[izym]

> The 1ms is network overhead, witch is nothing considering how long it takes for the image to render on the screen.

Yes, latency for a single bit of payload to ever hit the client. The rest of the data then needs to be transferred as well. Transferring that amount of data constantly would be terrible with regards to battery life.

> The CPU and the GPU could be several miles apart.

How exactly? Would you send draw commands over the internet?

[z3t4]

More data sent via carriers means they will be able to negotiate better deals and peering. So when 10Gbe hits the consumer market, they will let their users have 10Gbe downstream bandwidth, like a HDMI cable. With some compression you could stream 60 fps + sound. Sure it would use a lot of power, but the usage would foremost not be mobile devices, more like TV's for gaming, and workstations for other programs, BC to PC full circle. You could probably do a lot of optimizations like using a vector format, to drastically decrease bandwidth. I do not know exacly how it can be implemented technically, please give me break ... Some times things can be turned upside down and it's really hard to change your way of thinking, like HDDs gets faster then RAM or networks is almost as fast as internal buses. Not saying it is like that today, but it might be soon and it will be hard to break out of old design patterns.