An emerging type of Ethernet that supports full-duplex communication between stations on the network. Full-duplex Ethernet lets stations send and receive data simultaneously, thus giving it twice the maximum throughput of traditional forms of Ethernet.
Full-duplex Ethernet uses two lines for sending and receiving data simultaneously. It is used in point-to-point connections between stations and requires that lines be concentrated using Ethernet switches instead of hubs or repeaters.
Using switches and point-to-point connections means that full-duplex Ethernet avoids the collisions that can degrade the performance of standard half-duplex Ethernet.
A full-duplex connection on a 100BaseT network would thus have a theoretical maximum speed of 200 Mbps, but in reality full-duplex Ethernet tends to achieve only a 20 to 60 percent higher throughput than standard Ethernet.
Full-duplex Ethernet does not use the traditional Carrier Sense Multiple Access with Collision Detection (CSMA/CD) media access control method of traditional Ethernet since collisions cannot occur on a full-duplex, point-to-point link between two stations. Because of this, the distance limitations between two stations in full-duplex Ethernet depend only on the strength of the transceivers with respect to the medium used. Thus station-to-station distances for full-duplex Ethernet connections can be much greater than for traditional Ethernet networks. For 100-Mbps full-duplex links, this is generally around 2 kilometers over fiber-optic cabling.
Use 20-Mbps full-duplex Ethernet to connect two 10BaseT Ethernet networks over duplex single-mode fiber-optic cabling at distances of up to 15 kilometers. To do this, use a pair of half-to-full duplex converters at either end of the fiber-optic line. These converters should always be used in pairs, and they typically have an attachment unit interface (AUI) port that accepts the fiber-optic transceiver.
Graphic F-24. Full-duplex Ethernet.