What is the meaning of half-duplex and full-duplex Ethernet?

Half-duplex means that only one direction can be transmitted at the same time during transmission. After the data transmission of one party is over, the other party responds. That is to say, only one node can transmit at the same time. If two nodes transmit data at the same time, congestion will occur in the network.

Half-duplex Ethernet uses the CSMA/CD protocol to prevent collisions. If a conflict occurs, retransmission is allowed. If you use a hub to build Ethernet, you must work in half-duplex mode because the end station must be able to detect collisions. Half-duplex Ethernet - typically 10BaseT, with only 30% to 40% efficiency. Because a large 10BaseT network usually only gives a maximum bandwidth of 3 to 4 Mb/s.

Full-duplex Ethernet uses two pairs of cables instead of a pair of cables as in half-duplex mode. The full-duplex mode uses a point-to-point connection between the sender of the transmitting device and the receiver of the receiving device, which means that a higher transmission rate can be obtained in the full-duplex data transmission mode. Since the transmitted data and the received data are done on different cable lines, there is no conflict.

Full-duplex Ethernet does not create conflicts because it is like a highway with multiple entrances, not a road with only one entry, as provided by the half-duplex approach. Full-duplex Ethernet provides 100% efficiency in both directions. For example, you can get 20Mb/s transmission rate with 10Mb/s Ethernet running in full-duplex mode, or increase the transmission rate of Fast Ethernet by 200Mb/s, which is very remarkable. However, this rate is sometimes referred to as the rate of aggregation, which means that you need to get 100% efficiency, just like things in life, which cannot be fully guaranteed.

Half-duplex means that the sending and receiving of the same media is asynchronous. In contrast, full duplex has separate transmit and receive paths. Full-duplex links are the key to extending Fast Ethernet (100Mbps). A full-duplex link network segment cannot exceed two devices, and can be a network card or switch port. Note: It is not a relay hub port, and the hub does not have full duplex mode. This is because the hub is part of the collision domain and it enhances the collisions received by other ports. Only two NICs can implement full-duplex communication. For full-duplex mode when there are more than two NICs, the switch must be considered.

10BASE-T and 10BASE-FL have separate transmit and receive paths, and full duplex can be performed depending on the complexity of the network card or switch port. If these interfaces are configured in half-duplex mode, the synchronous detection of receiving and transmitting will trigger the collision detection. The same interface is set to full duplex, and collision detection is disabled because full duplex does not follow the shared CSMA/CD rules.

The configuration of the full-duplex link is correct. When the site is configured in full-duplex mode, the port of the site or switching hub sends frames in a manner that ignores the CSMA/CD protocol. If the other end is set in half-duplex mode, it will detect the collision and cause other problems, such as CRC error, the speed of the network drops, and the advantage of Fast Ethernet disappears.

As mentioned earlier, the network range at 100 Mbps has shrunk due to collisions. For twisted pair segments and switch ports, the longest distance of the segment is 100 meters (within the collision domain). The problem is that on fiber ports, the length of the segment is 2 km for multimode fiber and 15 km for single mode fiber. In the half-duplex mode, it is limited by the collision domain, and the network segment distance is 412 meters. Therefore, only in full-duplex mode (CSMA/CA is ignored), the extension of the fiber segment can reach its limit. In the fast industrial Ethernet mode, switch technology is recommended. For fiber ports under Fast Ethernet, full duplex is recommended.