EXPANDING THE ETHER: The invention of the network bridge

[@alankirby7580]

This is a great episode! Thanks for helping document this piece of Ethernet history.

[@klzgh]

the man, legend

[@donwald3436]

Damn imagine inventing STP over a weekend and writing it up in another two days..... she's one hell of an engineer!

[@gorillaau]

It's also pretty amazing that a proposed standard can be given to someone else who didn't need further clarification on how it should work. Nothing left out, nothing contradicting some other part.

[@Trenjeska]

And then make it into poetry.

[@RobertOrton]

She also did not have a computer constantly nagging her about updates, offers, spell checks, stupid word suggestions, annoying posts and goddamn advertising. She used tools that let her think, concentrate and express her creativity. We have lost a lot of that with our modern technology’s monetisation and enshitification of everything. 2 generations of engineering excellence like hers has been fucked by the likes of Jobs, Gates, Zuc, Bezo, Musk and the conga line of wannabe assholes they inspire.

[@nohphd]

If there was a Nobel prize for networking, Radia Perlman would probably have two.

[@dus10dnd]

@@nohphd Let's nominate her! I have the first and second editions of the books on the way to my house, now.

[@Torbjorn.Lindgren]

Side-note: The "vampire taps" were actually a big IMPROVEMENT over the original method of connecting thicknet - vampire taps were much easier to do and could be done while the network was live! Really early on it was done using techniques similar to thinnet, IE cut the cable completely at each connected machine - using the equipment he showed for putting on connectors and the restore the network by plugging in both into the AUI transceiver (either directly or via a T junction like thinnet). This is also why the terminator was usually added as a separate entity instead of being crimped on directly (would be faster), having a end-connector meant it could be extended by just adding a new section.

[@LogicalNiko]

Yes vampire taps were such a blessing. Otherwise every time you connected a system you generally took a segment of the entire network offline. And thicknet quickly moved into office spaces before thinnet ended up coming out.

[@leosmith848]

We called them bee stings

[@imjustabill247]

The details on tapping and terminating the thicknet from John was awesome. Keep up the outstanding history work on this channel!!!

[@PeterWillard]

Horror story... when non-network people would just *screw in the TAP* without drilling the hole. Network Down! Network Down!

[@johnreedCaroline]

​​​@@PeterWillardHi Pete, I remember the SANS server you managed for us. That was a busy time...

[@PeterWillard]

@@johnreedCaroline It was... I went on to manage DNS services for the southeast after NETsupport was killed. I left HP in 2015 at the time of the great split.

[@nickwallette6201]

@@PeterWillard I often wondered how these vampire taps I'd always heard about avoided bridging the shield and the conductor. This was the first time I've actually seen the process, ever. Sounds like the process is not too dissimilar to the proper way to pierce an ear, for example: Your goal is a clean cut around a perimeter, not just punching a hole through.

[@PeterWillard]

@@nickwallette6201 Once you understood the topology of the baseband cable, you would know that simply "screwing the pin" into the cable would be a bad idea. The cable had a lot of shielding.

[@ptefar]

A small nit: store and forward does not refer to storing the Mac address, that would be a "learning bridge". Store and forward is in contrast to cut-through bridges that start forwarding as soon as the destination mac address is received. Store-and-forward receives the whole packet first and then tries to transmit it, when it. This allows garbled packets to be discarded and not take up time on the receiving network segment.

[@phil7084]

All bridges (afaik) have to be store and forward, as directly forwarding a packet on the destination network would result in A. joining the collision domains together (i.e. if the destination segment is busy, you would have to signal a collision on the source segment to inform the sender the packet was not transmitted successfully), or B. dropping packets silently when the destination segment was busy. As neither of these was desirable, some memory was allocated to storage of packets for forwarding, usually with a staleness timer, to allow packets that weren't successfully forwarded to eventually be dropped in order to free up the memory buffer. From experience, as the memory buffer was small it could cause random drops between the segments if one was busy, which was always an interesting challenge to diagnose (as 100 Mbps networking was being introduced and the original 10 Mbps bridges were getting older and less reliable)

[@ptefar]

@@phil7084 Please see e.g. en.wikipedia.org/wiki/Cut-through_switching for a discussion about cut through switching. It was very common for a while before wire speeds were such that a single packet delay was less of an issue. If the outbound port was busy the packets were buffered, waiting for it to go idle. But the cut-through bridges would start early, if possible. The major downsides were: an inability to deal with differing wire speeds (not an isue at 10baseX as that was the only speed) and forwarding packets with broken FCS or otherwise corrupted wasting time on the outbound segment.

[@phil7084]

@@ptefar I'm aware of cut through switching - but it was introduced for ethernet switching, using custom silicon. The early bridges used shared memory between the network adaptors.

[@ptefar]

@@phil7084 I agree. I remember cut thorugh bridges starting to appear in the late 1980s. My point in the original comment was that I have never heard store-and-forward being used to refer to a self learning bridge like what whas mentioned in the video (as in storing the mac address). It was always used as a contrast to cut-through. I never menat to imply that early bridges were cut-thorugh, they were indeed not.

[@nohphd]

@@phil7084Kalpana switches have entered the conversation…

[@checksum00]

I'm the CTO at a datacenter. I had no idea how Thicknet worked physically. Not so many people have that knowledge still today. I was very well aware of vampire taps, AUI, etc. but never saw in done before.

[@Darxide23]

I may be among a small group of the youngest people to still know how Thicknet and vampire taps work and I'm confident that I could get a network up and going today with just a little bit of a refresher. I just turned 43 and the school I went to circa 1999/2000 still taught some of the older tech specifically for students to get into a legacy IT job. I even learned COBOL. Problem is, people who have those jobs just never retire and job positions have wait lists of candidates who will die of old age before their number gets called. When I went back to school at a 4 year university in 2013 for network architecture and design to bring myself up to speed (I'd been out of the networking game, focused more on the consumer side of things for the past decade) this stuff was nowhere to be seen. Every course was modern and modern only. No more legacy IT courses of any sort.

[@PeterWillard]

What a great video, it brought back a lot of fond memories. Bridges were a godsend as we started to create larger facility networks. I originally worked for Digital's internal facility networking group in New England and was one of the people responsible for installing this new Ethernet Baseband cabling in a number of Digital's cable plants. I later moved to the Southern Area USA, going to work for Customer Service, and ended up working in John Reed's group. While I was also a cable plant designer for multiple customers, it was John Reed who was our resident expert, particularly with broadband and baseband networks . I was a great pleasure to have worked with John.

[@kote315]

Oh wow! I studied networking and I remember being told about thick coaxial cable and "vampire" connectors. But since this was in the late 2000s, the technology was already very outdated and no one had seen it in real life. Thanks to your video, I now not only know what it looked like, but also how to connect such a cable. And I'm glad I will never have to do that.😄

[@theserialport]

Hahah we thought exactly the same... really neat to see, but glad that's in the past

[@DangerousPictures]

I went to school until 2018 and we still learned about coax and vampire connectors. On the other hand CIDR was never mentioned and we still learned about classes

[@James_Knott]

@@DangerousPictures That's one thing that really rubs me the wrong way. Address classes have been obsolete for over 30 years. However, originally there were no classes. Every address was what eventually become class A, with 8 bit network address and 24 bits for the device. These days everyone should be moving to IPv6. I've had it on my home network for over 14 years.

[@autohmae]

@@DangerousPictures that's just sad.

[@c1ph3rpunk]

One of the first networks I worked on in the late 80’s was all DEC, the building “backbone” was all thicknet. Vampire taps every $X feet that dropped down to terminal servers that VT320’s connected to. That cabling was horrid.

[@MN-Hillbilly]

Exceptionally researched video. Subscription well earned. As someone who has been in the game since 1990 I cringe when I watch a video with factual inaccuracies and I didn't notice a single one. Well done.

[@James_Knott]

1990? You're a newcomer! I first started working with LANs in early 1978, on the Air Canada reservation system. They used proprietary tech from Rockwell Collins, which used Time Division Multiplexing (TDM), instead of packets, over a loop. There were 2 speeds, 2 Mb and 8. I first came across Ethernet in the 1980, connecting some VAX 11/780s with 10base5. In 1989, I hand wired some Ethernet controllers for Data General Eclipse computers, so they could be connected to the network.

[@MN-Hillbilly]

@@James_Knott I started my professional life in the military in Panama interfacing communication equipment that wasn't designed to work together. We used what could be thought of as exotic to make it happen. It was an interesting time with all of the competing technologies before everything shook out to standardize on Ethernet. I think it was around 1993 we received our magical "packet switch" and how we could interface it with our teletype circuits. Interestingly to this day I service timing and scoring equipment in sport stadiums that still use current loop technology.

[@benjamin3044]

I legit just learned about plenum vs non-plenum cabling while studying for my CISSP. What timing!

[@yannickberrios]

An outstanding video about the roots of ethernet, impressive research and narrative. Greetings from Bolivia!

[@theserialport]

Thank you and bienvenidos!

[@BrynnHarrison66]

I started with DEC in 1988 as a Firmware Engineer in the Wide Area Communications Group in Reading UK. I worked on the firmware for the line cards in the DECnis 500 & 600. Great memories. I was there (DEC/Compaq/HP/HPE) for over 32 years. This for me isn't history it the journal of my career :-) Great Job Guys.

[@johnreedCaroline]

DECnis. I loved the OSI command line, separating characteristics from status...

[@majoryoshi]

As someone going for their bachelor's in networking, tutors networking topics, heck even installs networking infrastructure for some clients, I never knew just where it came from, only vaguely hearing about concepts like Store-and-Forward switching and STP. Having only ever seen RJ45 and fiber in my life, I never knew just how thick Token Ring was. Massive props!

[@gloriousstereo]

Token Ring is a different standard than the "thicknet" shown here.

[@nohphd]

I’m an old dog, started out with Novel 2.12 and IPX. Physical layer was Corvus (Like AppleTalk but using IPX) and later StarLAN1 and StarLAN 10, both 10BaseT precursors. Did a lot of 10Base2, 10Base5 and (the totally obscure) 10Base36. I like to joke that my troubleshooting 10BaseX put my kids through college in the late 1990s and early 2000s. I’d like to compliment your presentation. As somebody who lived through all this, I didn’t see a single error or misunderstanding, which is exceedingly rare for KZbin presentations. My hat’s off to the very fine job you did.

[@nickwallette6201]

The ending really illustrates the stark contrast in philosophy that early computing pioneers had: It's better for a product to gain traction than to be the sole benefactor of its success. The open nature of these protocols is what made networking not only ubiquitous today, but I would argue, to make it successful at all. Everyone now is so focused on market share and licensing revenue that dominant protocols and technologies burn bright and then burn out, as they're replaced by whatever's the next flash in the pan. Nothing has 40+ year staying power, because every solution is an opportunity to create a source of recurring revenue rather than a technology that will endure. OTOH, many of the pioneers that contributed so much to the open network as we know it have sacrificed themselves to competitors that implemented their ideas better, faster, more conveniently, and at lower cost. They lived as legends and died as martyrs.

[@germancaperarojas4023]

Wow! Man, you break it with this episode. This is one of the best recopilations on the history of the Ethernet standard I've ever seen. Today even engineers say misnomers like "ethernet cable", so your video is a "must see" for anyone studying the field of networking. Keep the good work!

[@MontegaB]

It's awesome that you're doing networking history. It's an often overlooked aspect of computer history that deserves more attention!

[@mndodd]

The thick-net annular ring is marked at the quarter wave interval. Termination needed to be made at that interval as well. 10base2 was ever so much more forgiving.

[@theserialport]

Didn't realize even the terminators were subject to that requirement, but that makes sense.

[@mndodd]

@@theserialport Nice photo of the Kalpana switch Want one?

[@theserialport]

Absolutely

[@colinstu]

@@mndodd Please hook him up with one!

[@James_Knott]

@@theserialport Actually, it doesn't. A terminator appears as an infinitely long cable. I suspect the designers were just being overly cautious.

[@Myfatheredward]

You guys unveil so much historical information, and your delivery makes for easy listening, and learning.

[@ToomsDotDk]

always fun to add the taps on live thick cable, being careful not to short it as that will take the network down

[@wilsonle61]

OMG! I used to install thicknet, bridges, VAX 11/780s for GE Syracuse Military Division. The equipment supported the Over The Horizon Backscatter Radar. I had totally forgotten this. It was a blast to see it again. DEC Equipment was built like a tank, and costed as if god plated!

[@aubreyadams7884]

So want to show my current networking students this video, but all they will see is a 70yo guy's nostalgia! 😃

[@Xsiondu]

49.9k and happy to sub. This is my mind of nerding out. I'm staggered that you just plugged it in and it just got down to business. Astonishing!

[@LaLaLand.Germany]

This was rather thrilling! I belive I saw a bridge like this years ago in a trash heap on the curb. But I couldn´t make heads or tails of it so I left them. Yes, plural… But I took another lan box that had coax ("cheapnet") plugs and a serial port as well as the same led config like the others- that´s why I remember. It too powered up on plug in and it was to talk to over serial. I gave it to a SETI guy, he instantly knew what it was and happily gave me 20 Schekel. But now I see these and something comes back… And they were European versions, ran on 240 Volts- I don´t know how many were sold to outside markets but 8K Dolares was a big oompf to spend. Yeah, thanks, this is brilliant! Happy New Year Y´all!

[@Solar_and_Security]

Just bought both Radia Pearlman's Books, Interconnections, and Private Comms in a public world. I had no idea of STP's history. Great Video. Congratulationson the 50k subs. Thanks!

[@binarydinosaurs]

I started working around the time this was invented, but had no idea about the 5-4-3 rule. We were putting thicknet in industrial premises often using stupid length ladders that you just wouldn't be allowed to do now. Up top with crimps, vampire taps and H4005 transceivers to drop AUI cables down to workstations on the factory floor. Later we discovered the 5-4-3 rule when we tried to link 2 buildings together and it was a mess until DEC told us about the DEREP and (I think ) DEREN repeaters which solved the problems. Plenty of other DEC comms boxes used the 68k cpu too - WANrouters, DECservers etc. Proper workhorse. I still have a pair of DEREPs though obviously they don't have a use any more :)

[@typxxilps]

7:00 hats off -- what a great demonstration of a simple click nowadays when we plug an ethernet cable into the socket. Of cause he showed the cromping and connecting too, but what he was showing were tools now used at home to build a 48 V diy home storage battery, same for the cable cutter and the hydraulic pump. Really great content even though I have never seen that, I was only aware of the outlet where we were used to have token ring which was a mess back then.

[@r000tbeer]

Very cool. We had a Synoptics LattisNet hub at an old job that would cause intermittent problems. We finally realized it was simply the hub itself, and that it was very close to the Ethernet standard but not exactly and tossed it to the curb. No more problems after that. Thankfully I never had to deal with 10base5!

[@deneb_tm]

great video! perfect combination of depth and clarity imo. i know networking, but thicknet had so far eluded me also super neat editing trick while going over the LEDs on the device - just want to make sure that doesn't go unappreciated!

[@TheRetroGamingPrincess]

Computing History AND DEC history into one video? Combining two favourite topics! How can this be! FANTASTIC video, Absolutely fantastic~!

[@dogratco]

Excellent. I began working with Ethernet as soon as the IEEE 802.3 spec was finalized, mostly with troubleshooting. With 10Base5, one of the 50 Ohm terminators needed to be grounded. After revolutionizing inter-processor communication, Ethernet drop cables going to multi-port transceivers became popular for attachment of distributed RS232 terminal servers, such as the DECserver. Plenum-rated AUI cables were too massive for the latch to hold the DB-15 shell in place. Hex nuts were needed in those installations. Richard Siefert came up with the latch system, that worked very well for PVC AUI cables.

[@MrRf81]

I remember when broadband became super popular and every office had an IT expert... I made a lot of money on "slow network" calls where I swapped out hubs for switches to avoid collisions. Almost forgot about that "fun" until I watched your video. Thanks for the memories.

[@djmichaelm3301]

Awesome! I love the history of IT. We’ve come such a long way in a short time and it’s easily forgotten. When I saw the DEC LAN Bridge 100 in the video, it reminded me of “The Box” made by Piped Piper on Silicon Valley. Even more so when you took off the plastic housing. Haha. Thanks for the great content!

[@theserialport]

Thanks for watching!

[@BrandonNedwek]

New Serial Port and clabretro videos on the same day?! 🤓

[@TonathanHawk]

Phenomenal. As someone currently learning network+ this is exactly what I needed to see to actually understand these technologies. Thank you!

[@BrianFitzgerald-x5q]

Worked a lot with the Lanbridge100s. A later innovation was the availability of a fibre optic interface for one or both of the ports (they were modular from the start) to allow networking between building with electrical isolation, a problem with long thickenet runs. The other other end of the connection was a 10base5 Dec Repeater with one fibre interface. They could reach almost 1km used FSMA ports, and were almost indestructable. The Lanbridges were usually in some central location (datacentre) with a rack of repeaters (10base5 fibre plus 10base2 multiports) and a coil of Thicknet in some remote dusty closet.

[@mpeskett]

Super cool video, I love these historical explainers. I think for newcomers to networking these days, it’s very easy to sort of gloss over words like ‘bridge’ without considering what it truly means. I think most would consider a bridge and a switch to be the same device. Generally speaking, I also think a lot of confusion can be cleared up by looking at the historical roots of technologies such as IPv4 and Ethernet, so I really rate your videos! Thanks a lot.

[@MN-Hillbilly]

My thoughts exactly.

[@MatthijsvanDuin]

The distinction between bridge and switch is just an ill-defined marketing thing, the standards consider all such devices to be bridges.

[@Thiesi]

6:13 - I've cut quite a few coax cables for satellite receivers, so that process looked very familiar to me - but boy do these crimping tools mean business!

[@techheart6090]

You realize that some of us pause and read the messages right - thank you, Serial Port... I love this ish.

[@colinstu]

AWESOME VID! Been wanting someone to do something like this for years, happy to finally see it! And excited for the next vid, see you do your own thicknet... then delve into thinnet... CAT3, other old hub history/early switching etc. EXCITE!

[@DuncanGlendinning]

Great video & story. Its been about 37 years since I worked on a product which implemented 802.1D!

[@yurikz9]

you should make a playlist with these networking videos in chronological order! would be interesting to watch tech progress this way

[@andremantovani]

again, you rocking in a deep analysis and a great video so that you cant see time passing by! congrats!

[@YvanHuneault-1]

Fascinating and very well done.

[@BryanSeitz]

Loved the episode! Such nostalgia.

[@airysquared]

I’m learning networking. This is interesting stuff. I feel like I understand bridges way better now.

[@tomschmidt381]

Thanks for the trip down memory lane. I was aware of Ethernet thickwire but never had to deal with it. My introduction to Ethernet was thinwire, along with the battle of IBM Token Ring, and Apollo's Domain networks. In the late 1970s early 1980s I was designing inhouse test equipment and wanted to try Ethernet but as you posted back then the interfaces were very complex and expensive so ended up using current loop. Today due to high levels of silicon integration Ethernet and WiFi interfaces are almost free.

[@zibifranz2429]

Thank you very much for this very helpful and educative document! By the way Mark Kempf invented the first Bridge in 1983, so it took him 3 years to bring it to the market in 1986. I have been recently tracing the history of Ethernet, Bridges and Switches myself as a preparation for CISCO certification, and the history of the first Switch is also fascinating! In 1986 the first Switch was installed by Alantec at NASA, but it was Kalpana that brought it to the market in 1990. The first ATM Layer 3 Switch was announced by Ipsilon Networks in 1996, but the first Ethernet Layer 3 Switch was marketed by Nortel in 1998. Thank you.

[@pokrovsky-production]

Great channel, guys, love every episode!

[@RikMaxSpeed]

Greatly enjoyed this video, both for its very clear explanations and the amazing ingenuity on display at DEC - this is what you get when you have a company run by talented engineers, and sadly not so talented on the sales & marketing side.

[@orbitingeyes2540]

Oh, man! I remember those RG-8 ish thicknet cables, flaky vampire taps, and AUIs! Thankfully those quickly evolved to 10-base-T thinnet with Tee's, and then the twisted-pairs we know and love today.

[@georgegrubbs2966]

Love these videos. I know they must take a huge amount of time and expert;ise to create. Keep up the great work.

[@nasasts51]

I had the pleasure to meet Mitch a few years ago. He has a LOT of DEC parts.

[@foxale08]

Wanted to mention that the current direction in networking is to minimize STP. Some vendors have Multi-Chassis link aggregation implementations which explicitly disable it, others allow paired devices to appear as a single bridge. I've worked on large campus networks which only have STP on edge ports for loop detection purposes, if a BPDU is received the port gets disabled.

[@James_Knott]

There's also shortest path bridging, which is based on IS-IS.

[@foxale08]

​@@James_KnottIndeed. I've used it but it's not as popular as EVPN/VXLAN.

[@nickwallette6201]

Yeah.. about that... I distinctly remember my first gig with a consulting firm where we forklifted an entire enterprise network in one night. They had just bought Juniper EX switches, and the Juniper rep was going on about how we don't need STP anymore because Virtual Chassis. And then one of the cabling contractors plugged two uplinks into a pair of switches that had been stacked. The resulting network storm was epic. I noticed things going wrong in stages. First, remote hosts stopped responding. Then I couldn't ping the management IP of the core switch. Then my mouse pointer started moving at about 4 FPS due to the barrage of interrupts from the NIC. To the Junipers' credit, once we disconnected the looping uplink, every single switch came right back almost immediately. Those things can definitely handle wirespeed throughput! We got through the rest of that install, and the next day at lunch, I saddled up to the director of IT... "So.... can we go ahead and re-enable STP?" "Yep." "Good."

[@foxale08]

@@nickwallette6201 I've never been a fan of the virtual chassis/stacking route. Sharing chunks of an FDB with separate management is fine though.

[@nickwallette6201]

@@foxale08 Implementation often leaves a bit to be desired. While some are better than others, I haven't yet seen a vendor that pulls it off in an intuitive, trouble-free, simple way. And it rarely lives up to the promise of real, chassis-like high availability. But it sure beats having to configure 2 to 10 separate switches!

[@Gillymonsterproductions]

I just shared this to my troops! I love historical information on networking! Also just subbed today!

[@MicrobyteAlan]

Very interest. I'm a former dec field engineer. I started as an in-house field service engineer , working at MR01 1980 - 83. I installed many ethernet back bones and a lot of taping. Great memories , thanks.

[@PeterWillard]

@MicrobyteAlan I'm pretty sure MRO1 was one of the first places I installed "thick wire" Ethernet.

[@MicrobyteAlan]

@@PeterWillard I think it probably was. Didn’t know it at the time. Glad I was there.

[@AlanTheBeast100]

I recall this yellow cable running a loop throughout our division floor covering about 100 engineers, programmers, et al. It was not "in the ceiling" - but ran down the edge of the labs inside (opposite side of the hallway wall). Main reason was to connect us to the VAX-780 (converted to 785) - DEC-220 terminals to some mux device to the ethernet. I had a special card for my PC (Compaq 386 @ 16 MHz) and some sort of "virtual" disk on the VAX. Nifty. There was a bridge to another division in the co. (this evolved a few times before the yellow cable was finally abandoned when we moved to a new building.

[@Thatdavemarsh]

Awesome. I remember doing network and PC support in the early 90s as a summer job pulling out old 3com coax cards (10B2) and replacing with 10BT runs. Every so often I’d find the old adapter modules and rarely a chunk of abandoned Thicknet. This was a great history lesson.

[@James_Knott]

Back in 1997 I was on a project to convert some Province of Ontario offices from 10base2 to 10baseT. I wasn't involved with the cabling as my job was to change the computers to work on 10baseT. On some, it was just a matter of disconnecting the coax and plugging in the new cable. Others required changing the NIC. I recall some Xerox printers that had an AUI interface and so couldn't be switched, without getting the 10baseT/AUI adapter. Otherwise, it would have been necessary to get the 10base2 cards from Xerox.

[@Myfatheredward]

This channel is top tier!!

[@DiegoBarrios]

Awesome explanation of Paleolithic network history!

[@Techno-Universal]

I even remember seeing old standard outlet and light switch wall plates in my old trade college which was built in the early 1980s that had parallel/AUI connectors on them so they probably had old transceiver Ethernet taps in the wall cavities behind them! :)

[@PatrickBaptist]

Congrats your at 50.1k subs, just a day after posting saying you were hoping to get to 50k, you made it!

[@thatLion01]

Awesome video. Thank you. I really enjoyed this and learning about Ethernet

[@funkiam9214]

Brilliant channel!Greetings from EU

[@exponentmantissa5598]

I am a just retired engineer and I worked most of career with computer networks. At a company I worked for we developed the first TCPIP stack for the PC and the first TCPIP stack for DEC Vaxes. I worked with Ethernet from the early days and even some of the alternatives such as ARCNet and Synoptics Lattisnet which was the first computer network stack to run across twisted pair which was a HUGE deal at the time. I did a lot of work with integrating big iron onto PC networks so I have seen pretty well all computer protocols and equipment used from the 80s onwards. If you are ever after information let me know. I know quite a but about the history and am still in touch with my original colleagues.

[@theserialport]

Contact us over at serialport.org, we'd love to chat.

[@stevengill1736]

Pretty amazing stuff! Thank you kindly for the video....

[@teedot]

Love these videos, guys. Keep it up!

[@TeslaTales59]

Excellent bit of history. I was there in those early days and used an HP bridge and even LAN 10Base2 speed analyzers.

[@fishbone470]

Great history lesson.. Thicknet def. has some girth in those cables! lol.. we've def. come along way.. I have a 10g nic and 4G fiber coming right to my desk lol

[@retinaquester]

I am raised with in the office 10Base-T Coax network, but all cables were like TV-cable thinckness. With T-shape Connectors at every ethernetcard. Always funny to have a lose end (50hm) resistor. Took the whole network down, without having a idea where to look. Love the vid, and the old guy who still can do the tab today. Asif he did it yesterday on the job.

[@alexatkin]

That would have been 10BASE2. 10BASE-T refers to Twisted Pair, 10BASE-F being fibre.

[@retinaquester]

@@alexatkin Oops I stand corrected, you are right 10BASE2

[@_Mackan]

Man, that PCB filled to the brim with ICs is beautiful

[@PeterWillard]

Digital Engineers were quite adept at building gorgeous PCB's.

[@idahofur]

Oh, network names I have not heard about. When I first got into computers. Did tons of small very small thinnet (10-base2) installs. Upgrade Arcnet to 10-base2 and one place had Token Ring and also 10-base2. Then came along Twisted pair and never looked back. While going to High School I witnessed a install of a Thick Ethernet cable ran between 2 building to connect 2 dechubs together with misc. modules. Then going to the other 2 schools at each end of a town was connected Frame Relay. That was latter replaced with CIsco Wireless.

[@paulstubbs7678]

I came into all of this at the end of 10base5 thicknet, I have a transceiver - somewhere.... I worked for a telco in their repair shop, we lived on leftover and surplus network gear as that was not our primary job, this was when networking was not a given piece of infrastructure, you only got it if you could put up a good case for justifying it, this (at the time) very expensive kit. However we were all kind of nerds in that repair shop so we put out feelers to see what we could score, so we ended up installed some thicknet, before moving onto the later thinnet. Actually I'm listening to this video via an amplifier I built in an old network bridge box, kind of the same as the DEC with two AUI ports, but half the height. I didn't think I could ever use it, so it was kept as a really nice metal box.

[@JamesTenniswood]

Awesome, love this channel. Keep them coming! ❤

[@SirKenchalot]

Oh what a tease! Now I have to tune in to the next episode to see what happened to he cable.

[@RadioChief52]

Brings back long ago memories working with this technology in the late 1980's. Does anybody remember Ungermann-Bass and the 10Broad36 standard that modulated the data onto cable television channels? It was a pre-cursor to today's cable modems. As I remember, it took two full length PC cards to implement. One card was the RF card and it feed the modem card via an internal jumper. Our plant was wired with 3/4" aluminum coax using standard cable television amplifiers along the way. Exactly the same as you see on utility poles along the road. The cable had taps along the way where you would connect an RG-6 coax with an F connector directly to the RF card in the PC. At the time, we thought it was the cat's meow and the logical replacement to thicknet. A year later, UTP was introduced and the 10Brroad36 was out the door.

[@DECcomputers]

Thanks for that. It's interesting to see the Thick Ethernet installation procedure. I didn't know, that DEC was also the inventor of the network bridge :)

[@michaeldeloatch7461]

Excellent! Thanks. Ironic that I lived through network topology getting thinner, faster and nimbler even as my own personal hardware seems to be trending exactly opposite.

[@erintyres3609]

My workplace had more than a dozen Dec computers connected with "thin wire" Ethernet. A single RG-58 cable snaked under the floor, with a T shaped BNC connector on each computer, and a 50 ohm terminator at each end. If there was a break in the connection at any point, all communication would stop. If any length of coax was not "50 ohm" characteristic impedance, communications would stop. If a coax was too long and was neatly coiled, it acted like a choke and communications would stop. I learned to track down problems by breaking the network in the middle and seeing which half would function. It was a real life binary search, done with giant suction cups to lift the floor tiles. Except for all that, it worked well. It was eventually replaced by an Ethernet hub. The coax cables were replaced by regular Ethernet cables running to the hub, making the network much easier to maintain.

[@michaelhess4825]

I yanked out tons of thicknet in the mid 90's. And did a few taps even. Thinnet was a dream to run into! 😅

[@JohnWallace74]

Great video. Very interesting.

[@micahnightwolf]

It's crazy how much the modern internet and computer networks still depend on protocols that were designed over 40 years ago by a small group of people with cool ideas and enough money and skill to implement them.

[@BestSpatula]

I am VERY curious about why you only saw about 1.8Mb/s. The 3.9 microsecond required to make a forwarding decision should have been more than fast enough to forward at the required 14,880 packets per second rate for the 64 byte length you selected. What am I missing?

[@Richardincancale]

17:53 Good memories! I did lots of this stuff in the early 1980s, including using managed bridges pre-spanning tree days. The throughput of around 1.8 Mbit/s isn’t too bad when you consider that the maximum theoretical throughput of 10Mbit/s Ethernet is around 3.6Mbit/s given the Aloha rules using collision detection (listen before send).

[@RoyAntaw]

Such memories, we had a DEC LAN Bridge connected to our VAX 11/780, running VMS, in the eighties and this system was not decommissioned until the early nineties.

[@DanielleWhite]

I remember that classic DEC solid construction mostly from dealing with their terminal servers, several of which had the plastic clamshell while others just had the inner steel housing and were in racks. I also dealt with a few DELNI and DEMPR devices at that job.

[@knietiefimdispo2458]

I installed a few 10Base5 networks in 1985. One of them at home to interconnect with my brother one floor lower. Soon we repurposed it as antenna cable for ham radio. It was as good as RG213U 🙂

[@sanyr80]

I remember learning about thicknet and vampire taps in a networking class I took 25 years ago. The instructor didn’t explain it very well and had no props to show so what it was and how you deployed it went over my head. This video would have been perfect back then! 😂

[@compu85]

It's so great how ethernet gear from nearly the beginning is still compatible with modern devices.

[@minecraft2048]

Having modern ethernet NICs and Wireshark just works with that bridge is amazing

[@olivier2553]

The Ethernet packet has not evolved so much, older packets are still compatible.

[@nitehawk86]

2:25 Wow that bridge is huge, you could use it as an actual bridge!

[@jmonger]

I've never been so thankful for Cat6.

[@yellowcrescent]

I had always wondered how 10Base5 Thicknet coax was "tapped" -- very cool. I have used and terminated RG8 coax for 10Base2 before -- until recently, it was STILL used in Industrial settings, usually in a bus-type configuration between PLC stations. Although nowadays most stuff has moved over to copper or fiber Ethernet using EtherNet/IP (CIP), MELSEC IP, or Modbus over Ethernet. Mitsubishi/MELSEC only recently (Feb 2024) discontinued support for their 10Base2 and 10Base5 Ethernet I/O modules for their PLCs. BNC connectors are still a pain to terminate, but that thumb-thick Thicknet cable looks like a beast.

[@swebigmac100]

Amazing piece of gear. I think there were similar repeaters for AppleTalk networks...

[@bdhaliwal24]

These days networking has become a commodity and its easy to forget the serious engineering challenges that had to be solved decades ago. I'm glad you went into depth the history of this gear as well as the technical challenges.

[@Zizzily]

Yet another awesome video. Definitely gotta get 500 more subscribers so we can see it work!

[@philkarn1761]

Man, this takes me back....

[@N82SV]

We had one of these in the lab back in college. It allowed us to tie our lab lan into the university connection to the internet.

[@SB-qm5wg]

Those early tools were massive wow!

[@MLX1401]

Although the LAN Bridge is literally not that flashy, it has a really nice professional look to it that still flies. Could have easily been just a "beige box" doing its job, but the slightly convex and layered chassy with that carefully placed orange stripe and logo make it stand out as an object 😎