It's all about managing expectations and taking advantage of interest rather than fighting your customers...

I can understand why people would want to download TV shows and other similar content to watch at their leisure (like during a boring meeting or presentation at the office :)). The problem here is not that people want to do that - because that's an opportunity. You can make money out of such customer behaviour, if you apply some ingenuity. Think about it: people want to use your service, and they want to use it a lot. If you say "nah, you can't do that" you open yourself up to a competitor who will allow it, in part or full.

The problem is that such usage wasn't predicted before. The shared-bandwidth model for the network assumes that all we're doing is surfing our inboxes rather fast and then not using the connection for a while. With P2P and similar content, that design falls over. IP isn't perhaps the most efficient way to deliver large contigious data streams either, but that's another issue.

All telco and network architects I've spoken to have been aware of this for a long time now. They've got some wild prognosis on bandwidth usage increases, but also ideas on how to redesign the network to cater for high, continous utilisation. One such idea is to move the switching and routing closer to customers, so as to reduce the path lengths for data and keep utilisation local so to speak.

Ultimately, I'm sure there will be a levelling-off point in the future where there's sufficient bandwidth for the content, but we're a long way from that still.