In an alternate universe, I’d probably be a History Professor somewhere. The launch of the Classics Rock blog has given me the opportunity go take care of the history part, but at the same time, I’ve been lucky enough to serve as an Adjunct Professor for ITT Technical Institute here in Louisville as well. I’ve been teaching mostly networking classes lately and I keep noticing a consistent theme in the courses. Every course has spent a non-insignificant part of the curriculum discussing Token Ring. Once the center of a great technical debate in the 80’s and 90’s, the Token Ring vs. Ethernet wars are long settled. But it’s kind of surprising that Token Ring is still taught.
The text for the class calls Token Ring the “second most popular technology” for connecting local area networks. Which, I explained to the class was like saying that traveling by ship is the second most popular way to go to Europe. Yeah, it can still be done, but almost nobody does so anymore.
Token Ring revisited
Token Ring was developed in the early 1980’s by IBM as a way to connect PCs together. It was a direct competitor to the Ethernet standard that we use today and the ARCnet standard that disappeared in the mid-80’s.
If you remember your network topologies, Token Ring was a ring technology whereby data went from one computer to another in a ring format. Most often, rather than direct connection from computer to computer, you’d locate a MAU in a central wiring closet and then run lines back to each computer. A MAU is IBM-speak for Multiple Access Unit and they were roughly analogous to a switch today. The ring would actually function inside of the MAU. So you’d have a logical ring working with a physical star topology.
Token Ring presented some significant technological advances over Ethernet. There was a huge debate in the 80’s and 90’s about which technology should be used. As we know now, Ethernet won, but if you just looked at the specs, there was no reason why it should have.
First, there was the way that both technologies allowed computers to communicate on the network. Ethernet uses a thing called CSMA/CD, which essentially means that each computer waits before it transmits on the line. If noone else is talking, it transmits a packet. If by chance another computer transmits at the exact same time, a collision occurs, no data transfers, and the two computers wait a random amount of time before transmitting again. By contrast, Token Ring passes a token around the network that grants computers the right to transmit. If a computer doesn’t have a token, it has to wait until the token is freed. This ensures that no two computers can try to access the wire at the same time, making bandwidth usage more efficient.
Second, Token Ring allowed for larger packet sizes than Ethernet. The maximum packet size you could have in an Ethernet frame was 1514 bytes. Token Ring packet sizes depended how fast your line speed was. If you were running Token Ring at 4 Mbps, you could have a packet size of 4550 bytes. If you ran Token Ring at 16Mbps, you could have a packet size up to 18,200 bytes long. Obviously, larger frame sizes mean that you could transmit more data in each packet, reducing the total number of packets on the wire. This lead to less congestion in routers and network cards.
Finally, you had the line speed advantage of Token Ring over Ethernet. Even though Token Ring only first shipped at 4Mbps speed while Ethernet ran at 10Mbps, because of token-passing and frame sizes, you could actually get better performance out of a 4Mbps Token Ring installation. IBM then cranked the speed up to 16Mbps, which had the added benefit of gaining a marketing advantage. Even if people didn’t understand how 4Mbps Token Ring was faster than 10Mbps Ethernet, the 16Mbps speed would overcome that.
If it’s so good, why did Ethernet win?
Unlike OS/2 where IBM lost because it didn’t know how to market the technology, Token Ring’s failure in the market place had little to do with bad marketing. In this case, it probably had more to do with greed and advancing technology. Ethernet grew ever increasingly faster, going from 10Mbps to 100Mbps, to where we are now with Gigabit Ethernet and 10G Ethernet. Even though Token Ring was more efficient – to the point that on an unswitched Ethernet segment 16Mbps Token Ring could still be faster than 100Mbps Ethernet in some cases – eventually the increasing Ethernet speeds overwhelmed Token Ring. Added to that was the fact that vendors started introducing switches which eliminated the problems of collisions which slowed down Ethernet networks in the first place. This completely removed one of the major technical advantages that Token Ring had over Ethernet. Switching and 100Mbps speeds make 16Mbps technology obsolete.
Probably the main reason why Token Ring failed however was pricing. IBM charged too much for royalties to vendors that wanted to produce Token Ring cards and MAUs. This made all Token Ring equipment too expensive. A Token Ring card could cost 5 and 6 times as much as an Ethernet card. Add on the cost of more expensive cabling and MAUs, and Token Ring just priced itself out of the market.
Token Ring vendors tried to increase the speed of the technology to 100Mbps to overcome the Ethernet advantages, but by the time they did at the turn of the century, it was too late. Ethernet had won the day.
Is it still around?
I told the class that chances are they wouldn’t run into any Token Ring installations today. If they worked in larger organizations where IBM had sold a lot of equipment, there could still be some Token Ring installations still around. But in smaller organizations or to deploy for new installations? They might as well be taking a steamship to London.