Properties
7/19

The second layer of Ethernet communication provides important basic functions for controlled data transmission. Besides the uniform message structure, these include the addressing of nodes and the bus access method. All basic functions are implemented in the Ethernet controller, which is generally an integral component of a microcontroller today.

While symbols and symbol rates are spoken of on the physical layer, bits that are assembled into Ethernet frames are used on the second layer. The bit stream is typically transmitted between the Ethernet PHY and Ethernet controller with the Media Independent Interface (MII). This is an IEEE-standardized interface family that offers multiple variants for different transmission speeds.

The Ethernet controller first listens on the physical medium before a message is sent (carrier sense). This prevents a message from being overwritten if another node in the network is already sending. If the medium is free, the Ethernet controller can begin its own transmission.

Since multiple nodes may access the bus simultaneously with Ethernet (multiple access), collisions may occur on classic bus networks if two nodes begin sending at the same time. For these situations, each Ethernet controller has a collision detection function that is used to cancel a transmission. To prevent a second collision, a node starts resending only after expiration of a random time (backoff process). Each sender must calculate the time itself.

The complete bus access method is referred to as Carrier Sense Multiple Access/Collision Detection (CSMA/CD). The associated algorithm is implemented in each Ethernet controller. For physical layers in the automotive industry, collision detection plays a rather subordinate role. IEEE 100BASE-T1, IEEE 100BASE-TX, and IEEE 1000BASE-T all allow full-duplex data transmission. For this reason, collisions normally do not occur on these physical media.