Slashdot Mirror
Time Warner Cable to Test Tiered Bandwidth Caps
I Don't Believe in Imaginary Property writes "According to a leaked internal memo, Time Warner Cable is testing out tiered bandwidth caps in their Beaumont, TX division as a way to fairly balance the needs of heavy users against the limited amount of shared bandwidth cable can provide. The plan is to offer various service tiers with bandwidth fees for overuse, as well as a bandwidth meter customers can use to help them stay within their allotment. If it works out, they will consider a nation-wide rollout. Interestingly, the memo also claims that 5% of subscribers use over 50% of the total network bandwidth."
Where is the news in this? Canadian ISPs have had caps and over usage charges for years. I can tell on any day exactly how much bandwidth I've used and how close to my cap I am.
I don't see a problem with this, having usage tiers with costs depending how much you plan to use is fine. The problem in the past has always been claims of "unlimited" until you reach a magical, secret cap. I don't think users will have a problem with tiers as long as you make the exact numbers completely clear, and of course that you charge reasonable rates.
US ISPs have charged different rates for different speeds for a long time, how is this any different? It brings clarity to users.
I, for example pay $35/month and am told I get 2.5Mbps down, 760kbps up, and 30GB total transfer. And if I want to transfer more, I pay more. It seems reasonable to me.
There are several places where ISPs are apart from reality.
First, there's abuse of the term "bandwidth", which has nothing to do with the amount of data downloaded. Bandwidth is how much of a frequency range is being used on the wire to provide the service. That's it.
"Date rate" is how much data the bandwidth, encoding, compression, and such allow you to get out of the bandwidth. It's also what ISPs limit you to when they say "megabits per second" or "kilobits per second".
Total monthly data transfer available for an always-on connection at a certain data rate can be calculated as the data rate per second times the number of seconds per month. Capped usage for total monthly data transfer, which is what this article is actually about, can be thought of as the data rate times the number of seconds time the percentage of utilization. What they're wanting to limit here is that percentage of utilization of what they're selling you access to use.
One way to lower total monthly data transfer for a customer is to lower the data rate. That means things come down slower all month. Another is to limit the amount of time for which the line is fully utilized. Many business users of truly high-speed access pay for what are called "burstable lines". You get the whole DS3 or entire OC-12 or whatever type of line it is. You get billed with the understanding that you use a certain percentage of data rate or less a certain percentage of the time, and that the rest of the time you can use all of it without paying extra.
When I was in the ISP field and buying our backbone lines from bigger network providers, we typically leased lines with the first 15% or 25% of the full data rate included, with the stipulation that 5% or 10% of the time we could use all the data transfer the line had to offer without being billed extra. That meant that if we experienced abnormal peak demand, we didn't get our lines saturated. It also meant we didn't get soaked paying for peak capacity all the time. We in fact got a report each month showing the percentage, on average over each 5-minute increment, we used of the line's data rate the whole month. We could look at the chart being built (by MRTG) as the month progressed, too.
The reason total traffic is the way ISPs want to deal with end users is that it's easier to explain "you can move 20 gigabytes" than "90th percentile usage will be at no more than 30% of the data rate capacity of the circuit". Still, I think people would understand easily enough if they were told, "For 60 hours a month, you can max out your line. The rest of the time, you're going to be at 1 Mbps".
Not to excuse the cable company but they see it as that they're in a bind trying to trade off how many TV channels they can support (and how many analog ones in particular - (the sooner they die the better) with how many qams they dedicate to cable modems - and the expense of injecting the internet feeds in lower and lower down in the plant to support more and more customers with more and more bandwidth (ie sharing with fewer neighbors)
They shouldn't have ever offered 'unlimited' because as we all know it really isn't and for technical reasons can't be as the customer base increases - they're depending on statistical models which those 5% who use 50% of the resources (if that's a real number) break
> WTF do you do if 60GB will "cripple" you? Download 10 TV shows/day? Where do you put all the stuff?
When I lived in Toronto with 3 other engineering students, we ran over the 60 GB limit within a couple weeks when we were trying to limit our bandwidth usage (there was no surcharge, but frequent threats from Rogers to cut us off), with "regular" usage, where "regular" includes frequent multiple-ssh sessions with GUIs being displayed and ftp'ing bulk data to and from the school servers and so forth. If we would have been cut off, we would have been forced to commute to the school on all weekends and stay the night on many weekdays, and given the money a student can't afford and the time this takes it could well be crippling.
I agree with much of the rest of your point, though.
Um, I'd say that is accurate, I run an ISP, and without getting out my graphs and doing some basic math, I am tempted to say that is a _conservative_ estimate.
It is the same in tech support, 5% of my customers are the morons I hear from on a weekly basis. They account for about 75% of my total time spent on the phone.
--Nuintari
slashdot : where an opinion can be wrong.