359........
Here's some more on Hub/Switch duplexing differences from Linda @ HP Procurve Support
>>
By virtue of being repeaters, hub were never capable of performing at full-duplex. Hubs are collision-based devices. For example, a packet arrives, and the hub immediately floods it out all the ports. This negates the ability to full duplex. As you know, half-duplex can be likened to a walkie-talkie conversation; only one person may speak at a time. This is why hubs need to be half-duplex devices. The packet comes in, and is flooded everywhere. Nobody else may speak during this flood--otherwise, you get collisions. Most hubs don't even do error-checking, which is why it is so easy to propagate corrupt packets across the network.
Switches, on the other hand, are a little more sophisticated. It is possible to perform full-duplex on a switch, because it actually examines the packets minimally for errors, source MAC address, and destination MAC address. A switch will forward a packet to the one port where the destination MAC address is held in memory. The source MAC address will be added to memory for reference later--just in case someone wishes to send data to this device. The packet will then be delivered to the destination device, rather than the entire network. Of course, this is simplifying the process somewhat. Yet, this store-and-forward architecture makes full-duplex possible.
<<
Xenno