Hacker News new | past | comments | ask | show | jobs | submit login
Netflix: ISPs who charge by the gigabyte are ridiculous (arstechnica.com)
96 points by woodrow on Jan 27, 2011 | hide | past | favorite | 50 comments



This is not accurate. Any time you have multiple people sharing a physical link you end up with a marginal cost for bandwidth in any sensible economic arrangement. If you insist on having every part of your connection dedicated to yourself, you're effectively back to a circuit-switched network with the corresponding orders-of-magnitude price hike in providing it.

For example, in the UK the last mile is usually supplied by the incumbent (BT), who also (virtually) route that traffic over a private IP-based backbone to ISPs to provide the backhaul. They charge ISPs by capacity. So these ISPs effectively have marginal costs that vary throughout the day according to demand, since they have to pay directly for an appropriate (fixed) amount of capacity to be available.

We also have LLU (local loop unbundled) ISPs who still use BT's last mile but provide their own equipment directly at the exchanges, thus effectively providing backhaul straight from the exchanges. They may only have a fixed cost for bandwidth depending on how they provision those links. But if they are buying capacity from anyone else, they will be charged by how much they are buying, and suddenly they have a marginal cost again.

The article says that Netflix are complaining about the situation where Netflix have brought the traffic to the last mile. In this specific situation it makes sense. But in the general case, ISPs still have marginal costs in providing bandwidth.


The article is not about bandwidth but about the amount of data you can download at a reasonable price. People are actually willing to spend more time downloading something instead spending more money.


The amount of data, when aggregated over many people, adds up to bandwidth of network needed.

An ISP's customers, all added together, are downloading a certain amount per month. That is, a certain amount of data per time period -- which is bandwidth.


It certainly seems like a little bit of consumer education should go a long way, but without competition – or in the presence of collusion – we'll end up with something looking more like text messaging.

Everybody knows the cost to the carrier is only a fraction of what they charge, but there's no competitor willing to charge less. Education, unfortunately, seems to only go so far.


Would you also educate consumers about the true cost of web apps?


On Verizon's family plan, we pay 30 extra per month for unlimited texts for 5 people. 6$ per month for that seems like a fair price.



Let me clarify. At 6$ a month, I don't feel like I'm getting ripped off, even if it really doesn't cost them nearly that much to provide me the service. The t-shirt I'm wearing costs maybe $1 in raw materials and manufacturing, but it sold for $20. I got it 50% off for 10$. Am I getting ripped off? Should $10 for a t-shirt not seem like a fair price?


Indeed. People today expect to get paid for their creativity and the value they provide to customers or businesses, but they still expect to pay based on raw material cost. You can't have it both ways. It's silly to think you should get paid based on value but pay others based on cost.


You might if you found out that people in other Western countries with similar costs-of-living could buy the same shirts for $1.50.


> "Let me clarify. At 6$ a month, I don't feel like I'm getting ripped off"

But surely you can see where people who don't have full family plans have a completely different view of the issue?


> At 6$ a month, I don't feel like I'm getting ripped off

OP's point was about consumer education. Have you considered that maybe you should feel like you're getting ripped off?


I have. I'm not dogmatic about this.

Part of my other point was that the comparisons in the link to why SMS is ridiculously overpriced, in which the cost of downloading songs via SMS is calculated to be in the millions of dollars, is that that's a bogus argument. If you're sending that many SMS's, get an unlimited plan. If your carrier doesn't have an unlimited plan, switch to one that does. Consumer education is fine, but there's also something to be said about consumer responsibility.


That article is comparing apples to oranges. I perfectly agree that SMS prices are a ripoff, but to evaluate this the author should be looking at what it costs AT&T to send a text message - not at what Comcast charges you for wired broadband! After all, I bet that SMS is a wonderful deal compared to what it costs NASA to send .13 Kb to the ISS ...


That's some pretty bad math. That analysis uses the average characters/message for SMS, but uses the maximum characters/letter for postal delivery. The average is probably less than a page.

It also double-counts SMS costs (for sending and for receiving), but doesn't double-count the ISP bandwidth costs.

Still, I agree with its conclusion: there's no good reason for texts to cost so much.


They're clearly limiting this discussion to wired ISPs, where I think this is reasonable.

Wireless (cellular) users can look to satellite ISPs to see just how crazy things can be. On a lot of satellite systems, users are sold "unlimited" bandwidth, but then aggressive fair access policy, unpublished caps, etc. apply, so your unlimited 2Mbps x 256Kbps connection might actually only let you transfer 500 MB/month.

True hilarity happens on the L-band services like Inmarsat BGAN when either someone leaves Windows Update or other autoupdate turned on, or isn't aware of the per-MB charging. $3 to $25 per megabyte adds up quickly; I've personally seen a $160k/mo bill due to users downloading movies and music over a BGAN system (with the bill paid by a third party, obviously, and no mechanism in place to communicate costs to the users).


Now I wonder why Netflix might have a problem with that...

In any case, they're absolutely right. These plans are clear cash grabs, and they wouldn't exist if we had proper competition.


In Canada many people won't get Netflix[1] because of ridiculously low bandwidth caps. It's amazing how much lack of competition and corrupt regulatory bodies can stifle innovation.

[1] And of course, Netflix' catalog in Canada is very limited due to similar politics.


Indeed. I moved in with by g/f and soon after that, her Rogers Internet service went up in price and monthly cap got slashed to 60 GB/month. I subscribed to Netflix service for the 30 day free trail, but promptly canceled after I saw the horrid selection and b/w usage. Needless to say, we are switching to Teksavvy since I've been a customer of theirs since 2001


I am on Teksavvy and a Netflix subscriber. I found Netflix catalog to be quite good, it may not have any blockbusters from the past few decades but it has a very good selection of indie, foreign and old movies. Netflix Canada also has good selection of tv series e.g. Mad Men, Topgear etc. Its definitely worth it for monthly price which is half the price of a movie ticket.


I use Teksavvy as well. Slower internet, but no caps (for the time being).


Spoke too soon. With UBB, they are capping it at 25 GB starting in March. :(


In the west, we've got SHAW who recently lowered bandwidth caps without changing the price and are charging for any overages, and Telus who are not yet charging for overages but they try to lock you in on contracts and also have caps. For a 15/1 pipe you pay about $60/month.

The problem is since all of the major ISPs also provide television they are effectively attempting to get rid of netflix as competition by placing such restrictions on our access.

http://openmedia.ca/ has been campaigning to get rid of the caps in Canada. Quite a fair bit of interesting reading and it's nice to see it gaining momentum lately.

EDIT for the typo.


Say what you like about the Maritimes, but one thing is for certain: I'm so glad that the main ISP out here is Eastlink. 15Mbps is the stock standard, and for a bit extra you can double that (or, for a lot extra, you can go to 100Mbps). Admittedly there is a 250G advertised cap on the faster plans, but nothing advertised on the 15Mbps plan. Plus, we definitely don't have to deal with weird torrent filtering, shaping, interceptions or anything.

And it definitely is good enough for Netflix.

Just wanted to throw that out there... you never know, they may eventually creep westward and bring some real competition :)


oligopolies are a bitch


agree.... I hope the wireless careers can give some stiff competition to the wired incumbents like comcast, road runner and their ilk.... that did be nice!!!!!!!


You should actually be rooting for telco competition. Look at areas where Verizon FIOS and Comcast compete head-to-head. Not only is the broadband excellent quality, it's price competitive with almost any other European/East Asian nation.

There's no doubt in my mind that the U.S. can have excellent broadband at a reasonable price. Unfortunately, in most areas we only have the choice between pricy, midgrade cable internet and cheap, slow DSL.


I'm rooting for all competition; I don't have any reason to believe that telcos would be any less corrupt than cable companies. But if we had a choice between multiple wireless providers, multiple cable providers, and multiple telcos, we might stand a chance at a fair deal.

The Cable/Telco-FIOS spar is likely to evolve into a wink-and-nod conspiracy where companies all settle on similar high prices and compete only in the marketing arena (see SMS as one easy example).


This has been the status quo in South Africa from 1997. In 2009 I lost a legal battle via the Advertising Complaints Commission.

It boiled down to the wording. Uncapped means that they can shape your line to s*, and still sell it as a 4meg ADSL service.


Australia has the same thing. 'Unlimited' = full speed up to X gigs each month, after that you get dial-up speed only until the start of the next month.

Most plans are advertised by GB/month though.


Sad, really.


It's fun to look around at the limits in terms of how long it'd take you to do it with a 56k modem.

250GB, which they consider pretty high, would actually take a while:

http://www.wolframalpha.com/input/?i=250GB+%2F+56Kbps

But smaller values are less impressive. 10GB is only half a modem month.

Looking at the $1/GB thing gets more interesting. How much would it cost me to run a 54Kbps modem 24/7 for a month at $1/GB:

http://www.wolframalpha.com/input/?i=%2856Kbps+*+1+month%29+...


You're forgetting, or maybe just never used a modem (some people ask what the hell a floppy is), but the max speed you could squeeze out of a 56k was like 3/4kbps for any single download. That's if you were lucky. You would never, ever have transfered so much as 10gb in a month.

Things sure have changed. I remember praying to the modem gods that the connection wouldn't simply drop when downloading something as small as 200kb. The good ol' days -- you made every moment on the net count.

http://www.wolframalpha.com/input/?i=10GB+/+3Kbps 10 months.

Edit: What I just linked to above was incorrect. I should have used kilobytes per second, not kilobits per second.


the 56k stands for 56KBit/s. What you were seeing was 3.5 KByte/s which of course is still way off the theoretical 7 KByte/s you could achieve with 56KBit/s.

If you only got 3 KBit/s over a 56K modem, your download-rate would have been 375 bytes/s - a third of a Kilobyte per second. That's a bit more than what you could have reached with a 2400 baud modem from the eighties.

You get to your 10 months result by dividing the 10 Gigs by 3Kbit/s, while you would probably have ment to divide by Kbyte/s

http://www.wolframalpha.com/input/?i=10GB+%2F+3KBps

38 days.


Thanks for the clarification on my points. I was conflating the two.


I never once saw more than 48kb/s on any 56kbs modem I ever used (as i recall there were two slightly different implementations at the 56kb/s level, but neither one ever achieved that speed in the real world).


Are you sure you're not referring to this?

http://www.wolframalpha.com/input/?i=56+Kbps+in+KB%2Fs

With 8 bits per byte plus protocol overhead, a good rule of thumb is take the Kbps value and divide by 10 to get a realistic KB/s figure. It's been too long since the BBS/dialup days and I don't recall if some of the later modems implemented on-the-fly compression which would squeeze out even more bandwidth.


I hate the idea of paying per GB.

However, the foundation of this argument is baseless. All that discussion of how much it costs to provide the bandwidth really is irrelevant.

I've said it before and I guess I have to keep repeating: the price of a product is unrelated to its cost!

The price of a product is the point at which a producer is willing to make another unit of it, and at which a consumer is willing to buy that unit. Costs don't come into the discussion at all. Everything based on the perceived value of the product from each party's perspective.


You are ignoring competition. If the cost to produce X is 1$ and someone is charging 10$ then there is an opportunity for someone else to step up and charge 5$.

In an open and free market the cost generally works out to around 1.1x the cost of producing the quantity of product that customers want to buy. Which is why most large companies hate free markets and do everything they can to create monopolies either natural or legislative. Cable companies would love to charge 1$ for something it costs them 1c to make and they are trying everything in their power to make that happen.


And that's ultimately the discussion we should be having:

Pay-as-you-go, Unlimited, Monthly caps -- all are potentially reasonable ways to charge for network access. Provided there's competition.

And without competition, there's really no advantage to any of them. You're just picking a poison at that point. The monopolist(s) will get their profit out of you.


I'm from Australia where fixed quotas on wired (typically DSL) are de rigeur [1].

Several years ago the ACCC (Australian Competition and Consumer Commission) banned the use of the word unlimited on Internet plans unless they were truly unlimited meaning no overage fees and no traffic shaping after hitting a soft cap.

Compare that to the UK where they've de died that unlimited means what 90% of users use, which is as little as 5GB/month!

In the US you typically still have caps (hidden behind terms like "fair use"). So the only difference in Australia is the ISPs are upfront about it.

So in Australia the net effect has been:

1. You buy the plan most suited to your usage. In other words you pay for what you use.

2. You get what you pay for. If you pay for 1TB/month you get 1TB/month.

Both of these situations are much better than say Comcasf or Time Warner really only giving you 250GB/month or UK providers declaring you're a downloaded and moving you to a network that in the evenings is utterly unusable.

The fact is that bandwidth still has a marginal cost so it's not unreasonable to expect users to pay for that. Nor is tiered pricing based on usage a bad thing.

In New York I have one choice for high speed Internet: Time Warner (some areas have Verizon FiOS; not mine). There are two plans for $65 and $99 (iirc).

In Australia I have typically a dozen choices or more ranging from $20 to ¢150+ pe month based on the features I want. How can that possibly be a bad thing?

[1]: http://www.iinet.net.au/broadband/plans.html


Australia has a terrible broadband network.

I live in the US and pay 45$ a month for a 35mbps up/5mbps down on a fiber connection with no cap. The actual cost of me downloading 1TB works out to around 10$, but like most customers I don't break 200GB/month which is a small fraction of my bill, and the company is wise enough to realize overcharging for bandwidth is just going to piss me off.


Yes and how much of the US by population has the option you have?


I don't know the percentage, but more people than live in Australia. I game with people in Australia and even in major city's they don't have reasonably priced high speed internet. Granted, laying undersea cables is expensive but compared to other countries in the reign Australian internet sucks.

PS: I am not saying that people living in the middle of the country need to have high speed internet just that you could wire up the major city's at reasonable cost because that's where most people live. http://www.environment.gov.au/soe/2001/publications/theme-re...


Now they know how I feel about text messaging rates. Captive consumption sucks.


The real problem isn't even that they're charging by the gigabyte (although they are, and I strongly dislike it).

No, the REAL problem is that they preferentially exclude their own services from these charges/caps.

For example: If you get Bell's IPTV offering (in Canada), do you really think they're going to cap your consumption? What about Rogers on Demand? Is that subject to the same caps?

The answer is no, making their services more attractive than competitors like Netflix.

Ergo, the only solution is separation of concerns. Providers of pipe should not be providers of content. Ever.


I know, I know I say this every time, but man, I would love someone in New Zealand to come and tell my broadband provider that their top-of-the-line plan that I pay $80 NZD a month for that only gives me 40 gigs is BS.

Oh and we've already prioritised traffic too..If i was to buy tivo, then all my on demand data would be "free". I don't think Net Neutrality is even a remote consideration here - though there are many of us who passionately believe in it.


Welcome to Australia...


At the risk of sounding trollish, this is one of the reasons I absolutely love living in Slovenia.

30 euro a month for unlimited and unshaped 20/20 FTTH. :)

From what I understand it's like this in a lot of Europe actually, especially those scandinavian countries.


If charging by the gigabyte is ridiculous, what is AT&T data roaming on Rogers for $15/MB?




Consider applying for YC's W25 batch! Applications are open till Nov 12.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: