Slashdot Mirror
Comcast's Throttling Plan Has 'Disconnect User' Option
newtley writes "Comcast's new people, not protocols scheme may mean high speed for some, but by no means all. It's also created a draconian 'disconnect' option for use against anyone who fails to toe the Comcast line. But, says Robb Topolski, the Net protocol expert who originally uncovered Comcast's blatant efforts to control its customers, the plan does offer key take-aways, telling P2P users on Comcast how to do what they do without the risk of corporate interference."
The problem I have with bandwidth caps as offered by ISPs is that when the ISP is also the cable provider the bandwidth cap is anti-competitive with Hulu and other video entertainment sites. As far as I can tell this is prime territory for an anti-trust investigation.
IANAL but it seems to me that these caps are not because of P2P but put in place because of competition for the television audience. By capping the users Comcast seems to be trying to guarantee that their cable service is still viable.
load "$",8,1
I agree. The cap is a perfectly good idea. Giving users no way to see how close they are to their cap, and cutting people off for exceeding it, are terrible ideas.
I see no reason why I, a moderate internet user, should subsidize that guy down the street who downloads 1TB of torrents every month. He uses more, he should pay more.
But the way Comcast is going about it is stupid. They're trying to have their cake and eat it too, essentially. An explicit cap can lead to more traffic, since now people know what the limit is and what they're really paying for, and they may decide that they should use more of what they're paying for. I think they're trying to limit the top people without causing this sort of increase, and doing this by having an explicit cap that still happens to be vague and dire.
If you were to do this right, you should really have a system where many different caps are available. You'd have a default one, probably well under 250GB, that comes with a service that's cheaper than what they offer now. Then you can pay more to increase your cap. You'd be able to monitor your usage, get a warning well before you hit the cap, and increase your account's cap at any time just by requesting it. And if you do hit your cap, then your account gets throttled to dialup speeds until your 30-day sliding window average decreases below the cap level.
Of course this would make far too much sense so Comcast won't do it, but it's what they ought to do.
If you mod me Overrated, you are admitting that you have no penis.
All of the major cell phone companies give you a free text message and/or wireless web page that tells you as best as they can how many minutes/bytes you've used this billing cycle and such. Why Comcast can't do the same for their bandwidth limit is beyond me.
Then you haven't thought very hard about it from Comcast's point of view. It makes perfect business sense. Developing a customer-ready bandwidth usage meter has very real fixed and recurring costs to Comcast, costs which have no potential to increase profits now or in the future. If customers are going to switch to or from Comcast, it will be because of the cap, not because of the availability of a usage meter.
Additionally, an easily-viewable bandwidth meter would in all probability only encourage customers to get much closer to the limit than they would otherwise. It's fear-based policy. The more of their customers that decide "I'd better not download this movie/album/ISO/whatever, I might hit my bandwidth cap", the better. Comcast wants customers to stay in the dark regarding usage and be as conservative as possible in their internet activities, while still pretending to offer the full 250 GB.
there's more to it than the cumulative cap. they also have an elaborate throttling scheme based on how much you're currently downloading:
The issue will be their strange throttling scheme, which puts users in a "penalty box" for using more than 70% of available bandwidth in any 15 minute window and releases them from the box when their activity drops below 50%.
It has the net effect of decreasing the effective sustained bandwidth. I don't have Comcast, and I think the cumulative limits are fair, but this strikes me as unfair. What if I don't come close to the monthly limits, but I'm streaming/DLing something that will take longer than 15min? If congestion isn't an issue, why not let someone DL at the capabilities of their connection?
Relax. Have a muffin. Enjoy the show. --Slick, Sept 13th, 2007.
Oh, you mean on the Comcast lines that were partially funded by U.S. taxpayer dollars that were given to Comcast (and others) to get the Internet infrastructure to reach as many people as possible? And now those lines that were funded by the public are to become new profit centers under the guise of "network management"? And you don't have a problem with that?
How about instead Comcast actually do what they were supposed to do and build capable infrastructure that has enough bandwidth for everyone to do anything?
Personally, I would love for the General Accounting Office to take a nice, close look at Comcast's finances to find out exactly where that taxpayer money went to. Looks more like it went into Comcast's advertising budget so that they could oversell their capacity instead of putting it into the hardware that could have prevented all of this in the first place.
Verizon has millions of miles of dark fibre and have said numerous times that they have plenty of bandwidth as it is. What's Comcast's excuse?
The Overrated mod is for reversing inappropriate, positive mods, not for voicing disagreement with a post.
But then no bitching if all you can buy is 256kbps. Bandwidth isn't free and the larger the links get, the more pricey they are. You can see this with LAN hardware. It is damn near impossible to get 10mbit switches anymore, 100mbit is the minimum and those are cheap as hell. However gigabit goes up a good deal in cost. A 24-port 100mbite switch might run you $100. A 24-port gigabit switch from the same vendor is over $400. Ok well then 10gbit goes waaaay up. Now you are talking thousands of dollars to get a gigabit switch with even a couple 10gbit ports, and then several hundred per port to get the transceivers.
Now suppose you want to design a network for 500 computers on 5 floors (100 per floor) that gives 100mbit to the desktop. So you get a bunch of 24-port switches and hook them together. Turns out you need about 31 of them. 1 central switch, 5 floor switches and 25 access switches. Those are about $100 each so $3100 total. Ok great.
However you then decide you want everyone to always get their full 100mbit. So now you still connect the computers with 24-port 10/100 switches. However those switches need to have at least 2 gigabit ports (channeled together) on them for uplink, assuming you hook 20 PCs to each. So you now need 25 access switches, but each now costs $180. That's $4500 for for the access switches. Now on each floor, your floor switch has to be able to take 10 1gbit connections in and so a 10gbit connection out. For that you are talking about $2500 per switch for the switch and transceiver. So $12,500 for those. Then for your core switch, you need something with 5 10gbit ports. That is getting extremely high end, and is nearly $10,000 for the switch and transceivers. So for this solution you are talking $27,000.
Well that's a difference of $6/computer and $54/computer. Costs a hell of a lot more to do guaranteed bandwidth. Also this is just a small scale example. Now suppose you have 10 buildings that need connection, then 20 cities with 10 building complexes, and so on. Gets amazingly expensive if you have this "Everyone must have dedicated bandwidth" idea.
What's more, you'd find that for less than that, you could do something that's better overall. If you ran gigabit to the desktop, gig to the floor switches and then gig or maybe 2 gig to the core you'd find that in real usage, everyone would have faster transfers, and you'd pay a less than the dedicated 100mbit solution. Yes, it can get overloaded, however so long as people share it'll actually be faster for everyone.