3 ms·

Residential broadband is a business model that depends heavily on successful bandwidth utilization speculation. However since Netflix is extremely popular (clo

Residential broadband is a business model that depends heavily on successful bandwidth utilization speculation. However since Netflix is extremely popular (close to half of all internet traffic in the US), the pattern gets extremely predictable and so there is little advantage to be had in the core business of bandwidth arbitrage, since it's being arbitraged "upstream" of residential ISPs and baked into the larger pipe prices. If any service becomes sufficiently popular, the price of providing bandwidth for it to a large subscriber base approaches the characteristics of a dedicated bandwidth circuit for more and more nodes on the network. Of course, Comcast is using it's last mile power as leverage, but the core issue the economics of bandwidth speculation.

jberryman 13y ago
Also something I never see mentioned is that this sort of seems to be how popularity works; that is "Popularity" tends to be zipf distributed
jjoonathan 13y ago
A very interesting observation! However, bandwidth speculation would only be the core issue in a competitive market where price was determined primarily by infrastructure costs. What we have is the opposite of that. The core issue is leverage. Bandwidth speculation is a distant second, even though it's much more interesting to us tech folks.
adventured 13y ago
The best accounting I've seen indicates Netflix is at roughly 1/3 of North American traffic: http://allthingsd.com/20130514/netflix-still-eats-a-third-of-the-web-every-night-amazon-hbo-and-hulu-trail-behind/ http://allthingsd.com/20130514/netflix-still-eats-a-third-of... Similar to where they were at the year prior: http://allthingsd.com/20121107/netflix-has-plenty-of-competitors-and-none-of-them-are-close/ http://allthingsd.com/20121107/netflix-has-plenty-of-competi...
- Retric 13y ago
  It's vary time of day dependent. At 10AM EST Netflix is a much lower percentage than at 10PM EST.
  - 6cxs2hd6 13y ago
    > EST Also EDT. :) I usually just write ET to mean both. (And CT, MT, PT, etc.) Yeah I'm being picky. It's a pet peeve. Sorry. Edit: After a downvote -- actually, I'm not sorry anymore for pointing this out. Roughly half the year there is no such thing as EST. Think that's picky? Bite me. (Note that I didn't grammar police their use of "vary" instead of "very".)

jberryman 13y ago

Also something I never see mentioned is that this sort of seems to be how popularity works; that is "Popularity" tends to be zipf distributed

jjoonathan 13y ago

A very interesting observation! However, bandwidth speculation would only be the core issue in a competitive market where price was determined primarily by infrastructure costs. What we have is the opposite of that. The core issue is leverage. Bandwidth speculation is a distant second, even though it's much more interesting to us tech folks.

adventured 13y ago

The best accounting I've seen indicates Netflix is at roughly 1/3 of North American traffic: http://allthingsd.com/20130514/netflix-still-eats-a-third-of-the-web-every-night-amazon-hbo-and-hulu-trail-behind/ http://allthingsd.com/20130514/netflix-still-eats-a-third-of... Similar to where they were at the year prior: http://allthingsd.com/20121107/netflix-has-plenty-of-competitors-and-none-of-them-are-close/ http://allthingsd.com/20121107/netflix-has-plenty-of-competi...

Retric 13y ago
It's vary time of day dependent. At 10AM EST Netflix is a much lower percentage than at 10PM EST.
- 6cxs2hd6 13y ago
  > EST Also EDT. :) I usually just write ET to mean both. (And CT, MT, PT, etc.) Yeah I'm being picky. It's a pet peeve. Sorry. Edit: After a downvote -- actually, I'm not sorry anymore for pointing this out. Roughly half the year there is no such thing as EST. Think that's picky? Bite me. (Note that I didn't grammar police their use of "vary" instead of "very".)