The major providers like to pretend that they're in a wonderful competitive market where, if only they were allowed to discriminate on traffic more, there would be more competition, prices would be lowered, and investment in infrastructure would grow due to the competition.
This is a nice rosy view of things, but it's just not accurate. They're de facto infrastructure builders, and for their efforts they want a monopoly on that infrastructure, with all the powers it comes with (e.g. favoring some traffic over others). The problem is that everybody long ago realized that infrastructure monopolies are a bad idea because they destroy both competition and infrastructure growth. It's why major roads aren't owned by private companies, it's why the electric company can't charge you more if they don't like the brand of your TV, and it's why public transportation services like trains and buses can't charge you more if you're partners.
Letting private businesses compete in various niche ways is fine as long as they are (1) actually private and (2) actually in a working free market. Telecoms in the U.S. are neither. They're so heavily subsidized and regulated that they function more as pseudo-government entities who should be entrusted with very serious business of maintaining public infrastructure. But they should definitely be prevented from rampantly discriminating against their customers to profit. The whole point of infrastructure is that it is a common public good, available to everyone equally without discrimination.
>>" It's why major roads aren't owned by private companies"
In the US, many major roads are owned by private companies like Cintra [1] to give one example, and they charge tolls [2] on those roads.
It's become common practice in many US states that any new highway construction must have a toll on it and then the state sells the road to a private company.
This is actually a great illustration of the reason we have common carrier laws in the US.
Should we allow Cintra to let you go 100MPH instead of 70MPH if you pay a higher toll on the Indiana Toll Road? Maybe every 10MPH is an additional $1. Or maybe Cintra knows that Tesla drivers are rich, so their tolls could be set 10x higher than a Kia's toll. Or maybe people with Illinois license plates are on average more wealthy than people with Ohio plates, so they can pay more too...
I think it's unfortunately down to the incessant anti-tax rhetoric in the US. It's a lot easier to get a private company to finance a toll-road with a concession agreement than it is to pass bond measures or increase taxes.
I don't follow this logic. Toll roads allow the people who want to use a road to pay for it. Tax-funded roads take taxes from all the district's citizens to pay for roads that only some will use. Great if you use that road. Not so great if you don't.
This is insane. By this logic, public goods don't exist. Tax-funded fire departments are great if your house catches fire, not so great if it doesn't, etc.
It's more a recognition that roads aren't automatically public goods. Roads in general tend to benefit everyone, sure. But something like say the Georgia 400, which just exists to funnel suburban commuters into Atlanta? That's better off as a toll road, paid for by the people who use it.
Indeed. I have no children, why should I pay taxes to educate other people's kids? Just because they're our future laborers working on and within our common economy to keep things going after I've left the workforce.
Of course there are public goods, but your fire department analogy doesn't work. We have fire departments standing by in case there's a fire. So, sure, a fire department that services your neighborhood is great, but you probably wouldn't want to pay for one that doesn't service your neighborhood.
"About 70 percent of the construction and maintenance costs of Interstate Highways in the United States have been paid through user fees, primarily the fuel taxes collected by the federal, state, and local governments."
And about 1/7 of the fuel tax revenue is redirected to public transit, rather than highways.
A moderate increase in the fuel tax[1] could make the IHS completely self-sustaining for both capital and maintenance expenditures. Unfortunately, raising the gas tax is typically a political loser.
Local roads are predominantly paid for through local property and other taxes, but that's easier to justify, since everyone takes advantage of local roads even if they don't drive.
[1] At least until electric cars become commonplace, in which case I'd probably be in favor of a simple mileage tax which could achieve the same effect.
The real problem here is that many advocates of net neutrality frame their support of neutrality strictly in terms of enabling free-market innovation. Perhaps because they're afraid to lose the support of the many libertarians who still walk among us, net neutrality supporters are afraid to give a fully fleshed-out narrative of what should happen: the government either makes a massive direct investment in increasing information bandwidth (as it has done with transportation bandwidth), or it assimilates the telcos and makes them do so. Finally, the current wave of market-based innovation can continue, supported (as markets always are) by a robust public infrastructure. There is a properly dialectical relationship here: you can't have one of these moments without the other, even though they are contradictory.
Is this plan perfect? No, but it sounds a million times more preferable than Comcast and Verizon picking which startups live or die.
Looking up the numbers it's still a tiny, tiny minority of roads in the US. Cintra's worldwide holdings are ~1% of the total length of major highways in the US. Significant, but not all in the US. And many toll roads are publicly owned. Or under private contract for X years so the taxpayers didn't have to take out a bond or front $Y million for the project.
I had a bit of back-and-forth with Marc on this topic and I think he's underestimating the difficulty of regulating a non-neutral 'net such that competition (and, therefore, innovation) isn't stifled.
I can't help thinking that if, 20 years ago, Microsoft had struck a deal with the major ISPs to prioritise packets being downloaded to Internet Explorer over those being downloaded to Netscape browsers, Marc would have had a very different take on net neutraliy.
Exactly. I think that the woeful state of US broadband and wireless infrastructure today is the strongest proof of the failure of a poorly regulated private sector that effectively has captured telecom governance. That sunny world of virtual worlds and holographic telepresence will happen in Singapore a lot sooner than in the US if government allows the Internet to look like Cable TV.
I consider Andreesson to have zero credibility in this matter, for reasons stated above and more. Why not allow rich and powerful corporations to control access to US consumers? Our smartest consumer startups, denied US consumers, will start to build Twitter and Pinterest and other services for consumers outside of AT&T and Verizon's control, and we will be more and more irrelevant as a market and a technology innovator.
It's weird to me to blame the "state of U.S. broadband" on the telecom providers, and in the same breadth compare it to Singapore, a city-state where all 5 million people live in a single urban area of 270 square miles (i.e. a bit smaller than New York City with a population density a bit higher than San Francisco).
The state of broadband and wireless in the U.S. is proof only of the fact that it's much more expensive to build infrastructure to a huge, sparely-populated country with large suburban, exurban, and rural populations. Half of South Korea lives in the metro area of the largest city. About 30% of people in England live in the metro area of London. About 25% of people in Finland live in the metro area of Helsinki. Only 4% of the people in the USA live in the metro area of New York City. Beyond that, more than half the population of metro London lives in the city proper. For a typical U.S. city like Boston or DC or Atlanta, its more like 15%.
> I can't help thinking that if, 20 years ago, Microsoft had struck a deal with the major ISPs to prioritise packets being downloaded to Internet Explorer over those being downloaded to Netscape browsers, Marc would have had a very different take on net neutraliy.
Interesting point. However, imagine if regulators had dictated the web browser as a "common carrier" and defined exactly what it can and can't do, we might never have had all of the new tech that makes the modern web so much better than the IE5 days.
Twenty years ago, those major ISPs were using the telephone system (via dial-up), which was designated a common carrier. If ISPs had to convince everyone to buy into a new cable or dish transmission system because the tel-cos could block their traffic over the phone lines or charge inflated fees for it, where would we be now?
Interesting point. But I have too much faith in human ingenuity to believe that without a common carrier telephone system we wouldn't have an internet right now. Maybe the path and details would differ.
Faulty frame. Try it with other basic infrastructure: "The more money I think I can make by owning ROAD infrastructure, the more I will invest to build it now."
I do not see the flaw. The statement in quotes, while a simplification, is correct if a private company is making the decision. If a government is making the decision (typical with roads, not with broadband), there are a ton of other factors, although in an ideal world they would also try to do a similar NPV calculation. In any case, the government doesn't own broadband.
So the argument seems to be that the more money these companies believe they can make, the more likely they are to build infrastructure that will ultimately reduce the costs for consumers.
This is a highly idealized view, which does not mesh well with synonymous real world systems like cable TV and cell phone networks. Those two examples alone prove to me that these companies would prefer to collude to keep prices artificially high than to use their monopolistic infrastructures to provide cheap and high quality service to consumers.
Besides which, America has already paid $200 billion to these companies in the name of infrastructure improvement, and all they did with that was take the money and run.
What is the technical definition of net neutrality? I look at a lot of these articles and they seem to imply people are using the same term for multiple things.
Indeed. You can imagine pretty expansive definitions, and whatever definition you come up with can create potentially undesirable incentives.
I don't know whether Comcast is using IP to deliver their video-on-demand services to your cable box, but let's assume they are. That IP bandwidth is above and beyond whatever you're paying for as a "broadband" connection, but that's just accounting: It's all packets in the end. Is it "neutral" for Comcast to be able to use that bandwidth to provide you with movies and TV while restricting NetFlix to your capped/metered "Internet" bandwidth?
If NN would disallow such restrictions, then Comcast really is a low margin dumb pipe provider. Some people love that idea, but it doesn't encourage any of the existing broadband providers (maybe sans Google) to keep spending tens of billions of dollars to roll out faster and faster networks.
If NN doesn't apply to these services, then of course Comcast will simply package up as much of their IP-based applications and content as possible into proprietary interfaces with your TV and other devices while making the bandwidth available to "open" IP applications commensurately smaller.
Im not sure how the second case incentivizes them to build out their networks either. They have barely done that for the last decade or so (consider how expensive and slow internet access in the US is compared to much of the first world), so I'm not sure why they would start now that they have even more control.
"Barely done that for the last decade or so"? I don't think you remember what broadband connectivity options were available in your typical American suburb in 2004 or 1994. As rayiner pointer out in a thread a day or so ago, telecom providers are some of the largest capital spenders in the nation:
Because it improves the quality of delivery of their service (e.g. video-on-demand). The Internet user won't notice the improvement - their Internet speed would remain the same, by metering - but the actual pipe improves so they can deliver more/better video (and reap the profits from it).
What's are the points of comparison? It's not really fair to compare average speeds in the U.S. with those in South Korea or Japan, where residents are heavily concentrate in a handful of major urban areas. I don't think internet access in the U.S. is slow compared to say Australia or Canada, which are similarly suburbanized and spread-out countries. According to this report, we're 11th in the OECD for average download speeds: http://cyber.law.harvard.edu/sites/cyber.law.harvard.edu/fil... (page 30), ahead of Canada, Australia, and the U.K.
Network neutrality rules prevent the network providers from cutting preferential deals with content providers. The pipe providers must remain neutral regarding the data flowing through their pipes.
If you want to know what the world looks like without network neutrality just go back to the days of Compuserve, Prodigy, Genie, AOL, et al. Dismal does't even begin to describe the situation.
We could ultimately replace our internet infrastructure with something more distributed which makes it difficult for ISPs to determine what traffic they're carrying. Ultimately traffic analysis works right now because the ISPs can detect where the traffic is coming from. In a more distributed internet, that wouldn't be the case. If all packets are encrypted, it would also make it harder to snoop and shape traffic that way.
Bit Torrent already does this with Message Stream Encryption/Protocol Encryption, and it can potentially hide on port 80/443. But why stop at just Bit Torrent traffic? Why not build a new web out of something more akin to this?
Hiding the source or destination of traffic (which is one of the major things ISPs want to discriminate on) requires onion routing which will make things ~2x slower and more expensive. If we end up in a situation where a significant fraction of Internet traffic is being onion routed I would consider that a massive failure.
I actually wasn't trying to suggest onion routing, but a more peer-to-peer distributed internet, similar to the way bit torrent works. If a CDN was spread across multiple sites instead of concentrated at one place (ala S3), it would be very difficult for an ISP to track.
I don't know how Marc squares his free markets love with his presence at the table being substantially enabled by the government (internet + UIUC + NCSA).
Surely people have not forgotten what it was like pre-neutrality (Compuserve, Prodigy, Genie, AOL, et al)?
I don't really understand the fear. The Compuserves and Prodigys of the world died because they weren't open and it didn't take a government dictate to kill them.
Openness won in the early iterations of online services in part because there was a neutral medium over which you were able to connect to your choice of service: the PSTN.
I'm not talking about why they died (that's obvious, the non-corporate internet was much better). I'm talking about how lousy the offerings were when the pipe providers could do whatever they wanted.
Completely OT: Is there any reason why all their content is SCREAMING? I'm stuck on IE 8 in the office, do they have some js or something that produces normal text on non-outdated browsers?
This is a nice rosy view of things, but it's just not accurate. They're de facto infrastructure builders, and for their efforts they want a monopoly on that infrastructure, with all the powers it comes with (e.g. favoring some traffic over others). The problem is that everybody long ago realized that infrastructure monopolies are a bad idea because they destroy both competition and infrastructure growth. It's why major roads aren't owned by private companies, it's why the electric company can't charge you more if they don't like the brand of your TV, and it's why public transportation services like trains and buses can't charge you more if you're partners.
Letting private businesses compete in various niche ways is fine as long as they are (1) actually private and (2) actually in a working free market. Telecoms in the U.S. are neither. They're so heavily subsidized and regulated that they function more as pseudo-government entities who should be entrusted with very serious business of maintaining public infrastructure. But they should definitely be prevented from rampantly discriminating against their customers to profit. The whole point of infrastructure is that it is a common public good, available to everyone equally without discrimination.