ITWissen.info - Tech know how online

Ethernet

Ethernet emerged from a project of the companies Digital Equipment, Intel and Xerox in the seventies, which became known as DIX. This project aimed at sharing a transmission medium between several data stations with equal rights and was designed for the local area. The structure of this local area network was that of a bus to which all data stations could be connected. The bus concept was limited in its extent, in the number of stations that could be connected and in data rate to 3 Mbit/s and was therefore only suitable for the local area. The DIX project became Ethernet in 1982. The specification and standardization was taken over by the IEEE, which published the specifications DIX Ethernet V2.0.

Since Ethernet developed rapidly in the following decades and the term Ethernet is also used for high-speed variants, the original Ethernet is also referred to as such.

The classic Ethernet

Ethernet, specified by IEEE Working Group 802.3, is a local area network to whose transmission medium equal data stations are connected. The data stations work with a stochastic access method in which they detect signals on the transmission medium and then, if no other signal is present, use it for their own data transmission. The access method is CSMA/CD and can lead to collisions between the transmitted data signals.

The original Ethernet according to 10Base-5 operates with a transmission speed of 10 Mbit/s. All data stations have equal rights and are connected directly to the transmission medium. The number of data stations is limited to a maximum of 256 stations per Ethernet segment, the length expansion is limited by the signal propagation time and the frame length. As the segment length increases, the propagation time between the two stations furthest apart increases, which has a negative effect on the time required to detect a collision. All three parameters, the number of stations, the frame length and the segment length, determine the performance of Ethernet. The shorter the frame length and the higher the number of stations ready to transmit, the lower the performance will be. The frame length of the Ethernet frame affects performance in that the ratio of user data to frame length decreases as the frame size becomes smaller.

The permissible network expansion of an Ethernet depends on the Ethernet variant and thus on the transmission speed on the network. In general, the expansion decreases with increasing transmission speed. If we consider the classic Ethernet with a transmission speed of 10 Mbit/s, the permissible network expansion is determined from the round-trip delay on the transmission medium. This is determined by the propagation time between the two most distant stations. With a cable-specific signal propagation delay of 0.77c, this results in a propagation delay of 2.165 µs for a 500-m segment. Based on the minimum packet length of an Ethernet frame of 64 bytes and the transmission rate of 10 Mbit/s, the round trip time between the two most distant stations is 51.2 µs, i.e. the simple propagation delay is 25.6 µs. This round trip delay corresponds to an extension of about 5 km. Taking into account the signal propagation times in the transceiver cables and the delay times of the active components, Ethernet extensions of about 3,000 m are thus possible.

Ethernet layer model for classic Ethernet

Ethernet layer model for classic Ethernet

The functionality of Ethernet is ensured by the Ethernet layer model. In classic Ethernet, the data link layer is divided into the two sub-layers Medium Access Control (MAC) and Logical Link Control (LLC). The physical layer consists functionally of the Physical Layer Signalling (PLS), Access UnitInterface (AUI) and Medium Attachment Unit (MAU) sublayers.

Historical Yellow Cable for 10Base-5

Historical Yellow Cable for 10Base-5

The transmission medium of 10Base-5 was a thick coaxial cable, the so-called Yellow Cable, which was only used until the end of the 1980s due to the complicated and error-prone connection technology and was supplemented by a thinner coaxial cable, TP cables and optical fibers. Accordingly, the first Ethernet interface was 10Base-5 for a distance of 500m. It was followed by 10Base-2, 10Base-T for twisted pair and 10Base-F for fiber.

From the classic Ethernet with 10 Mbit/s to 400 Gigabit Ethernet.

Ethernet has undergone a rapid development and has established itself in all speed and application areas. While in the 1980s it was the 10 Mbit/s variants that used coaxial cable, in the 1990s the various TP cables were added, as was fiber optics.

Ethernet in the LAN, AN, MAN, WAN and industrial sectors

Ethernet in the LAN, AN, MAN, WAN and industrial sectors

In 1995, Fast Ethernet with 100 Mbit/s and the switching systems were added, and as early as 1998, Gigabit Ethernet (GbE), which is standardized, as is 10 Gigabit Ethernet (10GbE). In 2010, IEEE 802.3 standardized the HS variants 40-Gigabit Ethernet (40GbE) and 100-Gigabit Ethernet, and in the following years 400-Gigabit Ethernet. Ethernet variants with data rates above 100 Gbps are referred to as Terabit Ethernet (TbE), although 400 Gbps corresponds to a data rate of 0.4 terabits per second (Tbps).

As far as access methods are concerned, the classic CSMA/CD to Fast Ethernet was retained, only with Gigabit Ethernet (GbE) were certain restrictions made in this respect, and with 10 Gigabit Ethernet (10GbE) this method was abandoned completely.

Ethernet for a wide range of applications

In the meantime, solutions have also been developed for Ethernet in the metro and access area: Ethernet in the First Mile (EFM), Metro Ethernet and Midband Ethernet are interesting approaches for access and city networks. Carrier Ethernet, 40 Gigabit Ethernet and 100 Gigabit Ethernet are also available for wide area networks and are also used as backplane Ethernet in servers.

Application areas of Ethernet

Application areas of Ethernet

Nevertheless, it seemed that certain applications could not be realized. For example, lossless data transfer as required in data centers and storage networks. Lossless Ethernet was developed for this application and is used in Data Center Ethernet (DCE) and Fibre Channel over Ethernet (FCoE). Another seemingly limited application is the transmission of time-critical signals, such as those found in Internet telephony, live streaming and industrial applications. Real-time Ethernet and Industrial Ethernet have been developed for these applications, which operate with the lowest possible known latency. The networks used in manufacturing and automation use Ethernet-based techniques with real-time behavior. Some fieldbuses use Ethernet as the communication infrastructure for integrating distributed field devices, plant modules, and control devices, including EtherCAT, TTEthernet, Ethernet Powerlink, and EtherNet/IP.

Automotive Ethernet is also used in automotive technology as UTSP Ethernet, Unshielded Twisted Single Pair. The process is called Single-Pair Ethernet (SPE) and transmits data at 10 Mbit/s, 100 Mbit/s and 1 Gbit/s via a twisted pair of wires.

Informationen zum Artikel
Englisch: Ethernet
Updated at: 29.09.2019
#Words: 1886
Links: