Ethernet

Ethernet organizes link-layer data into frames carrying addressing and control information. An IP datagram can be frame payload, but Ethernet and IP addresses belong to different protocol layers. A host can use Ethernet locally and reach other link technologies through a router without requiring the entire Internet to use the same physical medium.

References: [1]

Early shared media required coordination among senders competing for one transmission resource. Switched links changed that organization, and full-duplex operation permits simultaneous sending in both directions. Performance depends on the medium, duplex mode, and switching arrangement, so the behavior of an early shared bus does not describe every modern Ethernet installation.

References: [2]

Frame formats allocate space to control fields and payload, with padding for minimum lengths and limits on maximum carriage. RFC 894 explains that Ethernet padding is not part of the IP datagram. Application data, upper-layer overhead, and link capacity must therefore be considered separately when determining how much information one frame can carry.

References: [1]

The IEEE 802.3 working group develops Ethernet standards for different media, distances, and environments. A shared framework supports interoperability, while specific connections still require compatible interfaces, encoding, and rates. Ethernet denotes both historical technology and an evolving standards family, making the particular form of a connection important to its technical description.

References: [2]