# Why token ring uses a physical star?

Initially, I thought that the reason token ring uses a physical star topology is to make up for the shortcomings of a physical ring. That is: In a ring, when a failure occurs at any point, the system goes down.

However, I have read the following about daisy chains in Wikipedia (emphasis by me):

By connecting the computers at each end, a ring topology can be formed. An advantage of the ring is that the number of transmitters and receivers can be cut in half, since a message will eventually loop all of the way around. When a node sends a message, the message is processed by each computer in the ring. If the ring breaks at a particular link then the transmission can be sent via the reverse path thereby ensuring that all nodes are always connected in the case of a single failure.

So it says that a message can be transmitted in both directions, which ensures the system not to go down in case of a failure at a point.

My question is, if such thing is possible, why this isn't used in token ring as well? If it was used in token ring, then token ring wouldn't need a physical star topology. It could be a physical ring as well.

So is there an error in this Wikipedia article or am I missing something?

Token ring may appear to be a physical star, but it really is a ring from the perspective of the cabling. A physical token ring cable has an outbound cable, and an inbound cable inside it. Breaking one of those internal cables will break the ring, and the MAU will bypass the cable, thus preserving the ring.

The token ring star can be thought of as a circle (ring) with various points pushed tightly toward the center of the circle. It is still a physical ring, just distorted:

The token travels around the ring, out to a workstation on the internal outbound cable leg, and back in on the inbound cable leg. If the workstation or one of the cable legs fails, the MAU heals the break, and the faulty link is isolated from the ring, preserving the ring.

• The MAU has relays at each cable connection point. If the electrical connectivity of an attached cable is broken (broken leg, shutdown workstation, etc.), the relay trips and closes the the cable off from the ring. Nov 8, 2015 at 17:15
• Oh I see. So token ring is actually not a star physically too. Then the remaining question is: Why token ring couldn't be wired as described in the Wikipedia article? That is: Connecting the nodes in a ring using an inbound and outbound cable between each nodes. Or even using a single cable between each node, provided that the same cable can be used for both receiving and transmitting. This way, we wouldn't need a MAU. (Well, at least on single point failures and/or failures on adjacent nodes.)
– Utku
Nov 8, 2015 at 17:24
• @Utku The MAU is the ring. If you didn't use a MAU, all the workstations would need to be powered up all the time, and any cable failure would break the ring, essentially giving you a token bus topology (a different IEEE LAN). What is described above (in your question) can handle a single break, but two workstations, opposite each other on the ring, being down would create two separate LANs which are unable to communicate. Nov 8, 2015 at 17:28
• Yes. Multiple non-adjacent failures would break the system in two separate LANs. But can we call it a token bus topology in that case? Because still, two non-adjacent computers on the same LAN cannot directly communicate. It seems like a lines to me.
– Utku
Nov 8, 2015 at 17:35
• @Utku, breaking the ring, as described in the article, will result in the path having ends (it will no longer be a ring with no ends). That means it is a bus. The end stations will need to know that the ring is broken beyond them so that they can send the token back the way it came instead of continuing on the normal path. This site isn't made for this type of discussion, and you should use Network Engineering Chat for more questions and further discussion. Nov 8, 2015 at 17:40