Every network is essentially a shared road. When dozens of devices want to send data over the same cable at the same time, something must decide who goes first and who waits. Without a traffic system, data packets would constantly crash into each other, forcing endless retransmissions and grinding the network to a halt. The rules that manage this traffic are called Media Access Control (MAC) protocols. They sit quietly in the background of every Local Area Network (LAN), deciding which device gets to transmit and when. Two of these protocols defined the early decades of networking: Ethernet’s CSMA/CD and the Token Ring’s token-passing method. Understanding how they work reveals not just how networks function, but why one of them eventually conquered the world while the other faded away.
Table of Contents
- What are MAC protocols?
- Ethernet and the CSMA/CD protocol
- Carrier sensing and multiple access
- Collision detection and the backoff process
- The Token Ring protocol
- How token passing works
- Priorities and reliability features
- Comparing Ethernet and Token Ring
- Performance and reliability
- Use cases and why Ethernet won
What are MAC protocols?
A MAC protocol is a set of rules that controls how devices on a shared network gain access to the transmission medium, such as a copper cable or fibre line. It operates at the data link layer, one of the lower layers of the network model, and works closely with the physical layer that carries the actual signals.
The core problem MAC protocols solve is contention. In a shared network, the medium is a single resource that many devices want to use at once. If two devices transmit simultaneously, their signals overlap and corrupt each other, an event called a collision. MAC protocols exist to prevent this chaos, ensuring that each device can send and receive data in an orderly way.
There are two broad philosophies for managing this access. The first is contention-based, where devices compete for the medium and deal with collisions when they happen. The second is controlled access, where a system grants permission so that only one device transmits at a time, eliminating collisions entirely. Ethernet represents the first approach, and Token Ring represents the second. Both fall under the family of IEEE 802 standards, which govern LAN and metropolitan area network technologies.
Ethernet and the CSMA/CD protocol
Ethernet is by far the most widely deployed LAN technology, and its dominance is rooted in its MAC protocol: CSMA/CD, which stands for Carrier Sense Multiple Access with Collision Detection. Specified in the IEEE 802.3 standard, CSMA/CD is the access method that traditional Ethernet networks use to share a common medium. The name itself describes exactly how it works, so it helps to break the term into its parts.
Carrier sensing and multiple access
Multiple Access means that many devices share the same communication channel. Carrier Sense means that before a device transmits, it first listens to the medium to check whether another device is already sending data. This listening step is the foundation of polite network behaviour. If the line is busy, the device waits. If the line is idle, the device begins transmitting its frame.
This “listen before you talk” principle dramatically reduces the chance of two devices speaking over each other. But it does not eliminate the risk entirely. Signals take a tiny amount of time to travel across a cable, so two distant devices might both sense an idle line and start transmitting at almost the same moment, before either has detected the other’s signal.
Collision detection and the backoff process
This is where the Collision Detection part becomes essential. While a device is transmitting, it continues to listen to the medium. If it detects that another device’s signal is interfering with its own, it knows a collision has occurred and that both frames have been corrupted.
When a collision is detected, the device does not simply give up. It transmits a special jam signal so that every other device on the network is aware that a collision has happened and the affected frames should be discarded. Each device involved then waits for a randomly chosen period of time before attempting to retransmit. This is called the backoff process, and the randomness is deliberate. If both devices waited the exact same amount of time, they would collide again immediately. By choosing different random delays, the devices avoid repeating the conflict.
One important consequence of this design is that CSMA/CD is non-deterministic. Because access depends on chance and timing, there is no guarantee about exactly when a device will get to transmit. Under heavy traffic, collisions become more frequent, and performance can degrade. For ordinary office tasks like file transfers and printing, this unpredictability rarely matters, but it becomes significant in time-critical situations.
The Token Ring protocol
Token Ring takes a completely different approach to the same problem. Developed by IBM and standardised as IEEE 802.5, it replaces contention with controlled access. Instead of letting devices compete for the medium, Token Ring uses a special signal called a token to grant permission to transmit. This eliminates collisions altogether, which was its central selling point in busy network environments.
How token passing works
In a Token Ring network, the devices are arranged in a logical ring, and a small data frame known as the token circulates continuously around this ring from one station to the next. The rule is simple: a device may only transmit data when it is holding the token.
When all stations are idle, the token is labelled as “free” and keeps circulating. A station that wishes to send data must wait until the free token reaches it. It then changes the token from free to busy by altering its bit pattern and attaches its data frame immediately behind it. The data travels around the ring to the destination, which copies the information and marks the frame to acknowledge receipt. The frame then continues around the ring back to the original sender, which removes it and releases a fresh free token for the next station to use.
Because only the token holder can transmit, and there is only one token, two devices can never send data at the same time. Collisions simply cannot occur. This makes Token Ring deterministic: every station is guaranteed a turn within a predictable amount of time, which is exactly what made it attractive for environments with heavy, sustained traffic.
Priorities and reliability features
Token Ring also includes a feature that Ethernet’s CSMA/CD lacks entirely: a priority system. The IEEE 802.5 standard allows messages to be assigned one of eight priority levels, encoded within the priority and reservation fields of the token’s access control byte. A station can reserve an upcoming token by entering its priority code, and only stations with a priority equal to or higher than the token’s current value can seize it. This lets important traffic, such as management messages, take precedence over routine data.
To keep the system healthy, one station is designated as a monitor. Its job is to ensure that exactly one token is always present on the ring and to reinsert a fresh token if the existing one is lost or corrupted. Some implementations also use a wire centre that can automatically bypass a failed station, improving the reliability and maintainability of the ring when a node goes down or a cable breaks.
Comparing Ethernet and Token Ring
When these two protocols are placed side by side, their trade-offs become clear. Each was strong in a different way, and the contest between them shaped the networking industry for years.
Performance and reliability
On paper, Token Ring looks superior under load. Because it is collision-free and deterministic, its performance remains predictable even when traffic is heavy. Ethernet, by contrast, can suffer from increasing collisions as more devices compete for a shared medium, which is why Token Ring was primarily used in networks where high traffic was anticipated. Token Ring’s ring structure and monitor station also gave it useful fault-detection capabilities.
Ethernet’s advantages lay elsewhere. It was simpler to understand, easier to install, and considerably cheaper. The cost of Token Ring equipment was much higher, driven largely by the price of its hubs, known as MAUs, and by the licensing royalties IBM demanded from manufacturers. This price gap pushed most vendors and businesses toward Ethernet.
Use cases and why Ethernet won
The decisive turning point came with the arrival of switched, full-duplex Ethernet. Early Ethernet shared a single medium among all devices, which is what made collisions a real concern. But Ethernet switches changed the equation entirely. A switch creates a separate collision domain for each connected device rather than forcing them to share one. With full-duplex links, a device can send and receive at the same time, and the practical problem of collisions essentially disappears.
Once this happened, Token Ring’s main advantage evaporated. Ethernet now offered both predictability and performance at a lower price, removing the very reason organisations had chosen the more expensive, more rigid token-passing model. Combined with its simplicity and its ability to scale from 10 Mbps to Fast Ethernet, Gigabit Ethernet, and beyond, Ethernet became the undisputed standard for LANs.
Today, Token Ring has all but vanished from real-world deployments, while Ethernet underpins the vast majority of wired networks. Yet Token Ring remains an important part of networking education. It demonstrates the controlled-access philosophy clearly, shows how deterministic networks can be built, and explains a genuine chapter in the history of how our networks came to work the way they do.
What do you think? If Token Ring offered guaranteed, collision-free access, do you believe its determinism could ever become valuable again in modern networks that demand strict timing, such as industrial automation or real-time systems? And does the story of Ethernet’s victory suggest that “simpler and cheaper” will always beat “technically elegant but expensive” in technology?
References
- https://www.ieee802.org/3/
- https://www.techtarget.com/searchnetworking/definition/Token-Ring
- https://quizlet.com/327897110/networking-flash-cards/
- http://www.networksorcery.com/enp/protocol/IEEE8025.htm
- https://www.how2lab.com/internet/computer-networks/token-ring
- https://patents.justia.com/patent/5553073
- https://www.tutorialspoint.com/data_communication_computer_network/token_ring_network.htm
- https://www.networkworld.com/article/2339489/token-ring-and-ethernet-differences.html
- https://www.geeksforgeeks.org/difference-between-token-ring-and-ethernet/
- https://www.justanswer.com/computer-networking/5cg45-features-token-ring-suitable.html
- https://www.ituonline.com/tech-definitions/what-is-token-ring-network/

Leave a Reply