Interesting People mailing list archives
Re: WORTH READING Third Major ISP AT&T Testing Bandwidth Caps in the Fall [with a comment by me djf]
From: David Farber <dave () farber net>
Date: Sun, 8 Jun 2008 10:21:01 -0700
________________________________________ From: Tony Lauck [tlauck () madriver com] Sent: Sunday, June 08, 2008 1:06 PM To: David Farber Subject: Re: [IP] Third Major ISP AT&T Testing Bandwidth Caps in the Fall [with a comment by me djf] Dave, Along this line, Bob Briscoe's work is relevant. I believe this link was previously published on IP, but here it is again for the benefit of those who missed it: http://www.cs.ucl.ac.uk/staff/bbriscoe/projects/2020comms/refb/fair_ccr.pdf This paper has two key ideas: (1)resources need to be allocated according to economic interests (e.g. customers) and not technical objects (e.g. TCP connections or IP addresses), (2)when ascertaining resource usage by customer for purposes of providing fair allocation, only resource usage by a customer that delays *other* customers is relevant. There are political, regulatory and emotional issues involved when large government regulated monopolistic corporations are involved. Rather than debate those here, I will give a simple example of a cooperative network among three friends. This will clearly illustrate the idle resource issue. Suppose that three friends live a long way away from IP access and they decide to pool their resources and build a cooperative network. For geographical reasons the only significant cost of this network is the cost of building and operating a shared access link. Being somewhat frugal, they purchase a link with limited bandwidth. It provides good service when only one user is actively transmitting or receiving packets, but noticeably poorer service when two or more users are active. User C is the first to use the service and quickly becomes used to good service. After a while A and B also begin to use the network, and user C starts to experience reduced performance. Now suppose user C becomes annoyed at his poor performance when he uses the network in the evening. He decides to do something about it and identify which of his friends is causing him the problem, hoping to persuade that friend to send or receive fewer packets or to kick in a larger share so that a faster access link can be afforded. He looks at monthly usage statistics and concludes that he and user A are "moderate" users, but that user B transmits and receives ten times as many packets, clearly "excessive" usage. He decides to complain to user B. User B points out that he *never* uses the network in the evening and that he is not affecting user C's performance at all. He does a lot of bulk transfers in the middle of the night while A and C are sleeping, accounting for his heavy monthly usage. He suggests that C talk to A instead. While the situation is complicated in a real-world ISP situation by contractual, economic and political factors, not to mention more complex technical factors such as multiple bottlenecks and many more customers, the basic principle is the same. Usage of otherwise idle resources can never be considered excessive. Tony www.aglauck.com David Farber wrote:
I am at a loss to see how such a cap will really help. The issue is not the amount of bits you move but when you move it. If the competition for the bandwidth is sleeping then you can send with no impact. If you try when the kids get home from school things are busy and so large transfers slow things down. All an issue of busy hour design. What such caps DO create is an additional cost for people who are transferring large objects across the net -- like HD programs LEGALLY. Several of such transfers can eat up your allocation and then if they charge say $1/gig -- a HD can cost you what $3 to $4. Now usually the cable operator (and the FIOS) uses another channel to deliver VOD so, if I understand it right, they have created, by such a capacity model, a CLEAR competitive advantage in favor of themselves. Neat way around the NetNeurality potholes. Am I missing something. Dave http://gizmodo.com/5014290/welcome-to-the-future-of-broadband-third-major-isp-att-testing-bandwidth-caps-in-the-fall AT&T chief tech officer John Donovan has told Wired that they're going to test bandwidth caps in the fall, making them the third of the four major ISPs to do so <http://gizmodo.com/378760/will-your-isp-f-you-in-the-a-bandwidth-hogs-beware>. (Verizon stands alone, but for how long?) He lays out the familiar rationale, a small group of users (5 percent) pillage the network (40 percent) and they've got to stop them. But then he slips what's probably the real reason they've moving to caps: "Traffic on our backbone is growing 60 percent per year, but our revenue is not." It is more or less accepted that a minority of users use disproportionate of bandwidth, but what they're using it for is changing. It's increasingly video, not BitTorrent <http://gizmodo.com/382691/10-percent-of-broadband-subscribers-suck-up-80-percent-of-bandwidth-but-p2p-no-longer-to-blame>. The whole pro-BitTorrent thing is a smokescreen <http://gizmodo.com/373162/comcast-n-bittorrent-bff-whats-good-what-sucks>, because BitTorrent is less and less of an issue—video, and increasingly, HD video will be the real one. (Along with any number of other increasingly bandwidth-intensive apps.) And it'll be more and more competitive with providers' TV offerings—we've already seen Time Warner cry about it <http://bits.blogs.nytimes.com/2008/05/29/the-real-fight-over-fake-news/>. But there's no legitimate way to block it and protect their content. They can, however, make it more expensive for you to download with bandwidth caps (which is conveniently net neutral <http://gizmodo.com/gadgets/net-neuterality/att-considering-scary-content+recognizing-anti+piracy-filter-for-entire-network-320689.php>). And that's what I think this is partially about—protecting their TV business, not just curbing voracious bandwidth appetites. Regardless of the motivations, it's definitely coming. Comcast's tests will probably start soon <http://gizmodo.com/5012735/comcast-starts-net-neutral-slowdowns-of-heavy-broadband-users>, Time Warner's are already underway <http://gizmodo.com/5012427/time-warner-monthly-data-caps-detailed>and regional ISPs have <http://gizmodo.com/377955/the-future-of-broadband-were-totally-screwed> been doing it for a while. It's looking very much like the future of broadband here. At least if we're using it less maybe the internet won't explode now <http://gizmodo.com/381782/att-the-internet-will-explode-in-2010>. [Wired <http://blog.wired.com/business/2008/06/att-embraces-bi.html>] ------------------------------------------------------------------------ Archives <http://www.listbox.com/member/archive/247/=now> <http://www.listbox.com/member/archive/rss/247/> [Powered by Listbox] <http://www.listbox.com>
-- "Difficulties can never be greater than your capacity to solve them." - P. R. Sarkar Anthony G. Lauck PO Box 59 Warren, VT 05674 Southface 5 (for UPS and FedEX) 81 Park Ave Warren, VT 05674 (802) 583-4405 (802) 329-2006 (FAX) ------------------------------------------- Archives: http://www.listbox.com/member/archive/247/=now RSS Feed: http://www.listbox.com/member/archive/rss/247/ Powered by Listbox: http://www.listbox.com
Current thread:
- Re: WORTH READING Third Major ISP AT&T Testing Bandwidth Caps in the Fall [with a comment by me djf] David Farber (Jun 08)