How Ethernet works
The Institute of Electrical and Electronics Engineers Inc. (IEEE) specifies in the family of standards called IEEE 802.3 that the Ethernet protocol touches both Layer 1 — the physical layer — and Layer 2 — the data link layer — on the OSI network protocol model. Ethernet defines two units of transmission: packet and frame. The frame includes not just the payload of data being transmitted, but also:
- the physical media access control (MAC) addresses of both the sender and receiver;
- VLAN tagging and quality of service information; and
- error correction information to detect transmission problems.
Each frame is wrapped in a packet that contains several bytes of information to establish the connection and mark where the frame starts. Engineers at Xerox first developed Ethernet in the 1970s. Ethernet initially ran over coaxial cables, while a typical Ethernet LAN today uses special grades of twisted pair cables or fiber optic cabling. Early Ethernet connected multiple devices into network segments through hubs — Layer 1 devices responsible for transporting network data — using either a daisy chain or star topology.
If two devices that share a hub try to transmit data at the same time, however, the packets can collide and create connectivity problems. To alleviate these digital traffic jams, the IEEE developed the Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol, which allows devices to check whether a given line is in use before initiating new transmissions. Later, Ethernet hubs largely gave way to network switches, their more sophisticated and modern counterparts. Because a hub cannot discriminate between points on a network segment, it can’t send data directly from point A to point B. Instead, whenever a network device sends a transmission via an input port, the hub copies the data and distributes it to all the available output ports.