Fun fact - the OSI model is actually a real networking model that was competing against TCP/IP and Ethernet back in the day. You had companies demanding OSI compliant hardware and software as a competing network standard. Of course, it never took off, but that's why things like TCP/IP and Ethernet and other networking technologies never fit neatly into the OSI standard. They never were supposed to. As a vendor you offered an OS with OSI compliance, bought OSI compliant hardware and wrote OSI compliant apps. But it was too little too late, and in the end, things like BSD sockets won out for the user APIs, TCP/IP for networking protocols and Ethernet. I don't think there was enough OSI "stuff" around to actually made a production network from.
Oh, dude, production grade OSI networks were a (big) thing back in the day, in Govt, Telco and Banking. As recently as 3G we were still using OSI Protocol stuff.
@@Cadenadian Happy to hear. Though wondering how much of what I said in context of this historic network technology is still in present-day educational material?
I built my first 10base2 network in a my apartment, where I couldn't drill any holes in the walls to pass the cable between rooms. Instead, I ran the cables in the open space between the floor and bottom of each room's door when it was closed. It worked pretty well except for when a cable wasn't flat on the floor and someone slammed a bedroom door, taking down the entire network.
I have the strong urge of playing with some 10base2 stuff now! I always used 10baseT on these cards, and I even have a hub with both connections, and never got it working on base2. I think I didn't know how to set it up back then. Great video!
I remember the old macs in my school had connectors like these. But we had no cabling to hook them up. So instead I hooked them up using their serial ports so that all 7 of the macs could access the same printer. A room with 6 black and white macs in the back and one color mac. A laserwriter at one end and a imagewriter at the other end. Previously these printers were only hooked up to one computer on each end and people were ferrying floppy disks around to print stuff out. Once I set it all up everyone was able to print to either printer from any of the computers (which was important because the imagewriter could do color but the laserwriter could not.) Now imagine all this is being set up by an 11 year old in 2000. I attribute my love of computers today to stuff like that. I look forward to seeing how the network was supposed to be set up vs how I actually accomplished it.
I've only ever seen base2 once in my life around 2009-10 when I helped clean out an old office building. Got me a good load of old hardware from that place, including a bunch of NICs, a few laptops and other goodies. I never did get a base2 LAN going - it literally took until today to realize that I'd forgotten the terminators. So close and yet so far 😀
Don’t let that stop you, 50 and 75 ohm BNC terminators are available for cheap on Amazon. I know they’re used for oscilloscopes and probably other RF applications.
When I first started in computers. I installed several of those networks. Including putting ends of cables. Due to crimper and ends being a slight different. The middle brass pin would not crimp 100% or even fall off. So I would solder it. Solved so many problems and I can't tell you how many times I fixed competitors cables. I did have one tell me that is why they only used prepackaged cables. With Novell and the size of the network . Using multiple nic cards was the way to do. Even if you have 15 / 15 in a single room to cut down on traffic. I was so happy when we moved away from that stuff. Got tired of getting a call about the network going down and finding out somebody munched a connector or unplugged a few stations. leaving the T in place. I did have one strange instance. I did a long run between a building. It started on its own card from the server. Worked just great and was in the proper length. But, I had one workstation that would not work right at the beginning of the run in the other building. So basically the cable went 50ft before the first station. Maye 30. Anyway if I ended the network in that place. It would work fine. When I plugged in the rest of the network. That one station did not work. I tried different cards and everything. I could even run the station to another spot and it would work fine. Used a coupler to go around the spot. It was fine. Finally I did something I could not tell if it would work or not. I put a T up in the ceiling and run one cable down. Probably say 10ft of cable. Then on the back of the workstation put a T and a terminator. That solved the issue. I did all the lan tests I know of and it was not problem. Worked great for years until I left computing. So the only 3 terminator segment. I always wanted to go back and try that in a test environment. Never did though. Once again due to cat 3 / 5 cabling.
I love these ridiculously obscure stories :-) As I had talked with one who ran these networks at the time, I heard a story like putting the terminator not directly onto the T-Piece, but only after an extra 30cm whiptail cable. For whatever reason that seemed to work for improving stability. I have one to share myself, although I must admit, I was still in my youth and not very profound on the network topics back then. When my school threw the 10BASE2 stuff out, I was lucky to help there, and got to keep some of this old gear. But I couldn't get my nodes talking via a 25 meters coax wire, that I had feed through behind the walls in my parents home. The obvious was to replace the wire, or the NICs, but neither didn't help. It worked however with a flying 15m wire down the hallway. My teacher didn't really understand why, ultimately recommending me to install a 10BASE2 repeater in between, which solved the problem and it worked flawlessly for years.
I bet long runs introduced a lot of ground loop problems. Current will flow if there's a difference in voltage and know what causes differences in voltages? Differences in height. Base-T is so wonderful because it uses electro-magnetic field coupling like a transformer, so not only it doesn't need any current in the cable, it is immune to that. I just think of a Cat cable like a split transformer that runs a long distance, with half in one machine and half in the another. And they also use differential signaling, so it's also immune to interference. Coax is just a pain the ass.
@@monad_tcp Working at a small 2 person shop. I always figured if I had the $$$$ network testing equipment. I'm sure I would have found the answer. All I had is one those cheap light testers. It worked good for one thing. If the light was red. Even a tad. You know you had a bad cable or crimp some place. I found one with 1 braided ground wire touching the inside of the middle pin. Plus this was all before the internet. So all of my stuff mostly I got from books. I could even put a couple in and the rest would work fine. Just that spot. Also I forgot to say it was replacing a 2mb Arcnet network. They had 1 active hub and tons of those little 4 port passive hubs. I do remember calling other people and trying to talk my boss into getting a repeater or something. But, all I got was. It is within the correct distance.
I have a serious soft spot for 10Base2. I ran a multinode BBS off of this in the early 1990s. I had as many as 10-11 computers hooked up via BNC at one time, and I do remember the terminators getting loose on their own occasionally :). I started with a .. "free" copy of Novell Netware 286 before moving to LANtastic. LANtastic turned out to be faster and more flexible for running a board. Thank you for this great video!
@@THEPHINTAGECOLLECTOR Since LANtastic ran on top of DOS it worked with everything as drivers were compatible. I had an ISA caching SCSI card that flew with LANtastic. Look forward to your video!
@@THEPHINTAGECOLLECTOR Did several Lantastic network. 5.0 and tons of 6. They even came out with a dedicated server version. Never used it. But it was only a few $$$ away from Novell Server. So I just used that for a dedicated server. Had one place that used Lantastic E 2mbps cards. Had to look it up. Says serial network cards? Then they had this box. I can't find the name of it. Used Parallel port I think to provide network access. Sold one and it worked but was slow. Customer was not happy. But, hey what can you do when they have a old laptop. Not sure what I did not use a PCMCIA. Might not have had that slot. To many years ago.
Me and a couple of friends hosted the very first LAN-party at our scouting group in the early 90s and hosted several others later on as well. Certainly in the early years, even though 10base-2 networks were kind of passe already by then, they were convenient to set up and use, until one of us obtained a (IIRC) 24 port 100mbps switch, which was a very big step up. And all the computers we brought to these parties... PCs, DEC Alpha Workstations, SGI Indy's, etc. Ah good times...
Great video! I did have some ISA network cards that were NE2000 Compatible in the past, but they needed a BNC terminator or the lights would disco saying no cable connection found. The LAN port on them also wouldn't function without it.
My first network was some cheap discounted Novell Personal Netware kits I bought at an office supply store when I was about 12 or 13. It came with the software and 2 NE2100 cards. Sadly they were not as popular as the NE2000 cards and didn't have as good support. But after a day or so of playing around and watching the included VHS tape I had an IPX network going at my house in my bedroom and could play some networked games and share files. 10Base2 was pretty cool and the cables were fairly cheap.
Ah memories... that was my 3rd network, but 1st "serious" one from my Amiga days. First was just SLIP/PPP, then i upgraded it to PLIP and sometime in the 90's i finally got a cheap 2n hand 10Base2 NIC for the Miggy and the PC. Boy, was 10Mb FAST...
My first network was in 10base2. It was marvelous. And we played multiplayer Warcraft 2. My mother was always upset when unrolled the cable down the hallway in between rooms. A new world was opened to us. NICs were expensive but was so worth it. 😎
It sucks that we didn't have the electronics recycling industry back in the day, because all the 10BASE2 gear that businesses threw out in the 90s could have served home users quite well as the disadvantages of 10BASE2 don't really matter for home users, and instead we mostly got stuck waiting for integrated router/switch boxes.
We had at school 12 IBM PS/2 model 60 (286) computers wired in network with coax cable that booted from a 486 computer that acted as server. At boot, we had a huge number of drives (only one was for us to save our work) but we had to save our work on a floppy. One day, I get a bootable floppy with Prehistorik 2 game and when the teacher was not paying attention, I rebooted the computer. Until the game strted, the nework crashed and the only working computer was mine, the others were freezed. I have no idea what really happened; later I find that that floppy had a boot virus (parity, if I remember correctly). Borrowing the idea with the coax cable, that same summer, I with my neighbour made a network to play Doom and Red Alert: Command and Conquer. We both had 486 computers and we ran Win95.
My first network was 10 base 2. No surprise, not having a clue, I couldn't figure out at first why my network wasn't working without T's and terminators 😅. Great video as always.
Thank you for a great review of this technology. I am getting ready to set up a couple of old computers with 10base2, just for old times sake….😄 I installed these back in the Windows 95 days. But, I am building a legacy PC network…..
My parents ran a business out of a Victorian era building. Their first computer, which we still have and works was in 1985. Eventually, they had IBM-compatibles and the need to network these became obvious. Coaxial cable was passed through the floors to create a 10Base2 network. The workstations were running DOS and Sage Accounts, with one workstation also as the server. Eventually, in the late 1990s this was replaced with a 10Base-T network.
I used to install and fix these 10base2 networks. Always had t connectors terminators and bnc repair connectors in my bag. RG 59 was most commonly used for ARCnet network topologies.
Thanks for bringing this in, and it totally makes sense for transitions between all wiring standards. I'll go spotting, would love to get hold of a 10BASE-T MAU then :)
@@THEPHINTAGECOLLECTOR We often used AUI ports on repeaters, 10BASE-T hubs, bridges and routers to connect fiber transceivers as we migrated backbones from 10BASE-5 backbones to 10BASE-F fiber backbones.
My only 'experience' with 10BASE-2 was at school where it persisted to the mid 90s, says a lot about government funding of schools. Of course a popular thing to do was to steal the terminator and bring down the whole network. I do remember collision problems with 10BASE-T though, our early lan parties were in the time when ethernet switches were still quite expensive so of course the only option for poor teenagers was a cheap hub. Any sort of multiplayer game would have the collision light solidly lit on those little hubs and while I remember it working it wasn't ideal, seems silly now given how cheap switches became only years later but that was how it was at the time.
Of course, the hunt for the missing terminator (that you do not know is missing) is part of any 10-Base2 adventure. But there are many more. Like using a non-standard mode of 10Base-T Alta Research cards to be able to transmit over long and low quality Voice grade cables a 10Mbps Signal. Convincing some dumbass that if the X.25 (and by extension, the FSCKing ROSE) packets do not pass the local machine, and the packets from the router reach the remote server in another city, the problem lies in the local machine, not in the remote server. Oh, good times. A few things you failed to mention (or I failed to hear): - 10Base5 had a maximum lenght of 500mts, so, in principle, you could have two hubs in offices separated 500mts, connected by 10Base5 (as a backbone), and then use 10Base2 in each office for all the computers for cheapness sake (ask me how I know). - If you used MAUs, you had access to 10BaseFL, which gave you a maximum lengh for the backbone of 1000mts. and then could use 10base2 again for cheapnness sake (never even tried). - The automatic crossover specification is pressent as far back as 100Mbps eth, but optional. As long as ONE of the parties involved has it, it will work, is mandatory in Gbps Eth onwards. - Good quality terminators had to be Grounded, but you could ground only ONE of the terminators. In practice, everyone (us included) bought cheapo terminators like the ones in the video, but I had quality ones from HP in my HP K300 Servers (a thinet used along 3 racks for heartbeat services). Ah, fun times. Too bad there was the "Palace Revolt" and not only the IETF disobeyed the IAB and did not use TP4 for address extension (datatracker.ietf.org/doc/rfc1347/), but also, did not chose IPX for their TUBA Needs (datatracker.ietf.org/doc/html/rfc1791), so, we ended up with the IPv6 mess, and we are still waiting for it to take over... Anywho, too many ramblings for one night. Over and out.
Thanks for your inputs and the time for writing those missed aspects. For the most, I was just looking into 10BASE2. The side-tracked mentions, specifically for 10BASE5, came in just naturally when touching the comparison aspects and for putting the AUIs in context. When doing those videos, it''s always possible to miss out a few things. I'm glad you added them, and you're right, I did in fact overlook the need for grounding the terminators. I'll fill that in in an upcoming episode!
my experience was running the cable next to AC power for like 40 meters and I couldn't figure out why it kept shocking me and frying network cards. lol!
Air is not the transmission for wireless signals. Spacetime is the medium through which electromagnetic waves can propagate. I know... being pedantic again. But I ran 10base2 back in the mid 90s. My roomate and I would play Need for Speed over the network, Warcraft, transfer files, and share a printer with DOS/WFW3.11 and later Win95. And we used IPX/SPX on the LAN because nobody used TCP/IP on the LAN in those days... it was still a Novell deaminated world and games almost only supported IPX/SPX.
@@BandanazX Yeah, that might be correct from your physicist point of view. It‘s also correct to say, radio waves don‘t need a medium as a carrier. Still, they are very much influenced by the medium they travel through, i.e. speed is affected by vacuum, an athosphere, water, solid objects of any sorts (speaking of reflection, refraction, diffraction and absorbtion). But that‘s not the point here, as in regards to the networking context, I think it‘s enough to provide the picture of „the air being the cable equivalent“ in the layer 1 context. Frankly, we‘re installing the WiFi APs here on earth, so the readio waves are emitted and traverse not only the athmosphere, but to some extend also through solid walls, windows, etc. So the analogy I used is anyway incomplete. Despite not correct in a pure physics definition, bit I‘m not teaching physics here, and I can‘t assume everyone wants a lecture, I‘m sure it‘s good enough to get the meaning. But then again, If I’m supposed to provide a physically accurate statement for radio waves, then I should propably also not be talking about electrical signals, but rather about Ions and Electrons moving through a conductor. Maybe I‘m generalizing and simplifying too much. I‘m propably victim by day-2-day work experience with people not caring for all the real things behind, them being totally happy with simple to understand (but sometimes bot fully accurate) analogies.
There's also OSI Layer 8 - the interface between the PC and the technician on the helpdesk trying to worko out what they broke. Data type - words, vague phrases and facepalming gestures.
One other little nicety of 10base2, the line should be grounded by one station - in a hub situation, it would be the hub, not sure how or if the cards had an option. 10baseT was an interesting path, at first, it really had nothing special, with repeater hubs just putting all the packets on all the wires, with a logical bus but physical star configuration - and then came full duplex, 10Mb each way, but woe betide you if you mismatched the duplex setting
Ohh, the minor electric shocks from 10base2 when the mains filters leaked electricity to ground on each station, and grounded outlet was uncommon... DEC OfficeConnect was great in this regard as it had insulated BNC connectors (both for cables and the T thingies).
I always thought that coax networking made more sense for wiring up buildings, because you don't need to lay a separate cable to the server room for each station you intend to connect. Makes it much more efficient, more akin to electrical wiring. Of course, the drawback is much worse reliability, and when the 100BASE-TX standard and beyond entered the picture, it was a no-brainer. But every time I see a room full of Ethernet jacks, I get scared about how it must look inside the walls 😅
@@kFY514 Sure, TP needs more cabling than coax. But in terms of scalability, 10BASE2 was too limited. TP and switched networks was the next logical step. It is in theory possible to reach as much as 10 Gbps over coax nowadays, but it would require equire a practiable solution to avoid the packet collisions on a LAN.
Unique MAC addresses? Not if it was a PC from Packard Bell in the early 2000's. They foolishly sold thousands of PC's with NICs with the same MAC. It caused absolute hell on the NTL Cable network which demanded a unique ethernet card MAC address for their cable modems to connect to 😯🤣
@@looks-suspicious Ah yes, mine of course as well. I didn‘t intend to make this into a physics class about radio waves, but more looking for a very simplified to understand analogy.
Critique: despite still being widely taught, the OSI network layer model does not map directly to the TCP/IP stack we use today, instead it represented a failed networked proposal (failed in terms of lacked any commercial success). The TCP/IP model has similarities, but is simpler and represents our modern networking (including 10BASE-2) much more accurately. Most egregiously, OSI had proposed three layers on top of the transport layer (session, presentation, application) where nobody actually implements such things in a TCP/IP stack. After the transport layer (eg, TCP, UDP, SCTP), network protocols (eg, HTTP or SSH) are consumed directly by applications to do as they wish.
Thanks for your thoughts. I was taught both models as well, so I fully agree with the notion especially about layers 5-7 in TCP/IP. Those feel really artificially imposed onto the OSI model, and it's sometimes a struggle to explain, why a certain protocol is L5, L6 or L7. However, in the context of this series, I intend in doing some comparison of different protocols besides TCP/IP, like NetBIOS, IPX, AppleTalk, ARCnet, as I go along. Some of those don't even have their own model representation, so they just treated in context of the OSI model. As such I prefer using the OSI model overall, as I believe it's easier to explain different protocols in a comparable and relatable way using the same high-level model, e.g. "TCP and UDP, DDP and SPX are all (OSI) transport layer protocols". Although I have to acknowledge that even for other protocols, OSI is not always a perfect match, for example for NetBIOS, there's simply no L6/L7 equivalent, and the underlaying NetBIOS Frames Protocol sits both on L3 and L4. However, not all my viewers are indepth network engineers, maybe not even everyone is from the IT field. So I'm always trying to find somewhat a balance between how deep I go and how to explain it briefly in a understandable but still correct way, so it applies to a broader audience. The OSI model, as you mentioned, may not be fully accurate for TCP/IP, but it's also not totally wrong.
Fun fact - the OSI model is actually a real networking model that was competing against TCP/IP and Ethernet back in the day. You had companies demanding OSI compliant hardware and software as a competing network standard. Of course, it never took off, but that's why things like TCP/IP and Ethernet and other networking technologies never fit neatly into the OSI standard. They never were supposed to. As a vendor you offered an OS with OSI compliance, bought OSI compliant hardware and wrote OSI compliant apps. But it was too little too late, and in the end, things like BSD sockets won out for the user APIs, TCP/IP for networking protocols and Ethernet. I don't think there was enough OSI "stuff" around to actually made a production network from.
Oh, dude, production grade OSI networks were a (big) thing back in the day, in Govt, Telco and Banking. As recently as 3G we were still using OSI Protocol stuff.
Currently studying for my COMPTIA Network + Cert. this was a great educational video that I’d say all networking students should watch!
@@Cadenadian Happy to hear.
Though wondering how much of what I said in context of this historic network technology is still in present-day educational material?
Great episode! I've never seen 10BASE2 in person, other than one old D-link hub and one old network card I have, so this was super educational for me.
I remember when we in school still had a BNC network in the classroom in 2004.
Fun times.
I built my first 10base2 network in a my apartment, where I couldn't drill any holes in the walls to pass the cable between rooms. Instead, I ran the cables in the open space between the floor and bottom of each room's door when it was closed. It worked pretty well except for when a cable wasn't flat on the floor and someone slammed a bedroom door, taking down the entire network.
I have the strong urge of playing with some 10base2 stuff now! I always used 10baseT on these cards, and I even have a hub with both connections, and never got it working on base2. I think I didn't know how to set it up back then. Great video!
You're welcome! There will be more hands-on next week.
I remember the old macs in my school had connectors like these. But we had no cabling to hook them up. So instead I hooked them up using their serial ports so that all 7 of the macs could access the same printer. A room with 6 black and white macs in the back and one color mac. A laserwriter at one end and a imagewriter at the other end. Previously these printers were only hooked up to one computer on each end and people were ferrying floppy disks around to print stuff out. Once I set it all up everyone was able to print to either printer from any of the computers (which was important because the imagewriter could do color but the laserwriter could not.)
Now imagine all this is being set up by an 11 year old in 2000. I attribute my love of computers today to stuff like that. I look forward to seeing how the network was supposed to be set up vs how I actually accomplished it.
I've only ever seen base2 once in my life around 2009-10 when I helped clean out an old office building.
Got me a good load of old hardware from that place, including a bunch of NICs, a few laptops and other goodies.
I never did get a base2 LAN going - it literally took until today to realize that I'd forgotten the terminators. So close and yet so far 😀
Don’t let that stop you, 50 and 75 ohm BNC terminators are available for cheap on Amazon. I know they’re used for oscilloscopes and probably other RF applications.
I've had my retro PCs connected to a 10base2 network for about a year now, it's so beautifully janky I love it
Watched from Slovenia! Great video. I remeber doing BNC networks back in the day. Even one or two commercially but we switched to rj-45 very fast
I will listen to you talk about this topic for hours. And thank you for the captions as well.
When I first started in computers. I installed several of those networks. Including putting ends of cables. Due to crimper and ends being a slight different. The middle brass pin would not crimp 100% or even fall off. So I would solder it. Solved so many problems and I can't tell you how many times I fixed competitors cables. I did have one tell me that is why they only used prepackaged cables. With Novell and the size of the network . Using multiple nic cards was the way to do. Even if you have 15 / 15 in a single room to cut down on traffic. I was so happy when we moved away from that stuff. Got tired of getting a call about the network going down and finding out somebody munched a connector or unplugged a few stations. leaving the T in place. I did have one strange instance. I did a long run between a building. It started on its own card from the server. Worked just great and was in the proper length. But, I had one workstation that would not work right at the beginning of the run in the other building. So basically the cable went 50ft before the first station. Maye 30. Anyway if I ended the network in that place. It would work fine. When I plugged in the rest of the network. That one station did not work. I tried different cards and everything. I could even run the station to another spot and it would work fine. Used a coupler to go around the spot. It was fine. Finally I did something I could not tell if it would work or not. I put a T up in the ceiling and run one cable down. Probably say 10ft of cable. Then on the back of the workstation put a T and a terminator. That solved the issue. I did all the lan tests I know of and it was not problem. Worked great for years until I left computing. So the only 3 terminator segment. I always wanted to go back and try that in a test environment. Never did though. Once again due to cat 3 / 5 cabling.
I love these ridiculously obscure stories :-)
As I had talked with one who ran these networks at the time, I heard a story like putting the terminator not directly onto the T-Piece, but only after an extra 30cm whiptail cable. For whatever reason that seemed to work for improving stability.
I have one to share myself, although I must admit, I was still in my youth and not very profound on the network topics back then.
When my school threw the 10BASE2 stuff out, I was lucky to help there, and got to keep some of this old gear.
But I couldn't get my nodes talking via a 25 meters coax wire, that I had feed through behind the walls in my parents home.
The obvious was to replace the wire, or the NICs, but neither didn't help.
It worked however with a flying 15m wire down the hallway.
My teacher didn't really understand why, ultimately recommending me to install a 10BASE2 repeater in between, which solved the problem and it worked flawlessly for years.
layer 1 issue.
I bet long runs introduced a lot of ground loop problems. Current will flow if there's a difference in voltage and know what causes differences in voltages? Differences in height.
Base-T is so wonderful because it uses electro-magnetic field coupling like a transformer, so not only it doesn't need any current in the cable, it is immune to that.
I just think of a Cat cable like a split transformer that runs a long distance, with half in one machine and half in the another.
And they also use differential signaling, so it's also immune to interference.
Coax is just a pain the ass.
@@monad_tcp Working at a small 2 person shop. I always figured if I had the $$$$ network testing equipment. I'm sure I would have found the answer. All I had is one those cheap light testers. It worked good for one thing. If the light was red. Even a tad. You know you had a bad cable or crimp some place. I found one with 1 braided ground wire touching the inside of the middle pin. Plus this was all before the internet. So all of my stuff mostly I got from books. I could even put a couple in and the rest would work fine. Just that spot. Also I forgot to say it was replacing a 2mb Arcnet network. They had 1 active hub and tons of those little 4 port passive hubs. I do remember calling other people and trying to talk my boss into getting a repeater or something. But, all I got was. It is within the correct distance.
I have a serious soft spot for 10Base2. I ran a multinode BBS off of this in the early 1990s. I had as many as 10-11 computers hooked up via BNC at one time, and I do remember the terminators getting loose on their own occasionally :). I started with a .. "free" copy of Novell Netware 286 before moving to LANtastic. LANtastic turned out to be faster and more flexible for running a board.
Thank you for this great video!
@@jrherita LANtastic is coming soon. Don‘t know it from own experience, but I hear many people saying it was good!
@@THEPHINTAGECOLLECTOR Since LANtastic ran on top of DOS it worked with everything as drivers were compatible. I had an ISA caching SCSI card that flew with LANtastic. Look forward to your video!
@@THEPHINTAGECOLLECTOR Did several Lantastic network. 5.0 and tons of 6. They even came out with a dedicated server version. Never used it. But it was only a few $$$ away from Novell Server. So I just used that for a dedicated server. Had one place that used Lantastic E 2mbps cards. Had to look it up. Says serial network cards? Then they had this box. I can't find the name of it. Used Parallel port I think to provide network access. Sold one and it worked but was slow. Customer was not happy. But, hey what can you do when they have a old laptop. Not sure what I did not use a PCMCIA. Might not have had that slot. To many years ago.
Another high quality video with loads of geeky accurate and well researched details.
Thx :) And you're welcome!
Me and a couple of friends hosted the very first LAN-party at our scouting group in the early 90s and hosted several others later on as well. Certainly in the early years, even though 10base-2 networks were kind of passe already by then, they were convenient to set up and use, until one of us obtained a (IIRC) 24 port 100mbps switch, which was a very big step up.
And all the computers we brought to these parties... PCs, DEC Alpha Workstations, SGI Indy's, etc.
Ah good times...
Lots of scene parties in 90s had coax networks. With eh quite long chains, sparks galore.
I'm enjoying your recent tangent in to old networking technologies, it definitely scratches a different itch
@@robmcleod2876 glad to hear :)
Great video! I did have some ISA network cards that were NE2000 Compatible in the past, but they needed a BNC terminator or the lights would disco saying no cable connection found. The LAN port on them also wouldn't function without it.
My first network was some cheap discounted Novell Personal Netware kits I bought at an office supply store when I was about 12 or 13. It came with the software and 2 NE2100 cards. Sadly they were not as popular as the NE2000 cards and didn't have as good support. But after a day or so of playing around and watching the included VHS tape I had an IPX network going at my house in my bedroom and could play some networked games and share files. 10Base2 was pretty cool and the cables were fairly cheap.
Ah memories... that was my 3rd network, but 1st "serious" one from my Amiga days. First was just SLIP/PPP, then i upgraded it to PLIP and sometime in the 90's i finally got a cheap 2n hand 10Base2 NIC for the Miggy and the PC. Boy, was 10Mb FAST...
My first network was in 10base2. It was marvelous. And we played multiplayer Warcraft 2. My mother was always upset when unrolled the cable down the hallway in between rooms. A new world was opened to us. NICs were expensive but was so worth it. 😎
It sucks that we didn't have the electronics recycling industry back in the day, because all the 10BASE2 gear that businesses threw out in the 90s could have served home users quite well as the disadvantages of 10BASE2 don't really matter for home users, and instead we mostly got stuck waiting for integrated router/switch boxes.
Good old LAN Party memories. 😍
And Arnold Schwarzenegger likes that Ethernet technology! 😂
Perfect timing! I nostalgia-bought a 3c509 NIC last week!
Hands-on next week!
drilling into the yellow cable for the tap was fun, as one small short between shield and center will take down the network
We had at school 12 IBM PS/2 model 60 (286) computers wired in network with coax cable that booted from a 486 computer that acted as server. At boot, we had a huge number of drives (only one was for us to save our work) but we had to save our work on a floppy. One day, I get a bootable floppy with Prehistorik 2 game and when the teacher was not paying attention, I rebooted the computer. Until the game strted, the nework crashed and the only working computer was mine, the others were freezed. I have no idea what really happened; later I find that that floppy had a boot virus (parity, if I remember correctly).
Borrowing the idea with the coax cable, that same summer, I with my neighbour made a network to play Doom and Red Alert: Command and Conquer. We both had 486 computers and we ran Win95.
My first network was 10 base 2. No surprise, not having a clue, I couldn't figure out at first why my network wasn't working without T's and terminators 😅. Great video as always.
Thx, appreciated :)
Thank you for a great review of this technology. I am getting ready to set up a couple of old computers with 10base2, just for old times sake….😄
I installed these back in the Windows 95 days. But, I am building a legacy PC network…..
@@johnvanwinkle4351 you‘re welcome! 🙏
My parents ran a business out of a Victorian era building. Their first computer, which we still have and works was in 1985. Eventually, they had IBM-compatibles and the need to network these became obvious. Coaxial cable was passed through the floors to create a 10Base2 network. The workstations were running DOS and Sage Accounts, with one workstation also as the server. Eventually, in the late 1990s this was replaced with a 10Base-T network.
I used to install and fix these 10base2 networks. Always had t connectors terminators and bnc repair connectors in my bag. RG 59 was most commonly used for ARCnet network topologies.
Many aircraft still use 10 base 2 coax networks between the various avionics systems. They can be a real pain to troubleshoot.
AUI wasn't only for 10BASE5. 10BASE-T MAUs were also available, which gave your old Ethernet card a 10BASE-T port.
Thanks for bringing this in, and it totally makes sense for transitions between all wiring standards.
I'll go spotting, would love to get hold of a 10BASE-T MAU then :)
@@THEPHINTAGECOLLECTOR
We often used AUI ports on repeaters, 10BASE-T hubs, bridges and routers to connect fiber transceivers as we migrated backbones from 10BASE-5 backbones to 10BASE-F fiber backbones.
I think there were even fiber MAUs.
My only 'experience' with 10BASE-2 was at school where it persisted to the mid 90s, says a lot about government funding of schools. Of course a popular thing to do was to steal the terminator and bring down the whole network.
I do remember collision problems with 10BASE-T though, our early lan parties were in the time when ethernet switches were still quite expensive so of course the only option for poor teenagers was a cheap hub. Any sort of multiplayer game would have the collision light solidly lit on those little hubs and while I remember it working it wasn't ideal, seems silly now given how cheap switches became only years later but that was how it was at the time.
Of course, the hunt for the missing terminator (that you do not know is missing) is part of any 10-Base2 adventure. But there are many more. Like using a non-standard mode of 10Base-T Alta Research cards to be able to transmit over long and low quality Voice grade cables a 10Mbps Signal. Convincing some dumbass that if the X.25 (and by extension, the FSCKing ROSE) packets do not pass the local machine, and the packets from the router reach the remote server in another city, the problem lies in the local machine, not in the remote server.
Oh, good times.
A few things you failed to mention (or I failed to hear):
- 10Base5 had a maximum lenght of 500mts, so, in principle, you could have two hubs in offices separated 500mts, connected by 10Base5 (as a backbone), and then use 10Base2 in each office for all the computers for cheapness sake (ask me how I know).
- If you used MAUs, you had access to 10BaseFL, which gave you a maximum lengh for the backbone of 1000mts. and then could use 10base2 again for cheapnness sake (never even tried).
- The automatic crossover specification is pressent as far back as 100Mbps eth, but optional. As long as ONE of the parties involved has it, it will work, is mandatory in Gbps Eth onwards.
- Good quality terminators had to be Grounded, but you could ground only ONE of the terminators. In practice, everyone (us included) bought cheapo terminators like the ones in the video, but I had quality ones from HP in my HP K300 Servers (a thinet used along 3 racks for heartbeat services).
Ah, fun times.
Too bad there was the "Palace Revolt" and not only the IETF disobeyed the IAB and did not use TP4 for address extension (datatracker.ietf.org/doc/rfc1347/), but also, did not chose IPX for their TUBA Needs (datatracker.ietf.org/doc/html/rfc1791), so, we ended up with the IPv6 mess, and we are still waiting for it to take over...
Anywho, too many ramblings for one night. Over and out.
Thanks for your inputs and the time for writing those missed aspects.
For the most, I was just looking into 10BASE2. The side-tracked mentions, specifically for 10BASE5, came in just naturally when touching the comparison aspects and for putting the AUIs in context.
When doing those videos, it''s always possible to miss out a few things.
I'm glad you added them, and you're right, I did in fact overlook the need for grounding the terminators. I'll fill that in in an upcoming episode!
my experience was running the cable next to AC power for like 40 meters and I couldn't figure out why it kept shocking me and frying network cards. lol!
Air is not the transmission for wireless signals. Spacetime is the medium through which electromagnetic waves can propagate. I know... being pedantic again.
But I ran 10base2 back in the mid 90s. My roomate and I would play Need for Speed over the network, Warcraft, transfer files, and share a printer with DOS/WFW3.11 and later Win95.
And we used IPX/SPX on the LAN because nobody used TCP/IP on the LAN in those days... it was still a Novell deaminated world and games almost only supported IPX/SPX.
@@BandanazX Yeah, that might be correct from your physicist point of view.
It‘s also correct to say, radio waves don‘t need a medium as a carrier.
Still, they are very much influenced by the medium they travel through, i.e. speed is affected by vacuum, an athosphere, water, solid objects of any sorts (speaking of reflection, refraction, diffraction and absorbtion).
But that‘s not the point here, as in regards to the networking context, I think it‘s enough to provide the picture of „the air being the cable equivalent“ in the layer 1 context.
Frankly, we‘re installing the WiFi APs here on earth, so the readio waves are emitted and traverse not only the athmosphere, but to some extend also through solid walls, windows, etc.
So the analogy I used is anyway incomplete.
Despite not correct in a pure physics definition, bit I‘m not teaching physics here, and I can‘t assume everyone wants a lecture, I‘m sure it‘s good enough to get the meaning.
But then again, If I’m supposed to provide a physically accurate statement for radio waves, then I should propably also not be talking about electrical signals, but rather about Ions and Electrons moving through a conductor.
Maybe I‘m generalizing and simplifying too much. I‘m propably victim by day-2-day work experience with people not caring for all the real things behind, them being totally happy with simple to understand (but sometimes bot fully accurate) analogies.
There's also OSI Layer 8 - the interface between the PC and the technician on the helpdesk trying to worko out what they broke. Data type - words, vague phrases and facepalming gestures.
One other little nicety of 10base2, the line should be grounded by one station - in a hub situation, it would be the hub, not sure how or if the cards had an option.
10baseT was an interesting path, at first, it really had nothing special, with repeater hubs just putting all the packets on all the wires, with a logical bus but physical star configuration - and then came full duplex, 10Mb each way, but woe betide you if you mismatched the duplex setting
Ohh, the minor electric shocks from 10base2 when the mains filters leaked electricity to ground on each station, and grounded outlet was uncommon...
DEC OfficeConnect was great in this regard as it had insulated BNC connectors (both for cables and the T thingies).
Wow so the electrical signal was the same for 10-BASE-T and 10BASE-2 ?
Would a modern day version of this using wifi cards with SMA to BNC adapters work?
I've got one of those Etherlink II cards (albeit 16bit) - don't have a use for it however with my 10Mb Intel card as it does 10base 2/5/T.
I always thought that coax networking made more sense for wiring up buildings, because you don't need to lay a separate cable to the server room for each station you intend to connect. Makes it much more efficient, more akin to electrical wiring. Of course, the drawback is much worse reliability, and when the 100BASE-TX standard and beyond entered the picture, it was a no-brainer. But every time I see a room full of Ethernet jacks, I get scared about how it must look inside the walls 😅
Local switches. For a room of 30 machines put a 48 way switch and a pair of cables back to the main switch (one active one backup)
@@TheErador Of course. But that's still 30 wires to route around the room instead of 1.
@@kFY514 Sure, TP needs more cabling than coax. But in terms of scalability, 10BASE2 was too limited. TP and switched networks was the next logical step.
It is in theory possible to reach as much as 10 Gbps over coax nowadays, but it would require equire a practiable solution to avoid the packet collisions on a LAN.
Unique MAC addresses? Not if it was a PC from Packard Bell in the early 2000's. They foolishly sold thousands of PC's with NICs with the same MAC. It caused absolute hell on the NTL Cable network which demanded a unique ethernet card MAC address for their cable modems to connect to 😯🤣
@@thingi Well, „theorically“ unique ;)
I encountered a few SuperMicro servers with identical MAC addresses during the early 2000s as well.
Air is the transport medium for radio signals? My physics teacher is turning in his grave.
yep, it is, and it's almost as good as vacuum!
@@looks-suspicious Ah yes, mine of course as well.
I didn‘t intend to make this into a physics class about radio waves, but more looking for a very simplified to understand analogy.
Insightful breakdown of how networking works 🛜
Critique: despite still being widely taught, the OSI network layer model does not map directly to the TCP/IP stack we use today, instead it represented a failed networked proposal (failed in terms of lacked any commercial success). The TCP/IP model has similarities, but is simpler and represents our modern networking (including 10BASE-2) much more accurately.
Most egregiously, OSI had proposed three layers on top of the transport layer (session, presentation, application) where nobody actually implements such things in a TCP/IP stack. After the transport layer (eg, TCP, UDP, SCTP), network protocols (eg, HTTP or SSH) are consumed directly by applications to do as they wish.
Thanks for your thoughts. I was taught both models as well, so I fully agree with the notion especially about layers 5-7 in TCP/IP. Those feel really artificially imposed onto the OSI model, and it's sometimes a struggle to explain, why a certain protocol is L5, L6 or L7.
However, in the context of this series, I intend in doing some comparison of different protocols besides TCP/IP, like NetBIOS, IPX, AppleTalk, ARCnet, as I go along.
Some of those don't even have their own model representation, so they just treated in context of the OSI model.
As such I prefer using the OSI model overall, as I believe it's easier to explain different protocols in a comparable and relatable way using the same high-level model, e.g. "TCP and UDP, DDP and SPX are all (OSI) transport layer protocols".
Although I have to acknowledge that even for other protocols, OSI is not always a perfect match, for example for NetBIOS, there's simply no L6/L7 equivalent, and the underlaying NetBIOS Frames Protocol sits both on L3 and L4.
However, not all my viewers are indepth network engineers, maybe not even everyone is from the IT field.
So I'm always trying to find somewhat a balance between how deep I go and how to explain it briefly in a understandable but still correct way, so it applies to a broader audience.
The OSI model, as you mentioned, may not be fully accurate for TCP/IP, but it's also not totally wrong.