Today a coworker asked me about the limit of Ethernet links over Cat5 cable. I mentioned the 300ft. spec, and got curious about why the spec is that way. I assumed the spec was put in place because of some property of capacitance or resistance or something that is over my head. It turns out, though, according to this Wildpackets article, that the limit is actually to allow for propagation delay. Two hosts on a long run of cable might start talking at the same time, but not receive the colliding transmission from the host on the other end of the cable until they have transmitted more data than can be stopped using early collision detection. So that got me thinking about a few questions:
1. Would it be possible to greatly bypass the 300ft/100m limitation if the systems on each end had two NICs, one Send-only and one Receive-only, with two cables running between the hosts? The propagation delay shouldn't matter then. I see no possibility for a collision in this case.
2. Are there any systems in place that currently use an idea like this to provide long distance high speed communication over copper?
3. If this method works, what would be the new maximum cable length, and what would be the next limiting factor?