Ethernet

Standard 8P8C connector, often called RJ45

Ethernet standards define wired connections and electrical signals at the physical layer, personnel format, and media access control protocols at the link layer of the OSI model. Ethernet is mainly described by the IEEE standards of the 802.3 group. Ethernet became the most common LAN technology in the mid-1990s, replacing such outdated technologies as Arcnet, FDDI and Token ring. The Ethernet Alliance community continues to develop the technology.

Main article: Time-Triggered Ethernet (TTE)

Ethernet-APL (Advanced Physical Layer)

Main article: Ethernet-APL (Advanced Physical Layer)

Ethernet LAN Switches Market

• Main Article: Ethernet LAN Switches (Global Market)

Ethernet history

2023: A number of major IT companies announce formation of global Ultra Ethernet Consortium

On July 19, 2023, a number of large IT companies announced the formation of the global Ultra Ethernet Consortium (UEC) to create a scalable and efficient communication stack for next-generation networks. They will be created with an eye on high-performance computing (LDC) systems and artificial intelligence. Read more here.

2016: Development of 6 new speed standards

In previous years, the world has received six Ethernet standards:

• 10Mbps,
• 100Mbps,
• 1Gbps,
• 10Gbps,
• 40Gbps and
• 100Gbps.

In 2016, the Ethernet community is working hard to implement new speed standards:

• 2.5Gbps,
• 5Gbps,
• 25Gbps,
• 50Gbps,
• 200Gbps and
• 400Gbps - for the next three years.

High-frequency Ethernet implementations have become a response to increasing bandwidth requirements (information transfer rates). However, a tenfold increase in Ethernet speed introduced new requirements for the physical transmission medium.

For 2016, the most common cable types are categories 5e and 6. Meanwhile, the development of wireless networks has resulted in the 802.11ac standard, which in certain cases can provide a wireless connection at a speed significantly higher than 1Gbps, which is the limit for backbone ports of wireless access points that simultaneously receive power through the same cable (Power over Ethernet). The fact that 70 billion meters of the released twisted pair of categories 5e and 6, for physical reasons, are not able to provide 10Gbps speed, led to the beginning of the development of Ethernet standards 2,5Gb and 5Gb BASE-T. These new specifications can be applied to the existing cable infrastructure, giving the required speed boost for enterprise applications using the 802.11ac standard.

2016 Ethernet Alliance Roadmap.

Meanwhile, while there was a rapid spread of data centers based on 10GbE and 40GbE, the developers came to the conclusion that the easiest way to reach the strip in the 100Gb was to replace 10 tracks per 10Gbps with 4 tracks per 25Gbps. There was also a need for Ethernet speeds in excess of 100Gbps, which led to the development of the 400GbE.

Work on this high-speed Ethernet standard differed from 40Gb and 100Gb development, as new technologies were needed to find a practical solution to 400Gb Ethernet. The following shows the options that can be taken into account in the development of new high speed standards and how they were applied to the development of the target specification in the 400Gbps.

Ethernet 400Gb Technology Solutions

A large number of optical fibers operating on the 25Gbps have been selected as the solution for operating at 100 meters on multimode optical fiber (MMF). In the case of Single Mode Fiber (SMF) solutions, operation at 50Gbps and above was chosen, together with high-order PAM4 modulation. And while the 500-meter solution uses 4-fiber PAM4 100Gbps, the 2km and 10km solutions have added additional optical lambda, and they use 8 lambda per 50Gbps.

While there was a debate about various technologies to achieve 400GbE, a different development strategy appeared. The development of 25Gbps transmission to support 100GbE, including operation on switch backbones, copper two-wire shielded cables, and multimode optical fiber, prompted the realization that 25GbE-enabled servers using data 100GbE could be used in the same way as 10GbE and 40GbE and lead to a new generation of large data centers. The standard required for this jump will be completed in the near future.

Together with the upcoming completion of this standard, and with data centers developing in big steps, it was discovered, again, that high-speed Ethernet was necessary for a new generation of servers outside 25GbE. And with new 50Gbps transmission technologies designed to support 400GbE, the choice is obvious - 50GbE.

This, however, poses another question: what will be the right network solution? Given the success of the 10GbE/40GbE, and the implementation of the 25GbE/100GbE, it was determined that the best solution for servers is the maximum data transfer rate, using a network solution based on a 4x increase in this speed. Thus, the industry is now in the early stages of 50GbE and 200GbE development. The figure below shows the situation in the Ethernet industry.

As mentioned, Ethernet is no longer being developed in the same speed format for use everywhere and everywhere. Instead, Ethernet speed families appear, based on various data technologies. The basic transmission rate and its quadruple increase: the first generation is 10GbE and 40GbE; the next generation is 25GbE and 100GbE; and the last generation of 50GbE and 200GbE.

Solutions based on 2x and 8x look logical and are implemented. Transmission standards 100Gbps PAM4 are under development, so the fourth generation of data center architecture is clearly visible on the horizon. Moreover, if we consider the development of an eightfold increase solution based on 100Gbps, the next Ethernet speed after 400Gbps will obviously be 800Gbps.

Ethernet continues to evolve and at this time we see how the industry has focused on developing many new Ethernet standards, increasing the flow of investment in the next generation of technologies to achieve this result.

During this time, John D'Ambrosia leads the Ethernet Alliance, a global association aimed at achieving success and spreading Ethernet technologies. He is also Huawei's chief engineer.

2012: IEEE 802.3-2012 approved

White Paper: IEEE 802.3-2012 Ethernet Standard

The IEEE Standards Association, part of the Institute of Electrical and Electronics Engineers (IEEE), approved the new generation Ethernet network standard - IEEE 802.3-2012 in the summer of 2012.

2010: Energy Efficient Ethernet (IEEE 802.3az IEEE Energy Efficient Ethernet)

Energy Efficient Ethernet (IEEE 802.3az) reduces power consumption by automatically and realtime adjusting power consumption according to actual network traffic generated by switches and other network devices.

In December 2010, HP announced that it would be the first to begin shipping products based on the new energy efficient Ethernet standard developed with HP. This will allow enterprises to reduce energy consumption and operating costs for IT equipment.

The new zl modules for HP E-Series switches are the first devices built according to the IEEE Energy Efficient Ethernet standard, which, in the absence of traffic, can go to sleep after the EEE devices connected to them. With HP's efforts to implement this standard, customers will benefit from reduced power consumption in both switches and endpoints, which should reduce total cost of ownership by 51%.

During periods of reduced load, energy efficient Ethernet devices go into sleep mode, in which power consumption is reduced, but connected devices can be activated as soon as data transfer begins. This results in significant power savings compared to conventional switches capable of performing only a limited correlation between power consumption and actual traffic.

Since network loads are predominantly explosive, network equipment is an ideal platform for implementing energy-efficient Ethernet standards that automatically regulate power consumption depending on traffic.

In the future, the IEEE Energy Efficient Ethernet standard was expected to be used in all devices, including servers, laptops, and wireless access points. This will reduce energy consumption and therefore reduce the cost of enterprise IT infrastructure.

1990: IEEE approves the standard 10Base-T

In September 1990, the IEEE approved the 10Base-T standard.

1985: LattisNet

The next step in the development of Ethernet was the development of the 10Base-T standard, which provided for Unshielded Twisted Pair (UTP) as a transmission medium. This standard was based on the development of SynOptics Communications under the general name LattisNet, which dates back to 1985. The 10Base-T used a star topology in which each station was connected to a central hub (hub). This implementation eliminated the need to interrupt the network during the connection of new stations and made it possible to localize the search for wire breaks to one hub-station line. Manufacturers have the ability to integrate network monitoring and management tools into hubs.

1983: IEEE approves Ethernet 802.3 standard

After the release of the first products, in June 1983, IEEE approved the Ethernet 802.3 and Ethernet 10Base5 standards. As a transmission medium, a "thick" coaxial cable was provided, and each node of the network was connected using a separate transceiver. This implementation turned out to be expensive. A cheap alternative using less expensive and thinner coaxial cable is 10Base2 or ThinNet. The stations no longer required separate transceivers to connect to the cable. In this configuration, Ehternet began a victorious march through the vastness of the ex-USSR. Its main advantages were the ease of deployment and the minimum amount of active network equipment. The shortcomings were immediately determined.

At the time of connecting new stations, the entire network had to be stopped. To disable the network, it was enough to break the cable in one place, so the operation of the cable system required technical personnel to show applied heroism.

1981: 3Com unveils 10Mbit/s Ethernet transceiver and Ethernet adapter for PCs

In March 1981, 3Com introduced a 10Mbit/s Ethernet transceiver, and in September 1982, the first Ethernet adapter for PCs.

1980: Forming Group 802 at IEEE to work on the standard

In February 1980, the results of DIX were presented to IEEE, where the 802 team was soon formed to work on the project. Ethernet secured its position as a standard. For the successful introduction of the technology, the next steps of the "parents" of Ethernet to interact with other manufacturers of chips and hardware were important - for example, the Digital development group presented the Ethernet chip and the source code of its software to Advanced Micro Devices (AMD) and Mostek. As a result, other companies were able to produce compatible Ethernet chipsets, which affected the quality of the hardware and reduced its cost.

1979: One of the developers creates 3Com

The key figure in the fate of Ethernet is Robert Metcalf, who in 1979 created his own company 3Com to implement his ideas, while starting working as a consultant at Digital Equipment Corporation (DEC). At DEC, Metcalf receives a task to develop a network whose specifications would not affect Xerox patents. A joint project is being created between Digital, Intel and Xerox, known as DIX. The task of the DIX consortium was to transfer Ethernet from a laboratory-experimental state to a technology for building new systems operating at a considerable data transfer rate of 10 Mbps at that time. Thus, Ethernet turned from the development of Xerox into an open and accessible technology, which turned out to be decisive in its formation as a world network standard.

1976: Theoretical basis for the development of technology

In July 1976, Metcalf and Boggs released the joint work Ethernet: Distributed Packet Switching for Local Computer Networks ("Ethernet: Distributed Packet Switching for Local Computer Networks"). Thus, a theoretical basis was created for the further development of technology.

1973: The Birth of Technology in the Xerox Lab

Ethernet's birthday can be considered May 22, 1973, when Robert Metcalfe and David Boggs published a memo describing the experimental network they built at the Xerox research center in Palo Alto. At birth, the network was named Ethernet, based on a thick coaxial cable and provided a data transfer rate of 2.94 Mbps.

In December of that year, Metcalf published his doctoral work "Packet Communication communication."

See also

• Economic downturn proves carrier-class Ethernet market stability
• Wireless Hardware Design and Deployment Guide
• Wi-Fi (Wireless Fidelity) - 802.11 wireless standard
• Fibre Channel over Ethernet (FCoE)
• Power-over-Ethernet