All of us agree that the D-Star is a great radio to work with. The radio offers many modes, and we have sent almost a giga-byte of data through it with our testing. Because we only have two units, however, we are not able test what happens during a collision.
Ethernet is a broadcast technology. When a computer wants to send a packet, it checks the network if it is busy, and if the line is clear, it will transmit the packet onto the network. When multiple computers are involved on a network, it is very possible for computers to want to access the network at the same time. When two or more computers transmit at the same time, it is called a collision. Collisions are resolved according to protocol by choosing a random time to retransmit the packet.
In the ethernet world, the re-transmission of the packet is done in milliseconds. In the D-Star world, we don't know. I would imagine it is much slower, due to key-up times of the transmitter.
I would like to get obtain two more radios and see how collisions are handled. On a larger network, it might be a concern. Then again, proper design can minimize the collision factor.
Update! I had the pleasure of talking with Matt Yellen from ICOM about collision concerns. He advised me that the TR switch on the D-Star units are sensitive down to the 10 ms range, meaning that the radios are responsive to "listen" to other packets. Matt also confirmed that the protocol in use on D-Star networks check for other stations attempting to make a connection. I am please to hear of this design. Matt also mentioned that other networks in use are also checking for collision response.