Site icon Tom Merritt .com

About Ethernet

KALM-150x150"

Computers talk to each other via networking, most commonly over a wired ethernet connection. Tom explains how that started and why it was such a breakthrough.

Featuring Tom Merritt.

MP3

Please SUBSCRIBE HERE.

A special thanks to all our supporters–without you, none of this would be possible. Become a supporter at the Know A Little More Patreon page for exclusive content and ad-free episodes.

Thanks to Kevin MacLeod of Incompetech.com for the theme music.

Thanks to Garrett Weinzierl for the logo!

Thanks to our mods, Kylde, Jack_Shid, KAPT_Kipper, and scottierowland on the subreddit

Send us email to feedback@dailytechnewsshow.com

Episode transcript:

To describe Robert Metcalfe’s life might sound stereotypical for an American who grew up in the 1950s. He was born in 1946 in Brooklyn. His Mom was a homemaker and his dad was a test technician working in gyroscopes. He graduated from Bay Shore High School in 1964 and got bachelors in electrical engineering and industrial management from MIT in 1969. He got his Masters in applied mathematics from Harvard in 1970 and was cruising to a Harvard Ph.D. in computer science. A true American success story.

Until he wanted to connect to the brand new ARPANet node at Harvard. Metcalfe didn’t take no for an answer. He made his own fate. He used his contacts at MIT and got a job at the school’s Project MAC, which would eventually become the Computer Science and Artificial INtelligence Lab. He worked on the hardware that linked MIT’s minicomputers to the ARPANet.

It seemed like that was his thing. He proposed APRPANet as a topic for his thesis, but Harvard rejected it. While he tried to figure out how to revise the thesis, he got a job out in California at Xerox PARC. While there he read about the ingenious way scientists in Hawaii had figured out how to connect computers between the islands. It was called ALOHANet. He helped them fix some of the bugs in their connections and folded that into his thesis. This time Harvard accepted it.

It definitely seemed like connecting computers was his thing. And if you’ve ever plugged an ethernet cable in to get the internet, you’re glad it was.

Let’s help you know a little more, about ethernet.

Ethernet is a computer networking technology that lets you send data from one machine to another. Typically it’s used for internet traffic, but it can also be used for local area networks, or LANs.

While we think of ethernet as the wires and the funky plug with the little clip, the form factor of the wire isn’t the truly important part. It’s the protocol that makes it ethernet. I could conceivably plug any kind of wire between two computers but you need rules but it would be chaos without rules about which computer is sending and which one is receiving data. If both try to send at the same time you get collisions. Heck even when you do have rules you get collisions. The protocol works because it has good rules not only to try to avoid collisions but how to handle them when they happen anyway.

Ethernet is a very good set of rules.

To use ethernet, systems will divide data into short pieces called frames. Each frame contains info on where it originated, where its end destination is and some error-checking data to determine if the data has been corrupted. If it has, the frame will be discarded and in many instances can be retransmitted.

Think about it this way. When we have a conversation, we don’t both usually talk at once. You talk and then I talk. If we both start talking at once, we stop and one of us lets the other repeat what they said. The ethernet protocol does that too. It basically says for machines to wait until they don’t detect any incoming data before they start sending data. And yes there are rules to stop one machine from dominating the conversation. You have to take breaks every so often. The protocol also has rules for what happens when an accident causes two pieces of data to collide and get corrupted. Sometimes that means ignoring them, sometimes that means asking for them to be re-sent. It depends how important the data is. There are also origination points and endpoints. That’s important when data goes through an ethernet switch, so it knows where to send which frames of data.

Let’s get back to our friend Robert Metcalfe. As we mentioned, while at Xerox PARC, Metcalfe read up on Hawaii’s ALOHANet, which solved the problem of inter-island communication in Hawaii. Starting in June 1971, ALOHANet worked by dividing outbound and inbound traffic into separate channels. Outbound traffic just broadcast everything as packets with each packet containing a destination. Packets were rebroadcast until it was acknowledged that they had been received. We have a whole episode on ALOHANet.

Metcalfe fixed a few bugs particularly about collisions in ALOHANet and included that work in his Harvard PhD Thesis. He used that knowledge at Xerox PARC when he and David Boggs started to develop a way of connecting computers by wire over short distances. At least short compared to islands separated by large distances of water.

Metcalfe was trying to make it easier for Xerox’s Alto computers to communicate. At one point he even called ethernet the Alto Aloha Network since he was using a lot of the methods from ALOHANet.

But the name ethernet was also there from the start. His ideas were first documented on May 22, 1973 in a memo called Alto Ethernet. He likened the ever present stream of data flowing through the wires to the “luminiferous Aether” once believed to be the way electromagnetic waves moved through space.

Ethernet itself first functioned on November 11, 1973.

Xerox filed for a patent on ethernet in 1975, listing Metcalfe, Boggs, Alto designer Chuck Thacker and distributed computing pioneer Butler Lampson as its inventors. Metcalfe and Boggs also published a paper on ethernet in 1976 that was used as the definitive reference work for it.

Metcalfe left Xerox in 1979 to start 3Com but he was able to convince DEC, Intel and Xerox to work together to make ethernet a standard so everyone could use it.

In February 1980, The Institute of Electrical and Electronics Engineers, IEEE, started project 802 to standardize Local Area Networks. Ethernet, along with Token Ring and Token Bus were the most popular of the different ways you could connect computers to each other in a network. The 802 group divided into three subgroups to develop standardization for each of the protocols. The group working on ethernet had the reps from DEC, Intel and Xerox and were called the “DIX” group. They called their standard the Carrier Sense Multiple Access/Collision Detection protocol or CSMA/CD. If you see that somewhere you can assume that means ethernet.

Ethernet originally worked at 2.94 Mbits/s. To make it more commercially acceptable, Yogen Dalal, Ron Crane, Bob Garner and Roy Ogus adapted the protocol to work at 10 Mbit/s.

The first standard at 10 Mbit/s was published September 30, 1980.

The first commercialization also came in 1980 but it was in March of the next year that Metcalfe’s 3Com shipped its first 10 Mbit/s Ethernet 3C100 NIC.

Ethernet becoming a standard would be essential to 3COM, so Metcalfe and David Liddle of Xerox Office Systems, supported Siemens efforts for an alliance in office communications and Siemens helped push for standardization of Ethernet ahead of Token Ring and Token Bus.

Siemen’s Ingrid Fromm, a representative on the 802 group established a Task Group on local networks in Europe’s standards body ECMA TC24. By March 1982, ECMA TC24 agreed on a standard for ethernet based on the IEEE 802 draft. A second version of the DIX group’s works with some tweaks was published in November 1982, known as Ethernet II.

The decisiveness of the ECMA combined with the technical completeness of the DIX group’s work consolidated opinions around ethernet which was approved in December of 1982. IEEE 802.3 was published as a draft in 1983 and ratified as a standard in 1985. International adoption came with the adoption of ISO 8802-3 in 1989.

The rest is upgrades. 100 Mbps ethernet came along in 1995 and Gigabit in 1999. 10-Gigabit ethernet arrived in 2002, and 40 and 100 Gps in 2010. Research continues but 40 is the best you can get for your home because you can’t get any devices for your home that can take advantage of more than that.

So how does it work?

Let’s say you’re sending an email. The data that makes up the email would be broken into packets. Each packet will have three main parts. The bulk of it is the data, the bit of your email being sent. The beginning of the frame is a header which has the Media Access Control address or MAC address of the machine it’s being sent from, the MAC address of where it’s going as well as some other info like IP addresses encryption etc. And the end of the Frame is a frame check value which is used to detect if the frame has been corrupted during transit. That packet is then encapsulated in a. Frame which has information to synchronize the sender and the receiver. This makes sure that even though the frames may arrive at of order everybody knows where the frame fits in the full message.

Once the Frame has been created, the transmitting station waits for the channel to go idle and then sends the frame.

At the other end the receiving station will check to make sure the frame has not been corrupted. If it has it will signal a collision so that the originating transmitter can re-send it. If it’s not corrupted it moves the data into the layer where the frames can be reassembled into the full data, in this case your email.

I skipped over cables earlier because they really are at the service of the protocol. But let’s talk about the cables available now and what they mean.

All ethernet cable use an RJ45 connector. It’s a bit of misnomer since it refers to the jack not the plug, and the jack has a key that twisted pair ethernet cables do not. But nobody really makes a big deal of that anymore. If you must know, ethernet actually uses a modular 8P8C interface, but try asking for that at the computer store and see how far it gets you.

The wire itself as I just mentioned is a twisted pair of wires with a plastic covering. Most ethernet cables are unshielded, known as UTP for unshielded twisted pair, making them more flexible. If you have problems with interference in your area for some reason, more likely in industrial situations than home use, you can get STP or FTP which stand for shielded twisted pair or foiled twisted pair. This has the tradeoff of making the cables thicker and less flexible.

There are 7 widely available types of ethernet cables and they mostly look alike except for the stamp that tells you which kind they are.

The different types can handle different data rates aka speed and bandwidth aka amount of data. I’m just going to give the speed numbers below since bandwidth numbers are less commonly used and won’t mean much to most people.

Cat 5 is the most common and can handle up to 100 Mbps. Cat 5e can do gigabit. Cat 6 which also does gigabit has more bandwidth than 5e. 6a can do 10 gig. Cat 7 does 10 gig with more bandwidth. Cat 7a does 40 Gig. And Cat 8 does up to 40 Gbps with more bandwidth.

If your ethernet cable gets damaged it may not stop working but instead get detected as a lower category. A pinch in a gigabit cat 6 cable might show up as a cat 5 cable.

Then there’s the ports. You may have an ethernet port on your PC or an adapter that gives you one on a laptop. There’s also ethernet ports on routers. And if you want to extend the range of Ethernet, or just need more ports than your routers, you’ll need an ethernet switch. The ports are rated by data rate. You’ll get the lowest of the speeds among all the ethernet ports in your chain. Those ports can be 1 Gbps, 2.5 or 10.

Ethernet is generally more stable, more secure and less prone to interference, even when unshielded, compared to WiFi. You may hear people refer to it as a “hard line” or wired connection. WiFi is pretty fast and reliable these days, but if you have any questions about the reliability of the connection you’ll want to go with ethernet.

Metcalfe received the ACM Grace Hopper award in 1980 for his work on ethernet. He stayed with 3COM until 1990 when the board passed him over for CEO and he left. He spent the next ten years writing a column for InfoWorld.

In 1995 he predicted the internet would collapse within a year or he would eat his words. The next year he blended a page with his prediction and swallowed it. He tried to get them to let him eat a cake with the words frosted on it. To no avail.

He eventually became a venture capitalist and teacher at UT Austin and in June 2022 returned to Project MAC at MIT now known as the Computer Science and Artificial Intelligence Laboratory as a research affiliate and computational engineer, working with the MIT Julia Lab.

So now you know where the inventor of ethernet is. And hope you know a little more about ethernet.

Exit mobile version