think anyone has big problems with Comcast's revised management plans (though depends on getting more details, as I recall). The problem with what Comcast did is that they did it secretly and w/out consulting with standards-making bodies, etc.
Open networks, not necessarily "neutral" ones. And that's why the FCC's 6th principle on disclosure is a good one. it helps "crowd-source" enforcement.
Posted by: publius | September 22, 2009 at 10:16 PM
To justme. It is not Terry Gross. It is Lisa Mullins. I had to get some help at home in dredging up her name.
Posted by: Doran | September 22, 2009 at 10:20 PM
More brilliant insights from The Radical Moderate:
There are some people with simultaneously so much invested in conservatism yet so much embarrassment at the failure of conservative government, that they insist cloaking themselves as "moderates" in order to continue their moral crusade to mindlessly bash liberals.
Posted by: Tyro | September 23, 2009 at 12:07 AM
As much as you might like it to be so, your conclusion does not automatically follow from your premises. Repeatedly asserting it as if it did really isn't persuasive.
Posted by: Catsy | September 23, 2009 at 12:57 AM
Somebody isn't paying attention to the reporting that IS out there.
Posted by: gwangung | September 23, 2009 at 01:09 AM
Tyro, I'm so flattered! Nobody's ever taken the time before to attempt to discredit me with my own blog!
Please note that I'm about 85% in agreement with what's being said here on network neutrality, and am only trying to point out that you can have things go petty far wrong if you're not very, very careful about how you regulate. I really didn't mean to trigger all your orthodoxy antibodies.
Well, OK, maybe I did...
Posted by: TheRadicalModerate | September 23, 2009 at 10:01 AM
With all due respect, 'RadicalModerate', Tyro's quote from your blog reveals *your* orthodoxy. In case you somehow hadn't noticed, the fundamental problem with political liberalism (in the US sense) is that liberals themselves have such a strong tendency to see all sides of an argument, that they can forget what they were working for in the first place. In other words, American liberalism ends up being sort of about nothing, in a strictly ideological sense.
For liberals, life is life, and politics is just a means to improve life. For insurgent Rightists (I don't want to call them 'conservatives'), politics is life, politics gives life meaning, emotional excitation, etc. Guess who has the more emotionally compelling rhetoric? Guess who is more disciplined politically? Not the liberals, who resemble a herd of cats...
Tyro's quote shows *you* to be extremely - almost comically - orthodox in your political views. And BTW, that does probably bias you about this network stuff, although I'm glad you commented here, because it generated an informative discussion (informative for ignorant people like me, at least).
This latter is kind of an object lesson as per my basic point: insurgent Rightists, from Rove on down, know perfectly well that liberals *aren't* very orthodox or rigid, and in fact, that the surefire way to defeat liberals is to appeal to the latter's 'but on the other hand' instincts.
You see it on this blog constantly [what follows is not a slam on topic drift, which is a feature, not a bug, of blogging]. Moderate liberal makes a moderately liberal argument-and-assertion in a post; doctrinaire GOP apologist responds with something provocative and factually dubious; liberal commenters fill up a chunk of thread debunking provocative, errant, statement; GOP apologist, while never admitting error, changes tack and picks an assumption to challenge, and liberals dutifully back up; when that has been exhausted, he picks another one and liberals back up again. I'd call it a form of 'sandbagging'. The GOP apologist doesn't argue (ie in good faith) much, if at all, on the merits of the original issue, because that's not his goal. The goal is to piss liberals off and to diffuse them - pretty easy to do.
I'd watch what you call the Kettle, dude.
[this is also not a slam on any particular conservative posters or commenters on this blog. If the shoe fits, wear it - if not, don't]
Posted by: jonnybutter | September 23, 2009 at 11:30 AM
And BTW, that does probably bias you about this network stuff
Sorry, that was very poorly worded. I mean that your political orthodoxy is relevant to network-views.
See! There I go!
Posted by: jonnybutter | September 23, 2009 at 11:36 AM
jonnybutter, tyro--
I apologize for the cheap shot at tyro. It certainly didn't further the debate, and I completely agree that you could construe what I did as an attempt to take everyone's eye off the ball, although that was not my intention.
Posted by: TheRadicalModerate | September 23, 2009 at 01:00 PM
thanks for disposing about this case, RadMod, but I was making a more general point. The above scenario is how orthodox insurgent Right types can come to accuse Democrats of being both weak/soft AND brutally rigid, communist AND fascist, etc. etc. People who get paid to enforce and re-enforce a ridiculous Party Line are one thing, but I don't understand why others do it.
Posted by: jonnybutter | September 23, 2009 at 01:17 PM
Hey, I'll certainly cop to my own set of center-right orthodoxies and prejudices. I certainly won't engage in any debate over why liberals are better than conservatives or vice versa. In my experience the level of craziness is pretty much directly proportional to the extremity of your position on the left-right scale, irrespective of which edge you're hanging ten off of.
Posted by: TheRadicalModerate | September 23, 2009 at 02:54 PM