There have been several network-technologies with ring-topologies in the past. Examples are:
1) Token ring, which was a competitor against Ethernet. It was championed by IBM. But it was quite a bit more expensive than Ethernet. And more complicated, which meant harder to troubleshoot. Ethernet won.
2) FDDI. A bit like Token-Ring in the sense that it also was a LAN-technology, with a physical ring. But they used optical technology, with fiber-cables, in stead of coper. Very expensive.
3) Sonet and SDH. Americans use slightly different standards, and call theirs Sonet. The rest of the world uses SDH. Technology for telcos, to build large WAN networks. Supposedly the biggest benefit is that when a ring "breaks", it wraps around and immediately heals itself. The downside of this is that the ring consister of a pair of rings. And during normal operation (non-failure) you use only one of that pair. Basically you pay half your potential bandwidth for the option for restoration. That's a ridiculous amount of overhead, maybe acceptable in LANs, but imho not in WANs. Still in use today though. Telcos don't mind wasting their customer's money.
Building a real ring in your office with cables can be impractical. So after a while, the Token-Ring folks build a device called a "Token-Ring Concentrator". The concentrator would be like an Ethernet Switch. Just one box. And every device in the network would have a cable running from the device to the concentrator. Physically it would look like a star-network. But inside the concentrator, there would be connections that would make it ring. If one device was powered-off, or a cable would break, the concentrator would notice, and short-circuit its internal circuits and repair the ring.
This is what it would look like logically:
As you can see, physically, outside the concentrator, it looks like a star-topology. But inside the box, logically, it's still a ring.
Hope this helps.
Btw, ring-topology network equipment sucks.