Comments on: Open Letter to Senator John F. Kerry — Price Tiers Benefit Consumers

By: PB

PB — Wed, 27 May 2009 09:40:35 +0000

you know what would benifit consumers more why not put a tax on the cable companies who implment caps say 5 dollors per customer this would be good for end users and id rather my money go to congress or al quida then the greedy demon known as time warner cable

By: Mark McKenna

Mark McKenna — Fri, 24 Apr 2009 15:10:44 +0000

Some good points. I would not be surprised if a form of tiered service would actually stimulate widespread availability of low-level internet service provision. Though Brian Swanson makes a good point with the capacity vs. data consumption argument, it remains that, provided bandwidth windows are staggered appropriately, the tiered pricing schedule will have the desired effect of bringing bandwidth consumption more in line with fees accrued.

However, it is undeniable that the bandwidth caps are absurd in scale. Even moderate bandwidth users like myself would exceed the ‘power user’ rating in the first week of service. The measurement of overall consumption in megabits is very misleading. 100 gigabits of transferred information over the network is roughly equivalent to 10 gigabytes of actual content, possibly substantially lower depending on how packet overhead is accounted for. Internet radio at a decent quality will generally stream at a content rate of about 100 kilobits/s (roughly 10 kilobytes/s); at this rate listening for 8 hours a day (a not-unreasonable expectation for the near future of internet usage, considering that many work all day with the radio on) would consume the entire allocation. Streaming video bandwidth consumption would begin at around 5x this number, and this does not account for any additional usage (such as email, web browsing, and miscellaneous file transfers). Additionally, peer-to-peer technologies (such as Skype, Grooveshark, BitTorrent, and several online television streaming sources), in which an end user’s computer may be fetching, processing and uploading information which is not directly related to that user’s personal use, could increase the amount of bandwidth consumed by a typical user hugely, likely without their knowledge.

In fact, the knowledge problem is even more serious than that; any user in the 1 gigabit tier could, at any time, through the discovery of any single online service, easily exceed their limit within a single day. As the internet is growing in breadth and technology constantly, anyone could catapult from a low-usage pattern to a high-usage pattern at any time. In this way the internet differs radically from other potentially consumption-based services such as telephone or cable television.

A parallel situation might be if a cable carrier were to mix high-definition and low-definition content into their cable service, and charge users by their bandwidth consumption. This could result in an unsavvy consumer purchasing a low-tier service level receiving a bandwidth limit warning on May 3rd because they started watching a new nightly television show–and in order to keep their bandwidth usage within their limit they would need to forego watching television for the rest of the month (perhaps not a bad idea from time to time, but problematic nevertheless)?

A much more reasonable tiered pricing model would scale the data allocation exponentially while still increasing cost linearly, and where the customer is simply charged for the next tier if they exceed their data allocation:

allocation price
1 gigabit (Gb) $15
10 Gb $25
100 Gb $35
1 terabit (Tb) $45
10 Tb $55

…and so forth. This is more reasonable in the context of the technology in question, because having links in the first place costs significantly, but having larger links (larger fiber bundles, or multiple bundles following the same path) is only slightly more expensive. It is also more reasonable in the context of information technology changes, which cause the average user’s bandwidth consumption to rise exponentially (through moving from text to images, images to audio, audio to video) more often than it rises linearly (through the actual amount of time spent in front of the computer).

By: gabe

gabe — Tue, 21 Apr 2009 18:54:25 +0000

By: Brion Swanson

Brion Swanson — Sun, 19 Apr 2009 18:22:59 +0000

Mr. Darby,

I appreciate your concerns and agree with several of your points in your letter, however I am concerned that you may be taking Time Warner Cable’s word from their press releases as solid fact including their own confusion about bandwidth capacity versus Internet data consumption.

Please allow me to offer a simple comparison to help illustrate the difference between capacity and usage.

Webster’s dictionary gives one definition of capacity as: “the maximum amount or number that can be contained or accommodated .”[1] Using this definition as it relates to the Internet and data, we say that the capacity of a network connection is equal to it’s bandwidth, for example 10 megabits per second (Mbps). Likewise Webster defines bandwidth as: “the capacity for data transfer of an electronic communications system.”[2] Please note, I take exception to their example as it is contrary to the definition.

Other things we are familiar with that have capacities are rooms (maximum capacity), highways (maximum number of cars at a time), and pipes (maximum amount of fluid that may pass per second). All of these capacities refer to simultaneous usage, not total usage.

The capacity of a room is not diminished when one person enters and subsequently leaves. It is only diminished as long as they remain in the room. Likewise a highway’s capacity for traffic is not reduced as cars enter AND leave, rather only when cars enter and remain.

So-called bandwidth caps as proposed by Time Warner Cable, AT&T, Comcast and others are actually not related to bandwidth at all. They do not limit the speed of your network connection - the maximum simultaneous data transfer - they limit the total amount of data.

That limit is like specifying that a room has a maximum capacity of 1,000 people and after 1,000 people have entered (and left) the room no more may do so. Or in the case of Time Warner, any additional people to enter the room will pay an additional fee.

This extra fee does not do anything to alleviate the problem of having too many people in the room at once. It only discourages people from going into the room in the first place for fear of being the one that goes over the 1,000th person limit.

I hope this illustrates for you how “bandwidth” caps as proposed are not a solution to a capacity problem.

I agree that perhaps an emergency tier should be created for highly time-critical information such as medical records to be delivered, but I strongly disagree with the notion that the access you pay for is somehow subsidizing another’s usage.

I also request you publish the raw data used in the studies to which you refer so they may be reviewed publicly as a scientific paper would be reviewed by others in the field before it is accepted.

Thank you,
Brion Swanson

[1] http://www.merriam-webster.com/dictionary/capacity%5B1%5D
[2] http://www.merriam-webster.com/dictionary/bandwidth

By: Brett Schulte

Brett Schulte — Sun, 19 Apr 2009 17:59:54 +0000

The problem with the proposed Time Warner bandwidth caps is the timing and the size of the caps strongly suggests that the real motivating factor is to slow the growth of disruptive technologies like streaming video services and VoIP. Sanford C. Bernstein, a research firm, puts conservative estimates of potential costs for users who view streaming online video for roughly 7.25 hours a week at up to $200 in overage charges a month under Time Warner’s proposed service. Bernstein’s estimates, mind you, are based on streaming video services alone and don’t include other media transfers, such as purchases from iTunes.

The solution is probably going to be legislation that will further open cable networks to competitive providers, and in doing so create an opportunity for other business models.