isen.blog

I think transfer caps are a great development for broadband competitiveness :-) The outrage (and news stories) that will ensue when people get huge bills for innocently downloading high def movies to their AppleTV or TiVo or other set-top box WILL attract Broadband Internet Access competition.

The other thing thing that will be absolutely hamstrung is the new online backup services like Carbonite that allow you to back up your hard disks to Internet-based disk farms... backing up a 250 GB... or 500 GB... or (shudder...) a 1 TB disk will end up exceeding ANY transfer cap.

Heck... imagine how much transfer will end up being used in soon-to-be home High-Def videoconferencing that's possible now that HDTV cameras and displays are pretty cheap! Cisco + Linksys + Pure Digital...

That kind of populist outrage could even give rise to more publicly-funded fiber networks like Lafayette, LA. I'm kind of stunned that I'm cheering on Verizon Communications, of all companies, as they SEEM to be doing the right thing with FIOS in 1) building FTTH networks with AMPLE capacity, and 2) not speciously implementing transfer caps. Yet.
 

The fact is that metering and caps are necessary to keep prices down to what users want to pay. Here's why.

Internet service providers buy backbone bandwidth according to the continuous capacity of the link (usually measured in megabits per second). That bandwidth can be very expensive (in our case, $100 per Mbps per month; in some ISPs' cases, $325 per Mbps per month or more). In certain big cities, you can sometimes get bandwidth at $3 per Mbps per month, but only if you purchase huge amounts of bandwidth and pick it up at a major hub. (In other words, the $3 per month doesn't include the delivery system.)

End users, on the other hand, want a palatable flat rate price for service. Typically, it's $30 to 45 per month (though what they really want is $15-20). And even at the lowest service level, they expect Web pages to arrive quickly. They want VoIP to work. And they want interactive games to be responsive.

And so, what ISPs must do to provide them with reasonable monthly rates is reminiscent of the old engineer's joke:

Q: How do you carry two tons of canaries in a one ton truck?
A: You keep beating on the sides as you drive along, so that at least half of the canaries are airborne at any given moment.

In other words, to satisfy consumers, an ISP needs to oversell bandwidth and rely on the fact that not all users are consuming network resources at the same time. The ISP can let users burst to very fast speeds, but you can't have all of them doing things which continuously take up megabits of bandwidth for extended periods of time.

ISPs can also leverage the fact that most users have a limited capacity to produce new content. They may be prolific writers, but this isn't even a blip on the radar because text takes so little bandwidth. They may be avid photographers, but even good still photographs take lots of time to compose, crop, and post. They may even create a lot of video, but since video is generally uploaded to servers close to the backbone (e.g. YouTube) for general distribution, this also does not create much continuous upstream traffic. For this reason, ISPs can buy more downstream than upstream bandwidth, and design their systems so that they have more downstream than upstream capacity.

Now, enter two classes of applications that don't lend themselves to oversale. One is P2P. Leaving aside for a moment the fact that P2P is overwhelmingly used for illegal activity, the two big problems with P2P are that it often has a 100% duty cycle and consumes precious upstream bandwidth. This wreaks havoc with the careful engineering that is done to reduce the cost of users' service.

The second application that causes problems is video streaming. An HD stream can easily take up several megabits per second... and users, who are used to leaving the TV on, don't realize just how much it would cost the provider if they left the stream on in the same fashion. So, they often stream content for hours -- EVEN WHEN THEY ARE NOT WATCHING IT. (The same is true of music; many users leave Sirius/XM or Pandora running all day even if they are out of the room.)

The problems are further compounded by the fact that the Internet is not a broadcast medium. When another user tunes into an over-the-air television station or even a station on a cable TV network, no new resources are consumed. But on the Internet, each video is an individual stream which consumes more resources and creates additional costs.

For all of these reasons, ISPs MUST either discourage continuous use of bandwidth-hogging applications (analogous to beating on the side of the truck in the joke above) or raise their prices. This creates a serious dilemma. They do not want to insist that customers pay the full price of backbone bandwidth plus a markup, because the higher prices would anger consumers and cause many of them -- especially those not financially well situated -- to drop their service. It would also discourage potential new customers from signing up.

ISPs likewise do not want to charge by the bit, because so many applications that download (e.g. virus checkers, which download updates frequently) are beyond the average user's control.

So, they have been forced to adopt a hybrid approach. They start with a low flat rate that applies when the user obeys the assumptions that allow cost reduction. They then add surcharges which -- if they're designed correctly -- will approximate the direct sale of backbone bandwidth on a "cost plus" basis to heavy users who push their connections to the limit.

Of course, all such schemes are approximate. The crudest approximation -- caps plus overage charges -- tends to discriminate against users who do lots of long, low bandwidth downloads, and also tends to cause nasty surprises (as do overage charges on cell phones). But as one makes the formula more sophisticated (and more accurate), it becomes less and less comprehensible to the customer.

The result: more and more naive consumers are already claiming that ISPs are discriminating against online video because it competes with their own offerings (at least when the ISP is a cable company) or leveraging market power. And as mentioned above, one member of the US House of Representatives, Rep. Eric Massa of New York, has gone as far as to say that any sort of metering should be illegal. "I firmly oppose capping Internet usage," he said in a press release posted at

»massa.house.gov/?sectionid=24&se···emid=205

"and I will be taking a leadership role in stopping this outrageous, job-killing initiative." The Congressman also claimed that metering and capping Internet service harms free speech, which is absurd on its face because virtually no user can "speak" enough to have his or her speech impeded by the caps.

I called and e-mailed the Congressman's office and asked, "Would you also be opposed to metering of electricity? Of natural gas?"

As of this morning I haven't gotten a response.
 

Brett is on the right track here, Internet is a utility and we should look at utility models which are usually base+metered.

I think where we get the breakdown is in the pricing as the Ars article mentions. We don't count Watts in our everyday life, but we are mindful (especially today) not to be wasteful and that there is a real cost. However, I also only pay pennies for killo-Watts, not some dollars for Watts which is what TWC and Charter want to charge.

We need a) REAL competition that drives price down or b) effective monopoly regulation that contains this to a reasonable cost-plus equation that benefits society more than executive pockets.

PS: The subsidy model is broke too, so give up the "services" pandering. We don't buy stoves from the gas company, soap from the hydro or bulbs from the electric utility, so why do I buy minutes or channels from my "ISP" with the belief that bits and pipes are free.
 

The Internet isn't a utility. You generally have one power company, one phone company, one cable company, one gas company in a given area. But you can have dozens of ISPs, all with very different offerings. Gas is gas, electricity is electricity. But Internet services can be as different as night and day.

And while people don't usually buy light bulbs from the electric company, they often DO buy appliances from the gas company. (Our local gas company does a brisk business in water heaters, ranges, etc.) And our ISP's customers want to buy computers from us and have them service them. (Unfortunately, we aren't set up to do that, but many small ISPs are.)

In any event, regulating ISPs like utilities would kill many of them, destroying the competition you advocate. And the rest would be commoditized; there would be no variations in the service, no innovation. You do NOT want to go there.