While this argument sells to the crowd, it's facile and doesn't hold up to scrutiny.
I'm really quite a layman in this area, but I've heard the overselling argument quite a lot, and I'd be interested in hearing why it might not be a valid one.
I have a 15Mbps home connection, which serves me brilliantly when I need to grab an ISO or watch a couple of videos. I have no illusions that this is equal to a business 15Mbps connection, which would be significantly more expensive. A T3 runs $3000 a month, best case. I pay $45 a month.
The original point that I disagreed with was the assertion that everyone should have a unmetered 6Mbps connection. Firstly, I don't want a 6Mbps connection -> I want a 15Mbps burst connection that best accommodates a normal user. Secondly that's enough bandwidth that if you filled it at Amazon's AWS, you'd be paying some $240 a month, and Amazon is at the crossroads of a number of primary connections, pretty much the cheapest place to possibly origin data.
Joe Homebody is many nodes down, through a lot of infrastructure and choke-points, and the cost for throughput is significantly higher: My neighbourhood has a large number of cable modem subscribers, and to aggregate the total sum up through each junction would lead to enormous bandwidth potential.
In the mobile space the situation is vastly uglier. I don't think most people realize how incredibly crowded the data space is.
I would love to know how much of the Japan rhetoric is actually based in reality.
It does hold up, the math is all there. However what it does NOT provide is much room for burst capacity.
What I am saying is that the carriers could simply set reasonable expectations, and provide the consumer the option to choose what kind of connection to the Internet they wanted. They might even be able to provide priority routing to a specific content provider, if that became worthwhile in the future.
IE:
You can get a 15Mbps burstable connection for $50/mo. For this you might get 15Mbps, or you might get 6Mbps, but we'll guarantee you never get less than 2.5Mbps. It might also be a 95th% style billing scheme.
You can get a 5/5 symmetrical dedicated connection for $65/mo. We might even be nice and let you burst to 8Mbps if the capacity is there, but you'll never get less than 5Mbps, and you'd likely never get more than 8 either.
For an extra $10/mo on either plan you can have a "friends and family" routing concept where any traffic to the content provider of your choice (Hulu or Yahoo or Google, etc.) is guaranteed to never be less than 4Mbps.
Instead, it appears that what they are driving at is selling you a gob of bandwidth for $x/mo., unless you want to use that gob to connect to a business deemed a competitor or threat, then it's 2x.