> Gillis Cashman, a managing partner at MC Partners in Boston, says it makes sense to charge extra to big content providers like Netflix, whose services at peak hours can sometimes consume more than 30 percent of total Internet traffic. Video is “significantly congesting these networks, and causing real issues for carriers where they have to spend a lot of money upgrading networks, and pushing fiber deeper into their networks,” he says. “There is currently no model for monetizing that required investment.”
This argument doesn't make sense to me. The ISP's customers are using the ISP's network by downloading Internet content. The nature of the content should be irrelevant to the discussion. The ISP's customers are purely responsible for that usage, and it is part of - from what I understand - the service they are already paying for. The ISP's network is only "congested" if it is incapable of delivering on the service that was advertised (assuming for the sake of argument, some particular speed of access).
Thus, how can Cashman argue that "there is no model for monetizing the required investment?" Isn't that exactly what ISP customers are buying when they pay for access to the Internet through the ISP? The onus should morally be on the ISP to make a good-faith attempt to build sufficient connections to any content that's popular enough to saturate the links providing it.
Furthermore, why isn't it naturally the ISP's responsibility to price their access sufficient to cover current usage and expected growth in usage? New use-cases for the Internet are constantly going to appear. It is the ISP's responsibility to provide Internet access with sufficient throughput that it is not congested for any reasons under the ISP's control; and to price their access sufficient for investment to achieve the throughput.
"The latest BigMac hamburger caused hungry residents to swarm McDonalds and caused drive-through lines to be congested today in Cloudville. Gillis Cashman, managing partner at McPartners, Inc. says that, despite selling happy meals in record numbers, delays in drive-through lines will continue, since since there is "currently no model for McDonalds to monetize the required investment" to upgrade its drive-throughs.
"Electricity supply companies have found that almost 30% of the total electricity capacity in their networks is being consumed by customer's light bulbs. In order to support future demand, the electricity supply companies will begin charging light bulb manufacturers for light bulb users electricity usage."
Who knows how well it would go over? I would be pleased to pay a metered rate, provided that the metering was standardized and easy to understand - like the electrical meter outside my home - and the pricing was fair. I would also expect there to be no speed tiers; everyone using a service gets the same nominal speed. I would not be interested in purchasing e.g. bandwidth blocks with f-you overage fees like what we used to have (still have? I don't do mobile data) with mobile carriers. That, I agree, would not go over well.
I think that the real problem is not perceived acceptance by customers, but that moving to metered billing makes the service look more like a utility. If I were running an ISP I would want to do everything I could to keep people from equating my service with water or electricity.
There's an efficiency argument against metering. Bandwidth is "free" unless service is near capacity, so introducing pricing that discourages use during that off-peak period can create needless problems. So maybe you only want peak-period pricing (kind of like how some cell providers give free minutes on nights and weekends).
This is already done at commercial providers. Those of us who operate networks are pretty well used to paying monthly rates based on 95th-percentile bandwidth utilization.
Internet service is metered. It's just metered in very large chunks instead of by the penny. Try going over your monthly bandwidth cap and see what happens.
The ISPs have total control over their subscribers' connection fees. If they set them too low to make a profit,* then they need to raise those fees and live with the consequences, not go trying to extort profits from new sources.
*I don't believe for an instant that the major ISPs are actually hurting, mind you. They just see an opportunity to make even more of a profit.
Ha! Sorry can't read statements like that without laughing/cringing. We all lost our "unlimited pipe" years ago .. in fact I'd argue we never had it in the first place only that the technology prevented us from hitting our limits in the early days.
That's one form of metered pricing, but I'd argue that granularity of a month is much too coarse to make much sense form a network utilization perspective. As I've said in other comments, pricing based on 95th percentile circuit utilization (which is common among commercial ISPs) is a much more practical way to balance efficient utilization with incremental cost.
For phones the demand is very different: a subscribers connection is either in use or not. When it is in use, it typically consumes no more than 64kbps at most. And it is easy to model expected peak demands in normal (non-disaster) scenarios because everyone that people might be calling tends to have a fairly low number of inbound connections, so the total peak bandwidth usage is still quite small - miniscule by todays standards.
When you're paid per subscription, and you at most carry 64kbps through a small local area before long distance fees kick in, it works.
Even so, in Europe (and I believe most of the world) landline usage is metered. It is the US model of free local calls that is unusual.
It's a bad analogy. If you buy a cell phone so you can talk to your grandma (who lives 1600 miles away) on the weekend, the phone company shouldn't be obligated to buy her a phone too. Not even if she only uses it for incoming calls.
If you're offering a service where you let other people interconnect their computer networks to each other, by definition you're only offering the service of connecting them to other people who (choose to) connect to the network.
If the two are to be connected, someone must pay. Consumers are screaming "they're violating network neutrality!" any time someone suggests they (the consumers) pay, so who is left?
Grandma just wasn't (until very recently) paying anything at all.
This is why we shouldn't conflate peering disputes with network neutrality issues. We're reading and hearing quite a bit that is abysmally stupid.
Comcast has no obligation to spend it's own money letting Netflix connect for free. The peering arrangement they had with Level3 was more than adequate to handle everything but the Netflix traffic, and now Netflix/Level3 is expecting Comcast to pay for the upgrades? Why? How is that fair?
I get sick of defending Comcast here. They are an abominable company that I hope goes bankrupt and out of business. But I keep reading really stupid things from people who aren't thinking through the implications of what they're saying.
The US operates a bill-and-keep [1] model for phone calls, which is why people pay to receive calls on their mobile phones, and why spam texts are more prevalent.
If a call is made from Verizon to AT&T, no money changes hands between them.
ISPs have typically operated a B&K model where the traffic is largely symmetrical between carriers. The argument arrives when the flows aren't symmetrical (such as streaming video).
As for Comcast paying to upgrade their network, there is precedent in the Cable TV model where they both pay for the content and pay to upgrade their network to receive, carry and bill it.
Networks derive value from the connections. If there was proper local competition then this would fix itself over the course of a contract cycle since a competitor would enter offering a network with a better value.
This is American capitalism, morality doesn't enter the picture.
More seriously, the McDonald's argument doesn't work because McD's business model is profit on turnover. If they sell more, they make more. Comcast's model is renting out wires. The money they make is the price set by the "market" in internet connections. Actually providing a service is a cost, and therefore they want to reduce it as much as possible without driving away customers.
Building more won't enable them to charge their customers more, so they're sending a rent bill to Netflix to see if they pay it.
"Actually providing a service is a cost, and therefore they want to reduce it as much as possible without driving away customers."
Brilliantly summarized. I recently had intermittent issues with my crappy Comcast (only viable option, where I live now) Internet connection. The technician who came over, after viewing my ping logs and things, immediately decided I was onto something, and was awesome. Like really awesome. He was representing the very best a very bad company. Felt bad for him. I directly told him "If you can, get another job, somewhere else [equivalent pay, benefits of course]."
I've had similar experiences with my [monopolistic, US] cable provider. Everyone who has come out to do any service has been surprisingly competent and very willing to do more than what needs to be done. I had a guy come out this past Sunday who didn't like the angle the cable was coming out the wall panel (sharp turn against furniture), so he moved the coax connection inside the wall. The only way I can figure it is they are heavily incentivized to solve it right the first time, so they like to cover all the bases.
"Building more won't enable them to charge their customers more"
Why not? If I'm noticing slowdowns due to saturation, and my ISP said "we're expanding in your area and upgrading equipment, capacity, etc, but it's going to add another $9/month on your bill" of course I'd pay it. A) I'd have no choice and B) I want better service.
And if you noticed slowdowns due to traffic shaping and your ISP with their government-enforced monopoly said, "We're racketeering. It's going to add another $9/month on your bill." of course you'd pay it. You'd have no choice.
Building more capacity and offering better service doesn't really enter into the picture in most American cities.
Point to even a single example or you earned every downvote you get.
I know a lot of people hate the "government," but it's no more rational than hating "corporations." Which government? Why? My impression of US government ISPs from college campuses is that most leave a wide open pipe for pretty much free use with few exceptions. Many forget that the Internet was entirely given away to private interests by the government. Most public ISPs are understaffed and may at worst limit something by accident while targeting illegal activity.
I suspect few people will - and there's a limit to this, if they add $9 every few months for a different service, eventually you will refuse to pay $$$/month for the service.
Imagine if every speed upgrade from 56k modems had also come with a price increase.
No, it's not, but if they can increment prices once, why not again? If they can announce extra billing for "infrastructure upgrades" every few years, why not more frequently?
Boil the frog as fast as possible. After all, where would it go?
And there's clearly some models since I can get 1gb internet in Seoul for like $40/mo and Google apparently has some kind of model a well. This just reeks of an attempt to try and build up support for greed and laziness.
Can we get rid of the FCC now it's pointless having it if they don't understand what is and is not the responsibility of the ISP regardless of content.
The entire FCC is entirely corrupt and in pockets of the cable companies.
It is not doing its job worse yet it is harming the system it was designed to protect.
"causing real issues for carriers where they have to spend a lot of money upgrading networks, and pushing fiber deeper into their networks"
I've been reading over and over again how the ISPs should already have this upgrade money allocated and available to spend on network infrastructure upgrades.
He can argue it because he has a lot of money (indirectly) invested in the outcome going his way. There will be winners and losers if net neutrality is abandoned and it looks like Cashman would be a winner. Most of the rest of us would be losers.
I hate when reporters find the most biased people in the debate and quote them as if they were an objective opinion. This reporter, to his credit, immediately follows the Cashman quote by saying "Some less financially invested observers have little sympathy for this argument" and then quotes Rob Faris, who may have strong opinions but at least isn't talking his book.
Every one of these articles should be mentioning the fact that in the 90's, the ISPs got $200 billion dollars to build a fast internet infrastructure that they never delivered.
Not only did they not deliver, but they intentionally sabotaged US efforts to stay up to par with other countries. US infrastructure is woefully behind compared to other nations (we're only recently starting to see fiber optics as a serious competitor to DSL). Even still, as old and outdated as US infrastructure is, the idea of a bandwidth limit or speed restrictions are imposed by the ISP, not actually a physical constraint. Sure, there are limits to the loads a node can handle in an area; But with each node typically serving 200 customers, at a conservative $30/mo... That's a $72,000 annual budget upgrade and/or expand THAT SINGLE NODE. A typical ISP will handle hundreds, if not thousands, of nodes depending upon a location.
We don't need fast lanes or bandwidth caps on the internet. We need an ISP that isn't ran by greedy tyrants.
> US infrastructure is woefully behind compared to other nations
Reconcile your claim with Akamai's State of the Internet, which shows the U.S. ranked #10 globally, at 10 mbps average connection speed: http://www.akamai.com/dl/akamai/akamai-soti-q413.pdf?WT.mc_i... (page 23). That number would put at about at #6 in Europe (see page 29), ahead of the U.K., Germany, and France, and not far behind Sweden and Ireland.
What bothers me about the data is the implicit distribution skews in the Peak and Avg numbers, but that's a problem with every line, so not material to the subject here. It may signal a methodology weirdness, but it's lunchtime and I don't feel like paging back to see if I can find something showing that Akamai is counting things it shouldn't be.
I always find it fascinating when people take the data and proclaim that America should be stacked up against nations with single digit millions of people living there, much less postage stamp size areas like Hong Kong or Luxembourg.
Most of the comparisons are hilariously ridiculous. The proper comparisons are: China, Brazil, Japan, Indonesia, UK, Russia, India, Germany, France, Turkey, Italy, Mexico - large scale industrialized nations.
Comparing Norway or Luxembourg or Hong Kong with the US, as though they can possibly be compared in any sane respect what-so-ever is absurdity to the ultimate degree.
The fact that we compete at all against tiny nations with 5 or 8 million people is a stunning accomplishment.
Half the population of South Korea lives in the Seoul metro area, which has the same population density as the city of San Francisco. The whole of Hong Kong is a single city, again about as dense as San Francisco. The whole country of Japan is about as dense as the state of Massachusetts, the 2nd densest U.S. state. It's about twice as dense as New York, the 7th densest U.S. state.
How do you get that? Massachusetts's average is 13.5 Mbps. Japan is 12.8. New Jersey, Maryland, and Delaware are 12 and up, above Switzerland, Sweden, Czech Republic, etc. The dense U.S. states compare very favorably to the top-5.
South Korea is the outlier because half the country lives in a single megacity with the density of San Francisco. The U.S. can't compete with that geographic advantage.
One thing to consider is that US households are quite often getting less than the advertised speeds (probably because ISP's don't want to bother installing more repeaters or it's difficult to do so). I.e. you can barely get 10 Mbps on a 20/1 Mbps package. That skews the average speed stats.
From my experience, that is not the case in the EU - you pay for 20/1 Mbps, you get those speeds +/- a few Kb. Which is also why many don't want to switch from DSL to fiber where it is available - "why would I need that, my Internet is fast enough".
No, Akamai's data is based on achieved speeds, not advertised speeds. It's the results showing European broadband to be faster and American to be behind that are based on advertised speeds.
The ISP's never got "$200 billion" in the 1990's. That's a total made up number, based on taking what ISP profits would have been had they been regulated as a utility, and calling everything over that "money given to ISPs."
The premise of deregulation was that it would lead to increased infrastructure spending. And it has: the late 1990's and the 2000's saw massive investment into cable and wireless. People assumed at the time the money would go into fiber, but demand exploded in wireless so investment went there instead.
They got tax breaks for $200 billion, so not made up. They talked about building fiber in congress to get the tax breaks. That's why people feel like they didn't live up to their bargain.
The number is calculated by starting from the premise that telecom companies should be regulated as utilities, and make regulated returns on their investment. It also starts from the premise that the price of features should be proportional to their cost.
From that premise, it computes $103 billion as the "excess profits" when the telcos are compared to other utilities. Another $78 billion is chalked up to "excessive depreciation." This, in turn, is based on a calculation rooted in "[a]ssuming that depreciation rates should have remained constant after divestiture." (Page 220). Of course the assumption that the depreciation rate should have remained constant before and after the internet boom is ridiculous. Another $25-50 billion is "cross-subsidization overcharging for long distance, DSL, and wireless." This estimate is based on the claim that the telcos added charges to phone bills that should have been accounted as costs to non-regulated services.
None of this amounts to getting a direct subsidy or even a direct tax break in return for building fiber. The whole point of the 1996 changes in the law was deregulation of the industry. The whole idea was that deregulation of the industry would lead to higher capital investment, which it did (just not in the areas people expected at the time). Nobody hid the ball on the fact that the 1996 law was about deregulation--everyone referred to it as such. Well, in a deregulated industry, if a company raises prices, we don't call that "excess profits." If they cross-subsidize their business lines, we don't call that "excess profits."
Did the telcos make a ton of money between 1996-present? Yes, they did. Demand for internet exploded, and demand for mobile exploded, and they invested the money to build the infrastructure that made all that possible. That's why they made all those "excess profits."
It is amazing - or very, very ironic - that you're able to justify your argument with an excerpt from the book entitled "$200 Billion Broadband Scandal," which appears to be aimed towards educating the reader about the large amount of wool pulled over their eyes re: the NII.
The book is where the $200 billion figure originated. Instead of throwing third-hand information around like everybody does in these threads, I went to the source to criticize its methodology.
The book, as a whole, is handwaving. It's based on two fundamentally flawed premises:
1) That the internet and mobile booms are unrelated to deregulation and should not be factored into the analysis. Of course, that's ludicrous. For decades up to the mid-1990's, telephone service was an industry with steady demand growing based on population growth. Then the internet happened, and mobile happened, and demand skyrocketed. At a basic economic level, why should we not, then, expect: 1) providers of the infrastructure underlying those boom industries to profit more after than they did before? 2) providers of the infrastructure underlying those boom industries to depreciate infrastructure and invest in new infrastructure faster than before?
The $200 billion figure is fundamentally rooted in assuming that the answers to those two questions should be "no."
2) That 1970's-style rate-regulation doesn't have an adverse impact on capital investment. If companies can make 30% returns investing in Twitter and 10% returns investing in laying fiber, then the money will go where the returns are. If these folks had been in charge of regulating wireless, we'd still be using 2.5G phones (but at least we wouldn't be "overcharged" for text messages!)
Were provisions for a particular level of broadband penetration and performance included in the 1996 Telecom Act? If so, was the inclusion of these provisions tied in any way to the more provider-friendly regulatory posture granted as a result of the Act?
Assuming the answer to first question is yes, were those provisions met? If not, was it because they were just unrealistic, or was it because of some other reason?
What I think: yes, yes, no and a little of both.
What I don't think: doubling down on deregulation at this point - which is what this fast-lane stuff seems like to me - will be a long-term benefit for consumers, which (as a consumer) is what I currently care most about.
To answer your questions:
- Of course infrastructure providers enabling a boom industry should profit from their efforts. I also expect the provider to upgrade their infrastructure out of these profits, not go around with their hand out trying to get everyone else to pay for it every upgrade cycle.
- I don't understand how this supports your argument. Isn't a part of the argument that they are not upgrading infrastructure fast enough?
- re: 70's style regulation... Did you say handwaving?
I'm not really disagreeing with you that companies should profit. However they want to try and do that is also fine, they basically have the right to try and earn profit however they want. But it's got to be OK to decide no, this method you're trying is unseemly and we're not going to permit it. That's what I think is happening here w/ Comcast, and I'd prefer if regulation was not changed (or established or whatever the appropriate term is) such that it's not only OK, but a big green light for _every_ provider to start running around doing the same thing.
Yet it has somehow become conventional wisdom, parroted in every thread on this subject.
> Were provisions for a particular level of broadband penetration and performance included in the 1996 Telecom Act? If so, was the inclusion of these provisions tied in any way to the more provider-friendly regulatory posture granted as a result of the Act?
No. The purpose of the act was deregulation of the industry, not giving regulatory concessions in return for particular performance outcomes. Whether deregulation is "provider-friendly" depends on who you ask. In theory, deregulation is consumer-friendly. E.g. whenever Uber or AirBnB come up, people on HN champion deregulation of those markets, on the theory that deregulation is good for consumers. That was the operative theory underlying the 1996 Telecom Act.
Now as for whether the pitch back then lived up to the results, the answer is of course "kinda." The state of U.S. internet is really much better than people say. Read my post about what Akamai has to say about actual connection speeds in the U.S. versus Europe (that we're ahead of the big European countries like the UK, Germany, and France): https://news.ycombinator.com/item?id=7709948.
What people fixate on is the lack of fiber in major cities, and that has much more to do with dysfunctional municipal politics than national telecom regulation. In Chicago, for example, many folks have a choice between Comcast and RCN (which has a mostly fiber network). This is not because Comcast didn't bribe enough alderman, but simply because the city tends to have a pretty functional and hands-off approach to regulation, which reduces the roadblocks to market entrants. The city also has never instituted rent controls, and unsurprisingly has pretty affordable housing for city of its size and density.
> What I don't think: doubling down on deregulation at this point - which is what this fast-lane stuff seems like to me - will be a long-term benefit for consumers, which (as a consumer) is what I currently care most about.
The problem with "pro-consumer" regulation is that it's often short-sighted. Throughout the U.S. in the 1970's and 1980's and 1990's, and in Europe in the 1980's, 1990's, and 2000's, industries were deregulated, and it was largely a positive thing. The biggest problem with "pro-consumer" regulation in telecom is that it makes investment in regulated telecom unattractive, which causes investment dollars to flow into areas that are more attractive.
When I say "1970's style regulation is discredited" I absolutely mean it. Governments across the political spectrum have backed away from that style of regulation and embraced ones better-rooted in economic reality. But utility and telecom regulation is rooted in that same failed ideology. It's the ideology where regulators think it's a good idea to set prices instead of letting the market do so, or tell companies that they can only achieve a 10% return instead of whatever the market will bear. This style of regulation ignores the basic fact that the economy is inter-connected. If you tell companies in an industry that they can't raise prices, all the capital will flow to an industry where they can.
That's exactly what happened with telecom. Post-1996, the most heavily regulated communications infrastructure was the telephone system. Cable was less regulated, and so was wireless. And what happened? All the capital flowed to upgrading cable and wireless networks, and DSL was essentially killed as a potential competitor because there was not sufficient profit motive in DSL.
Now, I'm not saying network neutrality is 1970's-style regulation, it's not. It's not rate setting, or regulated rates of return, or anything like that. But when people say "the telcos took $200 billion and didn't give us fiber" well that's absolutely a contention that only makes sense if you buy into the 1970's ideology: that the public would've saved $200 billion had the industry continued to be regulated the way it was pre-1996, but that despite these regulations we'd still have the level of infrastructure investment we do today.
If our government wasn't corrupt we should have used that money to build a government fiber service instead of giving that money to private shareholders.
I wish those articles referred to this by its proper name: it's not about creating an internet "fast lane", it's about creating "slow lanes" for people not willing to pay extra for the "normal lane".
On the good side, all this will mean in the long run is that the VPN industry will be booming. No bandwidth throttling via VPN, after all.
While they can't discriminate between types of traffic on a VPN, they can throttle all VPN traffic, and react to higher VPN bandwidth use. And probably will, if allowed to.
So if there were FCC rules that came out saying the ISPs couldn't have slower access than they currently do and they can't completely block access to any legal server, would you be ok with that?
Last I checked Wheeler said "if the ISPs abuse these new rules and put sites on the "slow lanes", I'll bring in the real rules (referring to Title 2 reclassification)".
So I wonder what's his reaction when he found out that the ISPs are already abusing the new rules, before the rules to allow fast lanes are even passed:
Non-USA internet users should be helping protest this in the USA. If its allowed to happen there, it'll set a precedent for the rest of the world.
"Say no to double-dipping!"
Fortunately, many countries in the EU (Germany, Ireland, France, not so much the UK) are happy to give you permission to build your startup here and fall under EU net neutrality legislation.
This entire problem comes down to the fact that internet isn't priced the way other utilities are.
You don't pay your water company a fixed rate based on the size of the pipe to your house. You pay based on water actually delivered. You don't pay your electricity company based on the maximum number of amps that can flow down the wire, you pay for the electricity consumed.
Internet should be paid in bytes delivered. Then all the incentives will be aligned for all parties.
High usage end users pay for more of the network than low usage end users, which they should because they are the ones putting strain on it.
The ISPs would then prefer to transport the larger content that users demand because they will be paid more for it. Thus they are also incentivised to create more and better infrastructure to deliver it.
Internet should be paid in bytes delivered. Then all the incentives will be aligned for all parties.
That sounds like the interests will greatly favor the ISP. They can leave prices the same, institute a monthly data cap then gouge users on their overages. Yes, I can definitely see how this would be a huge win for customers...
Unlike water and electricity, the internet isn't a finite resource in the sense that the data will 'run out' if we use it too much. The people generating content are paying on their end and the users consuming content are paying on their end. We don't need to be careful not to overuse the internet, lest we leave future generations without the ability to use it.
I agree but unlike electricity or water where quality of service is regulated (water quality etc) internet service is not.
We would need some sort of regulation to force ISPs to provide a pipe that doesn't suffer random packet loss or connection problems. In addition you would need some way to make ISP upgrade to better pipes or you will be stuck with 56k for ever.
I believe all new homes in the US are now equipped with a minimum if a 200A circuit by law which would equate to your internet speed minimum being regulated.
> I agree but unlike electricity or water where quality of service is regulated (water quality etc) internet service is not.
It is very regulated. Starting a new utility that relies on public property (easements for landlines or spectrum for over the air) requires cutting an epic amount of red tape. In fact, it is such a messing process that only the very politically well connected companies can even consider entering the market.
It's definitely regulated. It just isn't regulated in the way that you want.
Big difference is that most households in a neighborhood use about the same amount of water or the same amount of electricity. I'd feel safe saying that 95% of my neighbors fall within a factor of 2 of my usage.
Water and electricity usage is also very predictable.
But Internet usage can vary a lot between households and between days. I could easily see there being quite a few people consuming a factor of 10 less bytes than me and also there being some number using a factor of 3 or so more than me, and this can also change a lot from one day to the next.
I don't like this idea because I'm a power user and would stand to lose in this transition but I realise that's not a good argument.
Here are some thoughts:
Other utilities usually come with a base fee, ie. even if I were to consume no electricity I'd still pay some amount of money per month. The total cost is (base fee + usage fee). So it's not purely based on consumed resource amount.
If we were to transition to a consumption based fee model to realign incentives, it would stand to reason that the amount of revenues the telcos should get should not change significantly. I expect that the telcos would not be able to resist gouging their customers if consumption based fees were to comeback (competition might drive down prices... eventually).
Water and electricity are quite cheap (in first world terms). There's an incentive not to be grossly wasteful with either, but that's it, really. It'd be nice if internet consumption fees could work in a similar way.
There is a big difference between bandwidth and water. After being built (which I grant can be substantial), the bulk of the costs to the water provider is the volume of water provided. If my neighbor uses twice the water I am, he's costing twice the same and should be charged the same.
But not all bytes are equal. If I am pulling 1 Linux DVD at 3am and my neighbor is pulling down 2 per night at 3am, the costs he is placing on the system are not twice mine. The actual electricity costs between moving his bits and my bits are negligible.
Instead, the cost is in building out the pipe to handle the maximum load each evening when everyone is pulling videos down and/or Skyping to Grandma. And even then not all bytes are equal. I might really want the voice packets of my Skype session to go through right now, but be able to tolerate several second of lag on the video I'm watching from Netflix, while letting the video of my Skype call degrade as needed.
I don't think you can put ISP regulations into one simple sentence. They need a more complicated and thus expensive regulatory regime. And, yes, the more complicated, the more subject it will be to lobbying pressure from people with different views from you. It still strikes me as the least-bad solution.
EDIT: Water and electricity providers do care about peak usage somewhat. Water companies need to keep up pressure, and electric companies need to keep the electrons flowing no matter what. But the water company can tolerate a 5% dip in water pressure, and electric companies have ways of, say, disabling AC units to juggle demand during especially high periods of usage.
I think we need some way of discouraging the single biggest consumers of the shared resource (such as charging them more) during peak usage while still providing incentive for the ISP to increase capacity. Lots of people fear that the ISPs would just use this to increase revenue instead of to properly ration a shared resource. I totally understand that fear because I trust my ISPs about as far I can throw their headquarters. But other providers of shared resources with much more predictable demand patterns still have ways of dealing with demand surges, so I don't see how ISPs could hope to function in the long-term without some kind of peak-demand management.
I am based in the UK and it is exactly how it is charged for on the consumer side, I can choose how fast I want my delivery to be (1mb/s -> 80mb/s) and also the monthly total consumption (10GB -> 20GB -> 40GB -> "unlimited").
I don't see how that's much different than the usual here in the US. You're paying monthly for a connection, variable fee based on speed of that connection, and then you have a data cap. I don't see how that's paying based on consumption.
This method also incentivizes ISPs doing their damndest to get rid of the lowest consuming tier of customers, even if there is no other service available (seeing as how they've managed to monopolize service carrying in most municipalities).
What would the FCC/Comcast do if they didn't have a 900 lb gorilla to go after? What if consumers were using the exact same amount of high bandwidth services, except instead of it being mostly Netflix and Youtube, it were 50 or 100 similar services?
This might be something that is not just beneficial to ISPs. Probably the major content services want it too, because they can strangle upcoming competition at the expense of a few dollars.
So the 900lb gorilla might not be actually running away.
Netflix doesn't want this to go through, their market saturation is not nearly high enough.
76% of american's have some form of extended television service, cable/satellite.
Netflix covers around 10% of US households.
They see them selves as being in competition with these other providers. It is in their best interest to keep the cost of providing Netflix as low as possible until their % of households is significantly higher.
So when you talk about protectionism you probably need to back out a level. Because remember the big business of the companies pushing the "internet fast lane" is also television. They have layers of conflicting interests in their wish to constrain Netflix.
Pro tip for those unhappy with their ISPs: You get what you pay for.
If you pay for a consumer, price-oriented package, expect price-oriented service in return. There is a reason your office doesn't use consumer Internet, and that a 1.4 Mbps T1 costs $400/month, while low-end 6 Mbps DSL costs $30/month -- 4x bandwidth for <10% of the cost, what a deal!
I shouldn't have to say this on Hacker News, but what you pay ISPs for is quality of service and support. If you want that at home, pay for a business-class service. Comcast offers business-class to residences, for example; we've used it many times and the services and support have been excellent.
Is there really a reason that a crummy T1 still costs $400/month? Other than all the existing places that have a 10-year-old T1 where that used to be the fastest option, and the phone company doesn't want to drop prices.
Business class Comcast service does not get higher QoS on its congested links. I've seen Netflix degrade to the lowest bitrate and frequently pause to buffer on a business class connection before Netflix began its payments.
This is classic price discrimination. As television channel packages become less of a moneymaker for cable companies, they need to find new forms of revenue and methods to segment the market. In separating certain content from other, especially internet data which is all the same to them, the ISPs are able to better maximize prices. Truly a tragedy, and all the more reason net neutrality is vital to the future of internet freedom.
Well if you're talking about stuff that's used at physical checkout counters, you need really fast transactions. Nobody wants to sit there for an extra 20-30 seconds holding up the line. Even if customers can tolerate it, businesses will be much less inclined to support such a feature.
Who is that someone? Distributed systems don't need any particular speed of interconnect to work. In fact, it's one of the benefits of distributed system designs.
Yes, that's what I was trying to say. The "someone" is whoever was quoted in the article; that person does not understand your point.
But it remains true that I, as a consumer, want my payments to be processed quickly. So do most other consumers, I imagine. But this has nothing to do with "making sense". I just want it to happen.
If you look at the history of the phone, which this is very similar too, initially the phone companies charged you for everything. Long distance call, oh that will cost you. Anything at all that will cost you.
Then eventually a player comes in that disrupts the model. What free long distance on weekends and evenings. How is that even possible.
Then one day competition takes over, et I mean logic ;) and suddenly the phone prices make sense. Long distance calls are something of the past.
I believe ISP are going through the same thing. However since they can't easily start charging by the call what they are trying to do is charge not only the person who is paying to have a phone line but also the caller. Except that now the caller has to pay for their own phone line and a fee to connect to the end line, which is already being paid for by the end customer. Aka the last mile is double dipping in fees cause they can because there is no competition. Just like the long distance calls from before. Amazing how quickly the phone companies went from recovering costs to can make profits from no fee long distance calling when competition finally came...
This comment reflects a deep misunderstanding of the history of the PSTN and its regulatory and deregulatory cycles. Telephone service is also vastly easier to predict and model than Internet usage.
This is a terrible article. It's just VCs talking their books. I hate the FCC's slow/fast lane proposal but you can't write about it only from the perspective of those with financial interests in the outcome (especially without making clear the importance of those financial interests).
I know ISPs attract little sympathy and HN readers naturally do sympathize with startups and smaller organizations (as do I), but consider:
In open markets a seller charges what the market will bear in order to maximize profit. And of course it benefits society because resources are allocated efficiently, i.e. to those with the greatest demand for them (those willing to pay the most). Why should this market be any different? I'd expect that any business, discovering they had an asset someone would pay for, would charge as much as they could for it. There is nothing dirty or underhanded about it, even if it challenges the status quo we all are accustomed to (or the notion that everything on the Internet is free). And it's not like big ISPs are stomping widows and orphan children -- Netflix, for all its victimhood, can stand up for itself and I'm sure maximizes its profit wherever possible. (It may be different for startups.)
I don't want to be absolutist about it, but what I said is a factor. Other factors are that the ISPs benefit from public goods such as local monopolies, and benefit from all the innovation that created the Internet and the services that their customers are paying them for, so they do owe something to maintaining that ecosystem.
And the clear solution, as has been said before here, is customers paying for metered bandwidth, and also for latency or other SLAs if they want them (which long has been common for businesses ISPs) -- why should someone who mostly uses email want to pay for the same latency that a Netflix customer or gamer needs?
Not only do they benefit from public goods, but it's a market where the barrier to entry is so high that competition cannot happen efficiently. Now before our american libertarian friends come frothing at the mouth, the barriers to entry are absurdly high before any sort of government interference, especially in North America, and with a captive customer base, it's just inherently a bad market to showcase capitalism.
So with inefficiencies due to a lack of competition comparable to a government agency AND a profit motive rather than the public service motive that government agencies have, you've got the worst of both worlds. Companies that barely compete to offer good service with the explicit intention of sucking as much of your dollars as they can for a service you can't really go without.
So, if ISPs are free to charge extra for bandwidth used by premium content providers, are they also free to charge extra based on other factors?
In other words, could the Koch Brothers start an ISP that gave a break to conservative information providers and charged extra to competing political views?
seems like someone really wants to destroy the US economy further, with all the NSA spying fiasco, and now this. seems like the US government has an unlimited amount of bullets to fire on its own foot.
seems that some networking startup will solve this as its peer problem with too many video downloads but not enough upload traffic
that is the problem with net fairness via speeds,etc is that underlying network ecosystem is not static it will change and with that change comes new problems in being fair with speeds. etc
I know this argument doesn't make sense at first blush, but it does at second:
1. People don't like metered Internet (which is odd given we don't mind metered water and metered electricity).
2. ISPs want to keep advertising unmetered Internet.
3. Some % of the customers use a lot more traffic, and most of that is Netflix.
4. The ISP want to charge Netflix, which will in turn pass the costs to their customers, and so these customers who use more traffic will indirectly pay more to the ISPs for their traffic.
I wish they'd just lay it out simple like that. At least then we won't have to nitpick their statements for logical inconsistencies.
Or... maybe they can just switch to metered Internet. And this whole net neutrality issue will vanish in a puff of smoke.
The reason I would disagree with metered Internet is because based on the past behavior of the major ISPs it would result in even worse pricing schemes. I simply do not trust them.
To me the only way metered would work is if all these government sanctioned monopolies in communities are lifted so that any company can offer service to any customer.
Or, I suppose, the service is a utility governed by representatives of the community.
Harsh reality time: 2 choices. Either Netflix ponies up, or everyone has to pay per GB to their ISPs. I honestly don't think the average consumer wants to be constantly tracking how much BW they're using per website. That might work for a single person cell phone, but doesn't work for a family.
That's fine. Since the problem isn't last-mile, but interconnection, the per-GB cost is miniscule. I'll estimate $0.025/GB, but that's way too high, just to avoid getting to the point where one might realistically quibble over the real cost.
I could stream 10mbps video 24 hours/day for 30 days and my bill at $0.025/GB will be about $80 for bandwidth, on top of the ~$40 I pay for my existing connection, which we'll just assume we'll let the cable companies continue to gouge me on.
If I'm streaming high-quality HD video 24/7, I can live with a $120 Internet bill.
I'm not actually doing that, of course. In reality, even I (combined with 4 other family members on the same connection who aren't power users, but do watch some streaming video and otherwise make fairly active use of my connection) rarely use more than ~500GB/month. Even if I used a terabyte every month, it'd be about $25 extra. Meh.
In the end, it's not consumers that don't want this. It's the cable companies. They hate the idea, because it puts front and center the fact that they are and should be just a dumb pipe. They don't provide content, they provide a conduit. This is exactly the public perception that they fear.
> That's fine. Since the problem isn't last-mile, but interconnection, the per-GB cost is miniscule. I'll estimate $0.025/GB, but that's way too high, just to avoid getting to the point where one might realistically quibble over the real cost.
That's half the per-GB cost that Amazon and Microsoft charge for S3 and Azure transfers to the internet for those customers using the highest volume below the "call for pricing" level, and so get the lowest published rate. I don't offhand see any reason to expect Amazon and S3 to be paying more for bandwidth than last-mile providers do, nor for them to be marking it up to their customers more. Are you sure $0.025/GB is too high?
Very, very sure. I can pay far, far less than AWS and Microsoft charge for bandwidth in any US datacenter managing my own equipment.
Meanwhile, HN-darling Digital Ocean charges their VPS customers just $0.02/GB on overage. That's right, for overage above the 1-9TB they include with their plans, DO charges less than I'm estimating the cost to be for massive providers to exchange traffic at far larger and more cost-effective scales.
Linode charges $0.10/GB, and that's still less than AWS charges for the first 10TB. And again, that's after the 3-20(!)TB Linode's plans include in the base cost.
Bandwidth is cheap. The US has all the fiber capacity between POPs that it needs. Adding or upgrading interconnects between Comcast and Cogent/Level3/whoever is not laying fiber across the countryside, it's buying switches or turning on ports and running cables from one cage in a datacenter to another cage in the same friggin' datacenter.
We have had limited GB in our country with a family for years. It's works, it's as simple as occasionally checking how-much you have left, or just going over the limit (that happens to). You learn how much you can download, and to not waste it. It's a bit like the water bill, you can leave it open all day, but you shouldn't.
Apart from countries that do this, such as Australia, your ISP shows how much you have used. Atleast in Australia now you can get quite meaty plans (1TB+ a month) for "cheap".
This argument doesn't make sense to me. The ISP's customers are using the ISP's network by downloading Internet content. The nature of the content should be irrelevant to the discussion. The ISP's customers are purely responsible for that usage, and it is part of - from what I understand - the service they are already paying for. The ISP's network is only "congested" if it is incapable of delivering on the service that was advertised (assuming for the sake of argument, some particular speed of access).
Thus, how can Cashman argue that "there is no model for monetizing the required investment?" Isn't that exactly what ISP customers are buying when they pay for access to the Internet through the ISP? The onus should morally be on the ISP to make a good-faith attempt to build sufficient connections to any content that's popular enough to saturate the links providing it.
Furthermore, why isn't it naturally the ISP's responsibility to price their access sufficient to cover current usage and expected growth in usage? New use-cases for the Internet are constantly going to appear. It is the ISP's responsibility to provide Internet access with sufficient throughput that it is not congested for any reasons under the ISP's control; and to price their access sufficient for investment to achieve the throughput.
"The latest BigMac hamburger caused hungry residents to swarm McDonalds and caused drive-through lines to be congested today in Cloudville. Gillis Cashman, managing partner at McPartners, Inc. says that, despite selling happy meals in record numbers, delays in drive-through lines will continue, since since there is "currently no model for McDonalds to monetize the required investment" to upgrade its drive-throughs.