![OSI Model Complete Guide: All 7 Layers Explained [2026]](https://www.networkingsignal.com/wp-content/uploads/2026/09/auto/art-osi-model-complete-guide.svg)
OSI Model Complete Guide: All 7 Layers Explained [2026]
Complete OSI model guide — all 7 layers explained with protocols, devices, PDUs, and real-world examples. Essential for…
Ring topology connects each device to exactly two neighbours, forming a closed loop. Data travels around the ring from station to station until it reaches its destination. There is no central device and no shared cable, every link is point-to-point.
The critical detail is that every station regenerates the signal it receives before passing it on. A ring station is not a passive listener; it is an active repeater. That is what gives rings their long reach, and what makes a single failed station capable of breaking the whole loop.
How Ring Topology Works
- Station A wants to send to Station D. It places the frame onto its outgoing link.
- Station B receives the frame, checks the destination address, sees it is not its own, and regenerates and forwards it.
- Station C does the same.
- Station D recognises its own address, copies the frame, and typically sets a “frame copied” bit before forwarding it onward.
- The frame continues around until it returns to Station A, which removes it from the ring. This return trip doubles as delivery confirmation.
Token Passing
Classic rings avoided collisions entirely using token passing instead of contention. A small special frame, the token, circulates continuously. A station may only transmit while it holds the token; when finished it releases the token back onto the ring for the next station.
The trade-off against Ethernet’s CSMA/CD is worth understanding:
- Deterministic. There are no collisions, and the maximum time a station will wait for the token is calculable. That mattered enormously for industrial and financial systems.
- Efficient under heavy load. Ethernet’s throughput collapses as collisions multiply; a token ring’s does not.
- Wasteful under light load. A station with data still has to wait for the token even when the ring is otherwise idle.
Characteristics of Ring Topology
- Each device has exactly two neighbours, one upstream, one downstream.
- Traffic in a single ring flows in one direction only.
- Every station acts as a repeater, so signal strength does not decay across the loop.
- There is no terminator and no central switch.
- Access is controlled by a token, not by contention.
Advantages of Ring Topology
- No collisions. Token passing removes contention completely.
- Predictable latency. Worst-case delay is bounded and can be calculated in advance, essential for real-time control systems.
- Performs well under load. Throughput stays stable as utilisation climbs.
- Long reach. Because every node regenerates the signal, a ring can span far greater distances than a shared-medium bus.
- Efficient cabling for a loop-shaped site. A metro fibre ring passing several buildings uses far less fibre than home-running each building to a hub.
Disadvantages of Ring Topology
- One break stops everything. In a single ring, a failed cable or a failed station severs the loop and halts all traffic. Rings do not “route around” a failure unless you build a dual ring.
- Adding or removing a device breaks the loop. Any change is disruptive unless the hardware supports bypass relays.
- Shared capacity. Only the token holder transmits, so the ring’s bandwidth is divided across all stations.
- Latency grows with size. Every extra station adds another hop and another store-and-forward delay.
- No privacy between stations. Every frame physically passes through every station on its way around, rings are not more secure than other topologies, despite the claim being widely repeated.
- Hard to fault-find. The symptom (total outage) is identical no matter which link or station failed.
Dual Ring Topology
The fix for the single-break problem is a dual counter-rotating ring: two independent rings between the same nodes, carrying traffic in opposite directions.
- The primary ring carries traffic clockwise; the secondary carries it anticlockwise.
- In normal operation the secondary sits idle as a standby path (or carries a second traffic load, depending on the technology).
- When a link or node fails, the two nodes adjacent to the break wrap, they loop the primary ring onto the secondary, folding the two rings into one longer working ring that avoids the fault.
- Traffic continues, at the cost of a longer path and no remaining protection until the fault is repaired.
FDDI popularised this design, and SONET/SDH built carrier networks on it. Wrap protection is fast, sub-50 ms, which is why rings dominated telecom transport for decades.
Where Ring Topology Is Actually Used
Token Ring lost to Ethernet in the LAN and IBM’s 802.5 is long gone. But the ring is alive and well elsewhere:
- SONET/SDH, the classic carrier transport ring with sub-50 ms protection switching.
- Metro Ethernet rings running ERPS (ITU-T G.8032), which blocks one link to break the loop and unblocks it on failure. This is the modern replacement for spanning tree in ring-shaped provider networks.
- Industrial Ethernet, PROFINET MRP and similar protocols run rings on the factory floor because a cable cut must not stop the line.
- Fibre ring backbones connecting campus buildings, where the ring shape saves substantial fibre versus a star.
One Correction Worth Making
You will often see the claim that Spanning Tree Protocol uses the Bellman-Ford algorithm to pick paths in a ring. It does not. STP runs a spanning-tree algorithm: switches exchange BPDUs, elect a root bridge, and each non-root switch blocks all but its lowest-cost path to the root, turning the physical loop into a loop-free logical tree. Bellman-Ford is a distance-vector routing algorithm used by protocols like RIP, at Layer 3, and has nothing to do with STP.
Ring vs Star vs Bus
| Ring | Bus | Star | |
|---|---|---|---|
| Media access | Token passing | CSMA/CD | Switched, no contention |
| Collisions | None | Yes | None |
| Single cable fault | Whole ring down (single ring) | Whole segment down | One device down |
| Signal regeneration | At every node | None | At the switch |
| Adding a device | Disruptive | Easy | Easy |
| Latency under load | Predictable | Degrades badly | Low and stable |
| Used today | Carrier and industrial | Essentially no | Yes, the standard LAN |
Full comparison of all five layouts in our network topology guide.
Frequently Asked Questions
What happens if one computer fails in a ring topology?
In a single ring, the loop is broken and all traffic stops, the failed station can no longer regenerate and forward frames. Token Ring MAUs mitigated this with bypass relays that electrically cut a dead station out of the loop; dual rings solve it by wrapping onto the standby ring.
Is ring topology unidirectional or bidirectional?
A single ring is unidirectional. A dual ring is effectively bidirectional, with the two rings carrying traffic in opposite directions.
Is ring topology more secure than bus or star?
No. Every frame passes through every station on the ring, so any station can read the traffic. This is a persistent myth.
Why did Token Ring lose to Ethernet?
Cost and momentum, not technical merit. Token Ring was technically superior under heavy load but the hardware was proprietary and expensive, while Ethernet was open, cheap and rapidly got faster with switching and full duplex, which removed collisions anyway.
Do modern networks still use rings?
Yes, but as a physical fibre layout with protection protocols like G.8032 ERPS or SONET/SDH, not as a token-passing LAN. Industrial Ethernet also uses rings for resilience.
