These aren't analogous at all. The situation that Net Neutrality is about preventing involves four parties: a website, a comsumer, the website's ISP or datacenter, and the consumer's ISP. Call these W (website), C (consumer), D (the website's ISP), and I (the consumer's ISP). There are network connections between these parties, and the graph of who has a network connection to who looks like this:
W---D---I---C
The website and the consumer each pay their respective ISPs for connectivity. Those two ISPs typically have a connection between them, in an arrangement called "peering". If they don't, then instead they will both pay a third ISP (a backbone provider, B) to bridge the gap. So the graph of who pays who for network connectivity looks like one of these two arrangements:
W-->D I<--C
W-->D-->B<--I<--C
If the connection between W and D is too slow, then W pays for an upgrade. If the connection between C and I is too slow, then C pays for an upgrade. If the connection between D and I is too slow, then D and I haggle over who pays, under threat of losing business from W and C.
Now suppose a I, a consumer ISP, decides to block or slow down traffic coming from W. The key fact of this situation is that in the original state of affairs, there is no business relationship between W and I. The website only has business relationships with C and D. Basically, I is threatening to sabotage W's relationship with C, by degrading C's service, unless W pays up.
Jim, this is a really fantastic breakdown. thanks. The part I still don't understand is - how can I threatening W other than by violating their terms with C? Doesn't C, the party who pays I, ultimately pay for a service spec? Is the spec too loose such that it allows I to throttle specific W's?
I think there's a good argument for making the Internet be an utility like electricity is, but I think that also means there won't be anymore competition effectively (at least in terms of deployment), and they will upgrade their networks much more slowly. So the money for upgrades might have to come from the government (which actually did happen already, but with few results:
That being said, we also can't just say "let the free market deal with it", because right now there's no free market in the broadband space. ISP's have managed to basically outlaw competition from their areas, so before we even begin to accept the idea of "competition is the solution, not regulation", then we need to actually make that true, and make sure all such restrictions against new entrants are lifted, and generally make it much easier for new competition to appear and exist.
If it becomes super easy to move from one ISP to another the moment one of them pisses you off, and there are at least several good choices, then this could work.
> I think there's a good argument for making the Internet be an utility like electricity is, but I think that also means there won't be anymore competition effectively (at least in terms of deployment), and they will upgrade their networks much more slowly
It depends. Some of my academic work looks at network neutrality and ISP investment incentives in the long-run (as well as the effect on content providers and consumers) when the ISP is a geographic monopoly. The investment incentives (improvements to bandwidth) should theoretically improve over a net-neutral policy if network neutrality is dismissed, but it's not guaranteed to be all that great.
Intuitively, if we really wanted to see innovation, we should do away with the typical geographical (natural) monopolies we see with Comcast, et al. For example, for years Austin plagued with internet problems and (especially) relatively low bandwidth offerings compared to places like Dallas (with FiOS, etc.). However, quickly after Austin was chosen by Google as a place for Google fiber, AT&T began offering GB/S speeds (as part of Uverse? Not sure the exact name). The point is -- if we want to see improvements in the network, opening up competition will encourage firms to innovate.
A fun story to consider is how MCI started out against the monopoly AT&T.
The incumbents don't currently upgrade their networks where they have no competition. My neighborhood's DSL (USWest -> Qwest -> Centurylink) has been at the same speed since 2000, and Comcast wants some prohibitive amount of money for a SOHO setup, which is the only way they give you a static IP address. Regulation would probably be more efficient than the status quo right now.
Could you clarify the following and maybe provide a link?
"right now there's no free market in the broadband space. ISP's have managed to basically outlaw competition from their areas, so before we even begin to accept the idea of "competition is the solution, not regulation", then we need to actually make that true, and make sure all such restrictions against new entrants are lifted"
It would be nice to put together a spectrum of possible outcomes if net neutrality regulations were completely repealed. Here's what I think.
Peer-to-peer file sharing is gone or throttled way back. ISPs have no reason to support it since it takes up so much bandwidth and there are no stakeholders with enough money/clout to fight back.
Microsoft starts paying carriers for QoS for Xbox to use it as a differentiator against Sony. Maybe they try a similar tactic with Skype but maybe the carriers don't want to go along with that since they have their own voice offerings.
I don't think streaming music services are affected at all. The bandwidth is so low and it's relatively insensitive to latency. It seems like it would be excessively cruel for ISPs to seek out and block arbitrary streaming services.
Maybe video is different? Perhaps they could charge video services (Youtube, Amazon Prime, Netflix) slightly more for high def streaming to subscribers.
Sports. ESPN already has a sort of self-imposed reverse net neutrality thing with ESPN3. Hard to see much changing there. I don't think there really any VC backed upstarts here to speak of either.
"I've got this thing and it’s fucking golden, and I’m just not giving it up for fucking nothing."
Occasionally statistics are released by companies on how much traffic leaves their site when it is 100ms slower, or the percentage drop in sales conversion rates. Furthermore, the sensitivity of visitors seems to be increasing.
And yet I think it would be very difficult to argue in court the specific harm caused by 50ms of added latency in 2020. Consumer ISPs could easily offer 'prioritization' to companies for a fee. Perhaps a Facebook of 2014 could make a subtle case measuring the harm of low priority, or cover the nominal fees to get that prioritization, or get users to complain to their ISPs. However, a company equivalent to Facebook of 2006 on the other hand would just have its growth slowed by a degraded user experience.
Already the structure of the industry for consumer ISPs encourages rent-seeking behavior. The slow spread of gigabit-tier broadband is likely related to those dynamics. Altering the landscape for content availability and latency, so that it is not neutral and is likely to favor incumbents, seems like it does not optimize the internet for innovation or the growth of new internet properties.
You're saying you think ISPs would try and penalize all traffic by 50 or 100ms and let the big sites pay up to avoid the penalty? That seems like mustache twirlingly cruel and I'm not even sure you can do that with conventional network hardware (you'd have to buffer all that traffic in RAM).
On the other hand, if we're just talking about allowing certain companies to pay ISPs for prioritizing traffic for certain applications (VoIP, gaming), then I don't think that's such a horrible thing.
I'm not just saying that not doing so is equivalent to leaving money on the table, I'm saying that the net costs to the economy would be greater than the revenue collected from content providers. In a certain sense the revenue collected from content providers could be seen as offsetting revenue that would otherwise be collected from subscribers, and may lead to faster deployment of network improvements. However, that neglects the full cost to the internet as a whole, which would be the increased barriers to entry for new companies.
Blagojevich got in trouble because he pushed the envelope so far, but I used the quote because it reflects the reality that not cashing in on some opportunity is equivalent to losing money, and particularly so if you have a responsibility to shareholders.
Agree that we need to look at concrete deal terms in question and talk about what's actually anticompetitive vs. what is just sound decisions by self-interested economic actors. It seems to me that if torrent users are paying for connection parameters that account for the specifics of their usage, there shouldn't be any problems. I have tried to find good detailed accounts about the torrent vs. comcast battle, but haven't found any really detailed technical and economic accounts of what's been happening and why. Anyone have a good link?
> Suppose you want to compete with Amazon, and you want Amazon to pay the same UPS rates you pay. It may make sense for Amazon to acquire UPS, or build a competitor to achieve rates that are more natural at their scale. [...] Vertical integration like this isn’t necessarily anticompetitive. Amazon, for example, runs AWS, but doesn’t appear to have engaged in anticompetitive practices with prospective competitors who run on AWS. Nevertheless, the scenarios above could be more undesirable for competition than the initial unregulated scenarios.
The article resume is "don't be pro regulation, because X company might acquire W company and create a vertical company?", because that is really not smart.
If it makes sense for Amazon to acquire UPS or run it's own courier, since it's not illegal to do so, be sure that it will. The Net Neutrality has nothing to do with this. Corporations don't do any favors, last time I checked.
According to the author's reasoning we should let any bill pass by, because hey if we push hard on the other direction, only God knows what other solution might find corp-X to make money, better the devil you see and we could go on like this for ever.
Hence:
* We should leave the NSA do it's job, because repercussions of our action - of protesting - might be grave!
* We should have accepted the DRM on mp3s, because Sony might buy an ISP.
> Sure enough, Microsoft arguably did become complacent, and that resulted in the well-documented shift in power from Microsoft to Apple and Google. Anyone seen an antitrust discussion centered on Microsoft recently?
Microsoft was under DOJ scrutiny[1] up until 2007, and that's the timeframe when their competitors grew surprisingly strong. Coincidence?
Pretty sure the answer is yes and it has a lot more to do with what Apple and Google were able to do. They didn't gain share in any PC OS or PC apps markets where microsoft was dominant, they just made those markets less relevant.
And, yes, an even better solution is competition in ISP space.
But, we need to get money out of politics for that to happen, so we can't hold our breath!
(For example, big ISP's have bought senators to block cities from running their own free wifi networks, passing state laws prohibiting municipal ISP's!)
Interesting article, although right when I thought we were going to get to some interesting suggestions, the article ends.
So let me finish it for him, what we need is more dynamic and realistic competition in the last-mile ISP business. I believe point to point high-frequency bandwidth can help solve this problem. Metropolitan fiber with sub-leasing capacity to ISPs that carry the traffic is also an option. But the natural monopoly position of last-mile wiring has always been a major issue. Hence regulation of phone companies.
In dense cities with irregular geography, the wireless line of sight can work. Eg: Webpass, monkeybrains in San Francisco. In flatter cities, this may not work as well. In less dense areas, the installation cost usually means a natural monopoly, but towers, line of sight wireless may help there.
The problem is all the LOS wireless stuff has been more of a bespoke solution rather than a scalable one. Tech has helped reduce the hardware cost, but tuning and aiming antennas - does this work for a city like, say, St Lewis?
To summarize, you're saying that none of these look like legit solutions broadly, but a hodgepodge of different solutions in different locations can create interesting options for competition against the natural monopolies, but these natural monopolies may remain in some areas regardless?
Net neutrality basically says, these are the legally acceptable network protocols and architectures. It inevitably requires a specific technical definition of acceptable traffic patterns and network uses, either by Congress or by case law.
So much of what we think about neutrality is based on a hand-wavy assumption of natural monopolies, that our current lack of last-mile competition is inevitable. At different times, we thought the same of package delivery, voice calling, sending messages. The things that obviated them were not obvious, a la “faster horses”.
Thought about this way, neutrality legislation actually locks in the incumbents’ current arrangements instead of opening them to (so far unimagined) asymmetric competition.
No, net neutrality is just a restatement of the End to End principle - network intelligence should be kept at the hosts. The only things routers should be doing are forwarding packets based on daddr, and dropping packets when queues are full (and I guess egress filtering these days).
Having said that, I'm sure that any FCC-created net neutrality would devolve into blessing specific protocols, defining protected classes of businesses, specifically excepting p2p, etc.
So are you saying you think no legislation can help here, or do you think there is something that can enforce e2e but it just wouldn't get through in practice without dinging certain use cases?
In general, I don't think legislation ever helps. The political meat grinder perverts a grassroots call for sanity into more red tape that supports the incumbents' status quo. Keep in mind that the large companies you might be tempted to say would support net neutrality are actually incumbents and would love to keep competition from springing up, especially if they were based on compelling non-http technologies.
If you think otherwise, please take my first point and run with it. I personally think the only way forward is to get more encrypted peer-talking traffic-indistinguishable apps into more people's day-to-day use. Then there's an actual market demand for real Internet access.
agree. while people may sometimes have valid concerns, its the chain of reasoning that leads to proposed ways of addressing those concerns that often isn't grounded in realistic economics.
Now suppose a I, a consumer ISP, decides to block or slow down traffic coming from W. The key fact of this situation is that in the original state of affairs, there is no business relationship between W and I. The website only has business relationships with C and D. Basically, I is threatening to sabotage W's relationship with C, by degrading C's service, unless W pays up.