Ethernet - 1703

Overview
The term Ethernet refers to a family of buses in which the addressing, the format of the messages and access control
are identical (laid down in IEEE 802, [5]). Ethernet and the Internet Protocol (IP) were developed for data communication
between computers or peripherals which are locally separate and where, during operation, reconfigurations of the network through the addition of new users or the failure of users can occur. The Ethernet buses are identified by the following important features:
– The transfer rate is in the range of 10 Mbit/s up to 10 Gbit/s.
– Data transfer is possible via assorted media such as coaxial cable, twisted two-wire cable, radio or glass fiber.
– The technology involved is standardized and very widely used.
– Simple insertion and removal of nodes is possible.
– The time response in the case of real-time applications is not guaranteed.

Ethernet is used in series-production vehicles; e. g. in the BMW 7 Series, where it is used to input vehicle data at the end
of production.

 

Transfer system
The Ethernet versions differ in terms of transfer rate, physical design of the channel, and encoding. Coaxial cables,
twisted two-wire cables with one or more core pairs, optical fibers, radio paths, or even power-supply cables are specified as channels. Encoding differs accordingly. Originally, coaxial cables in a bus topology were used as the medium. Here, the transceivers of the nodes were connected either directly or with T-pieces to the cable. Today, twisted two-wire cables are widely used. The transfer rates have been in- creased from initially 10 Mbit/s via the Fast Ethernet with 100 Mbit/s and the Gigabit Ethernet with 1,000 Mbit/s up to 10 Gbit/s.

 

Topology
The size of a network is limited by the fact that the signal propagation time between two nodes influences the arbitration process. This can be bypassed by subdividing into segments, which are connected by way of special components – hubs and switches. A hub functions as an amplifier which re-establishes the ideal signal shape of a bit if it has been corrupted by interference or dispersion on the transmission medium. A switch checks entire packets for correctness with regard to the checksum and directs packets without collision to another output if the destination address can be reached by this route. To this end, it must have the option of storing messages temporarily. As well as the costs involved in the hardware, a disadvantage when such elements are used is that the data stream is delayed. For this purpose, however, nodes with different data rates can be connected. Today, networks are usually designed in such a way that each node is connected
to the output of a switch, i. e. there is no direct connection between nodes. Switches themselves are in turn connected via a higher-level switch in such a way that a tree-shaped structure is created.

 

Ethernet protocol

Bus access
To transmit, a node checks whether there are signals on the bus. It starts to transmit when it deems the line to be clear.
Because of the signal propagation time between two nodes, the situation may arise where two nodes deem the bus to be
clear and start to transmit virtually simultaneously. The data frames transmitted in the process are destroyed. The nodes
identify this, abort their transmission, and wait a certain amount of time – different for each node – until they begin with a new transmission attempt. This destruction of data frames reduces the effective transfer rate to a tolerable extent provided bus utilization is not too high. This arbitration process limits the length of messages and the propagation time,
i. e. the reach. There are no priorities among the messages. A maximum duration for transmission can, therefore, not
be guaranteed. Each node adopts from all the messages those which contain its own as the destination address for further processing.

 

Data frame
Figure 11 shows the slightly simplified structure of a data frame. The preamble is a periodic bit sequence (101010 – 1011)
and thereby generates a signal to synchronize the receiver. Messages contain the address of their source and their des-
tination. Each network card has a unique address. The receiving nodes compare the destination address with their own
card address and accept the data frame if they match. With multicast and broadcast addresses, several receivers can also be addressed.