The Facts & Fiction of Bandwidth Caps

wjamesau writes "What's the deal with broadband caps, like Comcast's 250GB/month data transfer limit, which goes into effect tomorrow? Om Malik at GigaOM has a whitepaper laying out the facts and fiction about Comcast's short-sightedness (which other carriers are mimicking), and how it will impact the future Internet: 'Given the growth trend due to consumers' changes in content consumption, today's power users are tomorrow's average users. By 2012, the bill for data access is projected to be around $215 per month.' Ouch." The white paper is embedded at the link using Scribd; for a PDF version you'll have to give up an email address.

Comments

Re:Rates that high will force rerouting

[jmorris42]

> Great idea. Quick question: how will that wifi network connect to the Internet?

Ok, work with me here. Imagine the bandwidth cap drops to 100GB/month. Hard drives are still cheap and huge and will be cheaper and even bigger by the time this problem ripens. 802.11n will also be commonplace by then. Ok, so everyone participating in a neighborhood net is expected to buy the current reflashable linky, at least 1TB of drive and a 10dbi omni antenna, The AP does all of your bittorrent action, something ASUS is selling now, a browser plugin offloads all .torrent links to the AP, you monitor your downloads on a webpage it provides and when it hits 100% you access the files via a samba share.

Ok, now put this AP on a 10/8 net and it can see the neighbors and your outbound net. It's torrent client has been modified to prefer local peers by a ratio close to the number of members. It also assists in torrents a neighbor is working even if you aren't interested in the file, at a lower priority on the pipe to the outside world. It does something else interesting, it only caches the blocks it downloaded, thus distributing a cache of those files amongst the peers and greatly increasing the effectiveness of the cache. If you later decide you want one of those files your client gets the rest almost exclusively from the local nodes.

Now imagine a future where video over the Internet was about to launch but the cable companies and telcos squashed it in favor of their video on demand pay per view crap. Get fifty neighbors together and together they have an aggregate bandwidth cap of 5TB. If everyone is watching a totally different set of shows it won't help much, but there will almost certainly be a fairly good overlap. When a new episode of moderately popular show X is available the dozen or so people who want it will be downloading it in parallel across their net links and swapping the blocks across a much faster 802.11n WWAN aided to a lessor extent by the 38 peers who aren't interested in that program. And cutting the hit on their bandwidth cap by that same factor of 12+ but offset by helping download stuff you didn't want to help somebody else. And if anybody else later decides they want to watch it before it times out of the caches their cost is zero. By having one smart host do almost all heavy downloading it can know the caps and adjust it's activity to avoid hitting them.

Mabe trying ot cut the compitition out

[teknosapien]

What about subscription based services? what if I'm subscribed to MLB.com and and watch every game I can and use Vonage on a consistant basis to make calls and I stream my music online? what effect would this have on my bandwidth and would it move me away from competing vendors? Would I then find it more cost effective to drop Vonage and use Comcast's Phone service and watch my games via subscription through Comcast? I think there is more here than meets the eye and only after it's implemented will we see the true fall out. After all what better way to kill the competition than to make it impossible to do business in your area