Ring and Bus Topologies: Where They Are Still Used Today

Where ring and bus topologies are still used: metro fibre and G.8032 rings, MRP and HSR in factories, CAN, Modbus RS-485 and 10BASE-T1S buses.

Jul 26, 2025 - 12:34
Nov 7, 2026 - 05:16
 0  371
Ring and Bus Topologies: Where They Are Still Used Today

Ring and bus topologies have disappeared from office LANs, which now use switched Ethernet in a star. Both survive where they solve a specific problem. Rings protect carrier fibre and factory networks because traffic can go the other way round after a break. Buses connect many small devices on one cable in cars, machines and sensor networks.

This guide shows where each topology is still used in 2026, which standards keep them alive, and when you would choose one today. It belongs to our networking fundamentals cluster under What Is the OSI Model and Why Is It Still Relevant Today?, and the standards named here (ITU-T G.8032, IEC 62439, IEEE 802.3cg, ISO 11898) were checked against their publishers' material in November 2026. For the more common star and mesh designs, see why you would choose a star topology over a mesh.

Key takeaways

  • Physical bus Ethernet (thick and thin coax) and Token Ring are obsolete. IEEE withdrew the 802.5 Token Ring standards and disbanded the working group in 2008.
  • Rings are alive in metro and carrier fibre networks (SONET/SDH rings, Ethernet rings using ITU-T G.8032) and in industrial Ethernet (IEC 62439-2 MRP, IEC 62439-3 HSR).
  • A modern ring is not a single loop that dies when one cable fails. A protocol blocks one link in normal operation and opens it within milliseconds when another link breaks.
  • Buses are alive in vehicles and machines: CAN, RS-485 networks such as Modbus RTU, and the newer 10BASE-T1S multidrop Ethernet from IEEE 802.3cg-2019.
  • Software borrowed both ideas. Cassandra places nodes on a hash ring, and many event-driven systems use a "message bus".
  • For a general-purpose LAN or data centre, star and leaf-spine designs are still the right default.

Ring and bus in one minute

In a bus, every device attaches to one shared medium. A frame sent by one node reaches all the others, so the nodes need a rule for who may talk when (collision detection, arbitration or a coordinator). The cable needs a terminator at each end to stop signal reflections.

In a ring, each device connects to exactly two neighbours and the links form a closed loop. A simple ring fails if one link breaks. Practical rings either use two counter-rotating rings (as FDDI did) or run a protection protocol that keeps one link blocked as a spare.

It also helps to separate physical topology (how cables run) from logical topology (how frames flow). Token Ring was a logical ring wired as a physical star to a central unit. A switched Ethernet LAN is a physical star, but each VLAN behaves like one broadcast domain, much like an old bus.

PropertyBusRingStar (for comparison)
CablingOne trunk cable with short dropsOne link between each pair of neighbours plus one to close the loopOne cable from each device to a central switch
Single cable breakSplits the bus; devices beyond the break are cut offProtected rings reroute the other way roundOnly the device on that cable is affected
Media accessShared: arbitration, CSMA/CD or coordinated accessToken passing (old) or normal switched Ethernet with a blocked link (modern)Dedicated full-duplex link per device
Where it lives todayVehicles, machines, field sensorsMetro fibre, factory and substation networksOffices, data centres, homes

What happened to the classic LAN ring and bus

Early Ethernet was a true bus. 10BASE5 (thick coax) and 10BASE2 (thin coax) let every station tap into one cable, and all stations shared one collision domain. IEEE 802.3 has long since deprecated both for new installations. Structured twisted-pair cabling to hubs, and then to switches, replaced them because a single bad connector on a coax bus could take down the whole segment.

Token Ring (IEEE 802.5) and FDDI used a ring. A token circulated, and only the station holding it could transmit, which gave predictable access at a time when Ethernet collisions were a real concern. FDDI ran two counter-rotating fibre rings at 100 Mb/s so that a break could be wrapped around. Switched full-duplex Ethernet removed collisions altogether, got faster and cheaper every few years, and both technologies faded. The IEEE 802 Executive Committee disbanded the hibernating 802.5 working group in October 2008 after its standards were withdrawn.

So if a textbook says a classroom or small office "uses a bus topology", that description is decades out of date. A small office today has a switch, which is a star. The post on how LAN, WAN and MAN architectures use different topologies shows the current picture.

Where ring topologies are used today

Carrier and metro fibre rings

Telecom operators lay fibre along roads and railways, and a loop through several exchanges or towns is often the cheapest way to give every site two diverse paths. For decades SONET and SDH protection rings handled this. Many operators now run Ethernet directly on the ring, protected by ITU-T Recommendation G.8032, Ethernet Ring Protection Switching (ERPS).

G.8032 avoids a forwarding loop by keeping one link, the Ring Protection Link (RPL), blocked under normal conditions. When a link or node fails, the nodes exchange R-APS messages and the RPL owner unblocks the spare link. The Recommendation targets protection switching within 50 ms. Real recovery time also depends on how quickly the failure is detected and on ring size, so vendors quote ranges.

Industrial Ethernet on the factory floor

Production lines are long and narrow. Running every machine back to a central switch wastes cable, while chaining switches from one machine to the next and closing the loop gives redundancy with little extra cable. The IEC 62439 series standardises high-availability automation networks:

  • MRP (Media Redundancy Protocol, IEC 62439-2): one Media Redundancy Manager blocks one of its ring ports and opens it when the ring breaks. It is widely used with PROFINET. Recovery time depends on the configured profile and ring size, typically tens of milliseconds to a few hundred.
  • HSR (High-availability Seamless Redundancy, IEC 62439-3): each frame is sent both ways round the ring and the receiver discards the duplicate, so a single break causes no recovery delay at all. It is used where even a short interruption matters, for example in electrical substation automation.

Our post on fault tolerance in mesh networks covers the other common way to get redundancy, at the cost of more links.

Rings in distributed software

The ring idea also shows up in software that DevOps engineers run every day. Apache Cassandra maps every node to one or more tokens on a continuous hash ring. A key is hashed to a position, and replicas are chosen by walking clockwise until enough distinct nodes are found. Virtual nodes give each machine several positions so that data rebalances smoothly when nodes join or leave. Amazon's Dynamo paper described the same consistent-hashing ring. This is a logical ring of data ownership, not a cabling plan, but it is the "ring topology" you are most likely to troubleshoot in a modern job.

Where bus topologies are used today

CAN bus in vehicles and machines

Controller Area Network (CAN), standardised in ISO 11898, links the electronic control units in cars, trucks, tractors, lifts and medical devices over one twisted pair terminated at both ends. Nodes broadcast short frames, and an identifier-based arbitration scheme decides which message wins when two start at once, without destroying either. A bus suits a vehicle because it saves copper weight and connectors compared with wiring every unit to a hub.

Linux has native CAN support through SocketCAN, so an embedded Linux gateway can bring a CAN interface up with the same ip tool you use for Ethernet:

sudo ip link set can0 type can bitrate 500000
sudo ip link set up can0
ip -details link show can0

RS-485 and Modbus RTU in industrial sites

Many building-management systems, energy meters, PLC peripherals and solar inverters still talk Modbus RTU over an RS-485 multidrop bus. One master polls each addressed device in turn on a single twisted pair, with terminating resistors at both ends of the cable. It is old, slow and still everywhere, so engineers who connect operational technology to cloud dashboards meet it constantly, usually through a gateway that converts Modbus RTU to Modbus TCP or MQTT.

10BASE-T1S: the bus comes back to Ethernet

IEEE 802.3cg-2019 added 10BASE-T1S, a 10 Mb/s single-pair Ethernet that supports a half-duplex multidrop "mixing segment". The standard's baseline is at least 8 nodes on at least 25 m of cable. A mechanism called PLCA (PHY-Level Collision Avoidance, Clause 148) gives each node a turn to transmit, which avoids the collision problems of old coax Ethernet. The target markets are automotive zonal architectures and industrial sensors, where engineers want Ethernet and IP all the way to the edge without a switch port for every small device.

Message buses in software

The word "bus" is also used for software where producers publish events to a shared channel and any subscriber can read them, as in an enterprise service bus or an event bus. Like a physical bus, every subscriber can see every message on a topic. Unlike a physical bus, the transport underneath is ordinary switched networking. Our post on how networking concepts power distributed systems explores this overlap.

Ring and bus applications at a glance

ApplicationTopologyStandard or technologyWhy that topology
Metro and regional fibreRingSONET/SDH rings, ITU-T G.8032 ERPSTwo diverse paths with the least fibre; fast protection switching
Factory production linesRingIEC 62439-2 MRP (often with PROFINET)Redundancy along long, narrow layouts with little extra cable
Substation automationRingIEC 62439-3 HSRZero recovery time after a single failure
Distributed databasesLogical ringCassandra token ring, consistent hashingEven data spread and easy scaling
Vehicle electronicsBusCAN (ISO 11898)Low wiring weight; deterministic arbitration
Building and energy systemsBusRS-485 with Modbus RTUMany low-rate devices on one cheap cable
Automotive and industrial edge EthernetBusIEEE 802.3cg 10BASE-T1S with PLCAEthernet to small nodes without one switch port each
Office LANs and data centresStar / leaf-spineSwitched EthernetNeither ring nor bus is used here

When would you choose a ring or a bus today?

Choose a ring when sites or machines lie along a line or loop, each needs two paths, and running separate cables from every node to a pair of core switches costs too much. Always pair it with a protection protocol (G.8032, MRP, HSR or at minimum spanning tree). A loop of plain switches without one creates a broadcast storm within seconds.

Choose a bus when you have many small, low-bandwidth devices close together, cable weight or connector count matters, and a standard such as CAN, RS-485 or 10BASE-T1S already fits the devices. Keep the bus short, terminate it correctly and plan for the fact that one damaged trunk segment affects every node beyond it.

For everything else, from office floors to Kubernetes clusters, use a star or a leaf-spine fabric. If you are weighing combinations, read the trade-offs between hybrid and traditional topologies.

Common mistakes and limitations

  • Closing a loop without a protocol. Connecting the last switch back to the first without ERPS, MRP or spanning tree floods the network with looping broadcasts.
  • Missing or extra bus terminators. CAN and RS-485 need termination at the two physical ends only. Errors that come and go often trace back to this.
  • Mixing ring protocols. MRP, G.8032 and vendor ring protocols do not interoperate. Use one per ring and check every switch supports it.
  • Assuming a ring doubles capacity. With one link blocked, a protected Ethernet ring forwards traffic over a single path. Redundancy, not extra bandwidth, is the goal.
  • Treating a bus as secure. Every node on a broadcast bus sees all traffic. Vehicle and industrial networks increasingly add gateways and segmentation for this reason.

Why this matters for DevOps and cloud engineers

You will rarely cable a ring or a bus yourself, but you will meet both. Carrier links between your data centres may ride a G.8032 ring, which explains brief packet loss during a fibre cut. IoT and edge projects bring CAN and Modbus data into Linux gateways and then into cloud pipelines. Cassandra and similar databases expose their hash ring in day-to-day operations such as adding nodes or repairing data. Understanding the topology helps you read an incident correctly. Our DevOps training program covers Linux networking, cloud networking and the distributed systems built on top, in our Kharadi, Pune classroom and live online across India. For the wider role of DevOps, start with What Is DevOps?.

Frequently asked questions

Is bus topology still used today?

Yes, but not in office LANs. Bus wiring is common in vehicles (CAN), industrial and building systems (RS-485 with Modbus RTU) and the newer 10BASE-T1S single-pair Ethernet defined in IEEE 802.3cg-2019. Coax bus Ethernet such as 10BASE2 is obsolete.

Is ring topology still used today?

Yes. Telecom operators run protected fibre rings using SONET/SDH or Ethernet Ring Protection Switching (ITU-T G.8032), and factories and substations use rings with IEC 62439-2 MRP or IEC 62439-3 HSR for fast recovery after a cable break.

What is a real-life example of a ring topology?

A metro fibre network that links several telephone exchanges in a loop is a common one. Traffic normally flows one way round, and when a fibre is cut the G.8032 protocol unblocks a spare link so traffic flows the other way, with a design target of 50 ms.

What is a real-life example of a bus topology?

The CAN bus in a car. Engine, brake, airbag and body control units share one twisted pair, broadcast short messages and use identifier-based arbitration to decide which message goes first when two transmit at once.

Why did Token Ring disappear?

Switched full-duplex Ethernet removed the collisions that once made token passing attractive, while becoming faster and much cheaper. IEEE withdrew the 802.5 standards and disbanded the working group in 2008.

Does a ring network go down if one cable breaks?

A simple single ring does. Modern rings use a protection protocol that keeps a spare link blocked and opens it when a failure is detected, or, with HSR, send every frame both ways so a single break causes no interruption.

What is the difference between physical and logical topology?

Physical topology is how the cables run; logical topology is how data flows. Token Ring was a logical ring wired as a physical star, and a switched Ethernet VLAN is a physical star that behaves like one shared broadcast domain.

Should I use ring or bus topology for a new office network?

No. Use switched Ethernet in a star, with redundant uplinks or a leaf-spine design for larger sites. Rings and buses make sense only in the specialised settings described above.

References

What's Your Reaction?

Like Like 0
Dislike Dislike 0
Love Love 0
Funny Funny 0
Angry Angry 0
Sad Sad 0
Wow Wow 0
Mridul I am a passionate technology enthusiast with a strong focus on DevOps, Cloud Computing, and Cybersecurity. Through my blogs at DevOps Training Institute, I aim to simplify complex concepts and share practical insights for learners and professionals. My goal is to empower readers with knowledge, hands-on tips, and industry best practices to stay ahead in the ever-evolving world of DevOps.