Saving Ethernet: the birth of switching (or actually Bridge)

An interesting short article from IEEE spectrum on the two guys at DEC responsible for the underpinning tech of "store and forward" and what has ultimately led to we have today as the all dominant LAN (and WAN) switched Ethernet technology standard.

This fork in the road took "old" CSMA/CD Ethernet on coax and spun in into a something which successively conquered and smashed FDDI, Token Ring and ATM - which could all at various points in time legitimately claim technical, throughput and efficiency superiority over Ethernet.

"No one knows what would have happened to Ethernet had Mark not invented the learning bridge. Perhaps someone else would have come up with the idea. But it’s also possible that Ethernet would have slowly withered away."

IEEE Spectrum: How Engineers at Digital Equipment Corp. Saved Ethernet

Token Ring had its problems because of the way the signal passed through all the wiring and devices connected to the ring. An Ethernet switch avoided that problem.

Thanks for that, an enjoyable read. I have fond memories of setting up 10Base5 and 10Base2 ethernet networks, and also a Token Ring network for HP Workstations in a couple of Glasgow colleges in the late 80s and early 90s. Nice to see the Spanning Tree protocol getting a mention too - I used to teach that to students around the same time

In reply to a post by Michael_Chare:

Token Ring had its problems because of the way the signal passed through all the wiring and devices connected to the ring. An Ethernet switch avoided that problem.

Back when I was a bit more hands on with these things we had passive TR concentrators, that used IBM Type 1 cabling - extremely thick shielded twisted pair - VERY IBM-esque thing of course in the 80's and 90's. They used a connector that would "heal" the ring as you disconnected a station from the concentrator.

Later of course there were active TR concentrators, that regenerated the token, and physically used RJ45 connections as it was delivered over 'normal' Category 3 or 5 structured cabling, which supplanted all the Type 1 we'd been paid to install 5 years previously 😂. I have a vague recollection the ones we came across were made by Allied Telesyn, but don't quote me on that.

Good old days eh 😂

In reply to a post by potterer:

Thanks for that, an enjoyable read. I have fond memories of setting up 10Base5 and 10Base2 ethernet networks, and also a Token Ring network for HP Workstations in a couple of Glasgow colleges in the late 80s and early 90s. Nice to see the Spanning Tree protocol getting a mention too - I used to teach that to students around the same time

When we were naughty undergrads, we'd think nothing of causing mass havoc in a computer lab full of VT100' type terminals or Tektronix X-terminals that were cabled up with a mixture of thin and thick ethernet. Amazing how much carnage a lowly missing BNC terminator can do 😅

That was always a good practical demonstration for students who couldn't be convinced that reflections on improperly terminated transmission lines were a real thing - I seem to recall that a similar challenge existed on the earliest SCSI storage systems as well

I used to go to the IBM London South Bank office to listen to lectures by a man from IBM. I later saw him at some other event. I think IBM had dispensed with his services. I think IBM lost out when cheaper Ethernet replaced expensive SNA.

There is quite a good book you should read about how and early networked was hacked and how the perpetrators were found by a physics student who studied stars.

In reply to a post by Michael_Chare:

I think IBM lost out when cheaper Ethernet replaced expensive SNA.

I read IBM sold Token Ring and patents and technology to Cisco, whom of course buried it 🤣

In reply to a post by Michael_Chare:

There is quite a good book you should read about how and early networked was hacked and how the perpetrators were found by a physics student who studied stars.

Is that the Cuckoos Egg? A good read, but I don't think the right one.