Chapter 1. The Evolution of Ethernet

Ethernet is used to build networks from the smallest to the largest, and from the simplest to the most complex: it connects home computers and other household devices, but it also connects the building networks that support servers and wired desktop computers, as well as the wireless access points that support smartphones, laptops, and tablets. Ethernet provides the connections that make up the worldwide Internet and that connect the Internet to our workplaces and our homes.

Ethernet’s longevity is remarkable. The memo describing the network technology that became Ethernet was written in May 1973. There have been many changes as computers have evolved over the years, but Ethernet continues to be the network technology of choice. This is because Ethernet has been constantly reinvented, evolving new capabilities to stay current with the rapid transformations in the computer industry and, in the process, becoming the most widely used network technology in the world.

History of Ethernet

On May 22, 1973, while working at the Xerox Palo Alto Research Center (PARC) in California, Bob Metcalfe wrote a memo describing the network system he had invented for interconnecting advanced computer workstations called Xerox Altos, making it possible to send data between them and to high-speed laser printers. The Xerox Alto was the first personal computer workstation with graphical user interfaces and a mouse pointing device. The PARC inventions also included the first laser printers for personal computers and, with the creation of Ethernet, the first high-speed local area network (LAN) technology to link everything together.

This was a remarkable computing environment for the time, since the early 1970s was an era in which computing was dominated by large and expensive mainframe computers. Few places could afford to buy and support mainframes, and few people knew how to use them. The inventions at Xerox PARC helped bring about a revolutionary change in the world of computing.

A major driver of this revolutionary change was the use of Ethernet LANs to enable communication among computers. Combined with the development of the Internet and the Web, this new model of interaction between computers brought a new world of communications technology into existence.

The Aloha Network

Bob Metcalfe’s 1973 Ethernet memo describes a networking system inspired by an earlier experiment in networking called the Aloha network. The Aloha network began at the University of Hawaii in the late 1960s, when Norman Abramson and his colleagues developed a radio network for communication among the Hawaiian Islands. This system was an early experiment in the development of mechanisms for sharing a common communications channel—in this case, a common radio channel.

The Aloha protocol was very simple: an Aloha station could send whenever it liked, and then wait for an acknowledgment. If an acknowledgment wasn’t received within a short amount of time, the station would assume that another station had transmitted simultaneously, causing a collision in which the combined transmissions were garbled so that the receiving station did not hear them and did not return an acknowledgment. Upon detecting a collision, both transmitting stations would choose a random backoff time, and then retransmit their packets with a good probability of success. However, as traffic increased on the Aloha channel, the collision rate would rapidly increase as well.

Abramson calculated that this system, known as pure Aloha, could achieve a maximum channel utilization of about 18%, due to the rapidly increasing rate of collisions under increasing load. Another system, called slotted Aloha, was developed that assigned transmission slots and used a master clock to synchronize transmissions; this increased the maximum utilization of the channel to about 37%. In 2007, Abramson received the IEEE’s Alexander Graham Bell Medal for “contributions to the development of modern data networks through fundamental work in random multiple access."[1]

The Invention of Ethernet

Metcalfe realized that he could improve on the Aloha system of arbitrating access to a shared communications channel. He developed a new system that included a mechanism that detected when a collision occurred (collision detection). The system also included “listen before talk,” in which stations listened for activity (carrier sense) before transmitting, and supported access to a shared channel by multiple stations (multiple access). Put all these components together, and you can see why the original channel access protocol specified for Ethernet is called Carrier Sense Multiple Access with Collision Detection (CSMA/CD). Metcalfe also developed a more sophisticated backoff algorithm, which, in combination with the CSMA/CD protocol, allowed the Ethernet system to function at up to 100% load.

In late 1972, Metcalfe and his Xerox PARC colleagues developed the first experimental “Ethernet” network system to interconnect Xerox Altos to one another, and to servers and laser printers. The signal clock for the experimental interface was derived from the Alto’s system clock, resulting in a data transmission rate on the experimental Ethernet of 2.94 Mb/s.

Metcalfe’s first experimental network was called the Alto Aloha Network. In 1973, Metcalfe changed the name to “Ethernet,” to make it clear that the system could support any computer‚ not just Altos‚ and to point out that his new network mechanisms had evolved well beyond the Aloha system. He chose to base the name on the word “ether” as a way of describing an essential feature of the system: the physical medium (i.e., a cable) carries bits to all stations, much the same way that the old “luminiferous ether” was once thought to propagate electromagnetic waves through space.[2] Thus, Ethernet was born.

In 1976, Metcalfe drew the diagram shown in Figure 1-1, and it was used in his presentation at the National Computer Conference in June of that year. The drawing uses the original terms for describing Ethernet components.[3]

Drawing of the original Ethernet system
Figure 1-1. Drawing of the original Ethernet system

In July 1976, Bob Metcalfe and David Boggs published their landmark paper “Ethernet: Distributed Packet Switching for Local Computer Networks.”[4] In late 1977, Robert M. Metcalfe, David R. Boggs, Charles P. Thacker, and Butler W. Lampson received U.S. patent number 4,063,220 on Ethernet for a “Multipoint Data Communication System with Collision Detection.”

At this point, Xerox wholly owned the Ethernet system. The next stage in the evolution of the world’s most popular computer network was to liberate Ethernet from the confines of a single corporation and make it a worldwide standard.

Reinventing Ethernet

No matter how well designed a network system is, it won’t help you much if you can only use it with a single vendor’s equipment. A network technology has to be supported by the widest variety of equipment possible to provide you with the greatest flexibility. For maximum utility, your network must be vendor-neutral (i.e., capable of interworking with all types of computers and other devices without being vendor-specific). This was not the way things worked in the 1970s, when computers were expensive and networking technology was exotic and proprietary.

Bob Metcalfe understood that a revolution in computer communications required a networking technology that everyone could use. In 1979, he set out to make Ethernet an open standard, and Xerox agreed to join a multivendor consortium for the purposes of standardizing an Ethernet system that any company could use. The era of open computer communications based on Ethernet technology formally began in 1980 when the Digital Equipment Corporation (DEC), Intel, and Xerox (DIX) consortium announced the first standard for 10 Mb/s Ethernet. The original DIX standard was not copyrighted, allowing anyone to copy and use it.

This standard made the technology available to anyone who wanted to use it, producing an open system. As part of this effort, Xerox agreed to license its patented Ethernet technology for a mere $1,000 to anyone who wanted it. In 1982, Xerox also gave up its trademark on the Ethernet name. As a result, the Ethernet standard became the world’s first open, multivendor LAN standard.

The idea of sharing proprietary computer technology in order to arrive at a common standard to benefit everyone was a radical notion for the computer industry in the late 1970s. It’s a tribute to Bob Metcalfe’s vision that he realized the importance of making Ethernet an open standard. As Metcalfe put it: “The invention of Ethernet as an open, non-proprietary, industry-standard local network was perhaps even more significant than the invention of Ethernet technology itself.”[5]

In 1979, Metcalfe started a company to help commercialize Ethernet. He believed that computers from multiple vendors ought to be able to communicate compatibly over a common networking technology, making them more useful and, in turn, opening up a vast new set of capabilities for the users. Computer communication compatibility was the goal, leading Metcalfe to name his new company 3Com.

Reinventing Ethernet for Twisted-Pair Media

Ethernet prospered during the 1980s, but as the number of computers being networked continued to grow, the problems inherent in the original coaxial cable media system became more acute. Installing coaxial cables in buildings was a difficult task, and connecting computers to the cables was also a challenge.

A thin coaxial cable system was introduced in the mid-1980s that made it a little easier to build a media system and connect computers to it, but it was still difficult to manage Ethernet systems based on coaxial cable. Coaxial Ethernet systems employ a bus topology, in which every computer sends Ethernet signals over a single bus cable; a failure anywhere on the cable brings the entire network system to a halt, and troubleshooting a cable problem can take a long time.

The invention of twisted-pair Ethernet in the late 1980s, initially developed as a vendor innovation, made it possible to build Ethernet systems based on the much more reliable star-wired cabling topology, in which the computers are all connected to a central point.[6] These systems are much easier to install and manage, and troubleshooting is much easier and quicker as well. The use of twisted-pair cabling was a major change, or reinvention, of Ethernet. Twisted-pair Ethernet led to a vast expansion in the use of Ethernet; the Ethernet market took off and has never looked back.

In the early 1990s, a structured cabling system standard for twisted-pair cabling systems in buildings was developed that made it possible to provide building-wide twisted-pair systems based on high-reliability, low-cost cabling adopted from the telephone industry. Ethernet based on twisted-pair media installed according to the structured cabling standard became the most widely used network technology. These Ethernet systems are reliable, are easy to install and manage, and support rapid troubleshooting for problem resolution.

Reinventing Ethernet for 100 Mb/s

The original Ethernet standard of 1980 described a system that operated at 10 Mb/s. This was quite fast for the time, but Ethernet interfaces in the early 1980s were expensive, due to the buffer memory and high-speed components required. Throughout the 1980s, Ethernet was considerably faster than the computers connected to it, making a good match between the network and the computers it supported. However, computer technology continued to evolve, and by the early 1990s ordinary computers had become fast enough to provide a major traffic load to a 10 Mb/s Ethernet channel.

Much to the surprise of those who thought that the original CSMA/CD-based Ethernet system was limited to 10 Mb/s, Ethernet was reinvented to increase its speed by a factor of 10. Based on technology developed by Grand Junction Networks (later acquired by Cisco Systems), the new standard created the 100 Mb/s Fast Ethernet system, which was formally adopted in 1995. Fast Ethernet provides both twisted-pair and fiber optic media systems, and it became widely adopted, first for network backbones and later for general computing.

With the invention of Fast Ethernet, multispeed twisted-pair Ethernet interfaces could be built, operating at either 10 or 100 Mb/s. These interfaces are able, through an Auto-Negotiation protocol, to automatically set their speed. This made the migration from 10 Mb/s to 100 Mb/s Ethernet systems easy to accomplish.

Reinventing Ethernet for 1000 Mb/s

In 1998, Ethernet was reinvented again, this time to increase its speed by another factor of 10. The Gigabit Ethernet standard describes a system that operates at the speed of 1 billion bits per second over fiber optic and twisted-pair media. The invention of Gigabit Ethernet made it possible to provide faster backbone networks as well as connections to high-performance servers.

The twisted-pair standard for Gigabit Ethernet provides high-speed connections to the desktop when needed. Multispeed twisted-pair Ethernet interfaces were built to operate at 10, 100, or 1000 Mb/s, using the Auto-Negotiation protocol to automatically configure their speed.

Reinventing Ethernet for 10, 40, and 100 Gb/s

Not content to rest on its laurels, Ethernet has continued to expand beyond the original design constraints. Although it’s not possible to support the original CSMA/CD shared-channel mode of operation at these higher speeds, that doesn’t matter: virtually all Ethernet connections now operate in full-duplex mode, which does not rely on the CSMA/CD access control system.

The 10 Gb/s Ethernet standard, published in 2003, defined a set of fiber optic media systems operating at 10 billion bits per second. A twisted-pair 10 Gb/s standard was developed and published in 2006, providing 10 billion bits per second over Category 6A twisted-pair cables. Multispeed twisted-pair Ethernet interfaces can now operate at 10, 100, and 1000 Mb/s, and 10 Gb/s.

The 40 and 100 Gb/s Ethernet standard, which was published in 2010, defined both 40 and 100 Gb/s media systems. Since then, media systems have been evolving to carry 40 and 100 Gb/s Ethernet signals over fiber optic cables and short-range copper coaxial cables.

Reinventing Ethernet for New Capabilities

Ethernet innovations include not only new speeds and new media systems, but also new Ethernet capabilities. For example, the standardization of full-duplex Ethernet in 1997 made it possible for two devices connected over a full-duplex link to simultaneously send and receive data, thus allowing a 10 Gb/s link to provide a maximum of 20 Gb/s of data throughput.

The Auto-Negotiation standard complements the invention of twisted-pair Ethernet by providing the ability for switch ports and the computers connected to those ports to discover whether they support full-duplex mode and, if they do, to automatically select that mode of operation as well as automatically setting the highest link speed supported by both devices.

Another innovation has been the Power over Ethernet (PoE) standard, which uses the Ethernet cable that is providing data to also power the device connected to an Ethernet switch. This has become a widely adopted method for deploying wireless access points connected to Ethernet switch ports and drawing their power from the same cable that they use to send and receive Ethernet frames.

Ethernet Switches

The invention of full-duplex twisted-pair and fiber optic Ethernet coincided with the development of network switches, allowing network managers to build large networks based on switches and full-duplex links. Switches have Ethernet interfaces (ports), but the operation of switch protocols is not part of the Ethernet standard. Instead, the operation of switches is specified in the IEEE 802.1 series of standards, with the 802.1D standard providing the specifications for basic switches.

You can build a wide variety of networks with switches. There are switches designed for campus and enterprise networks, switches with special capabilities for data centers, switches that support carrier and long distance networks, and more.

Network design based on switches is a big topic with its own literature, based on the type of network being developed. There are books on campus and enterprise network design, as well as books on data center networks. This is a book on Ethernet standards and technology, and we don’t have the space to provide an in-depth treatment of the 802.1 switch standards and the topic of network design with switches for multiple network types. However, Part IV, including Chapters 18 and 19, provides an introduction to switch operation and a discussion of how switches can be used in network designs.

The Future of Ethernet

Ethernet has come a long way since 10 Mb/s Ethernet became the world’s first open standard for computer networking in the early 1980s. As you can see, the Ethernet system has been reinvented to provide more flexible and reliable cabling, to accommodate the rapid increase in network traffic with higher speeds, and to provide more capabilities for today’s more complex network systems.

Ethernet has been able to meet these challenges while maintaining the same basic structure and operation, and doing it all at a reasonable cost. This fundamental stability, combined with the ability to evolve to meet new needs, is at the core of Ethernet’s success.

[1] The IEEE Global History Network biography of Norman Abramson states: “While at the University of Hawaii, he led efforts that gave rise to the construction and operation of the ALOHAnet, the first wireless packet network, and to the development of the theory of random access ALOHA channels. ALOHA channels have yielded significant advancements within wireless and local area networking, with versions still in use today in all major mobile telephone and wireless data standards. This influential work also developed the core concepts found today in Ethernet.”

[2] Physicists Albert Michelson and Edward Morley disproved the existence of the ether in 1887, but Metcalfe decided that it was a good name for his new network system that carried signals to all computers.

[3] From The Ethernet Sourcebook, ed. Robyn E. Shotwell (New York: North-Holland, 1985), title page. Diagram reproduced with permission.

[4] Communications of the ACM, 19:7 (July 1976): 395–404.

[5] Shotwell, The Ethernet Sourcebook, p. xi.

[6] The vendor was SynOptics Communications, whose LattisNet was the first twisted-pair product.

Get Ethernet: The Definitive Guide, 2nd Edition now with O’Reilly online learning.

O’Reilly members experience live online training, plus books, videos, and digital content from 200+ publishers.