In a bus topology, the actual network speed experienced by users is not a fixed number but a performance level that declines significantly as network traffic and the number of connected devices increase.
While early bus-based networks like 10BASE2 and 10BASE5 Ethernet had a theoretical maximum speed of 10 Mbps, that bandwidth was shared among all devices on the single, central cable. This shared, contention-based design fundamentally limits speed and makes bus topology unsuitable for modern, high-speed networking.
The shared bandwidth and collision problem
The primary technical reason for the slow speed of a bus topology is that all devices are connected to and share the same physical cable, which serves as the "backbone".
- Shared medium: Any data transmission from one device is broadcast along the entire length of the cable. While only the intended recipient accepts the data, all other devices on the network must process the signal.
- Data collisions: With the legacy Carrier Sense Multiple Access with Collision Detection (CSMA/CD) protocol used in Ethernet bus networks, devices "listen" to see if the bus is clear before transmitting. If two or more devices transmit at the same time, their signals collide and are corrupted.
- Performance degradation: The devices must then wait a random amount of time before retransmitting the data. This process of detecting and recovering from collisions creates network latency, which increases exponentially as more devices are added or network traffic increases. The more users on the bus, the more collisions occur, resulting in slower speeds and reduced available bandwidth per device.
Factors that impact bus topology speed
The speed of a bus network is not determined by a single factor, but by a combination of physical and logistical elements.
- Number of devices (nodes): As more computers or other devices are added to the bus, the total network traffic increases. This leads to more frequent data collisions and a greater probability that devices will be forced to wait for retransmission, severely reducing the effective speed for all users.
- Network traffic levels: Even with a small number of devices, high-bandwidth activities—like streaming video or transferring large files—can congest the single cable and slow down the network for everyone.
- Cable length and quality: The physical length of the cable is a major limiting factor. Signals degrade over distance, and a bus that is too long will suffer from signal attenuation, leading to data loss and communication failures. Early coaxial bus networks were limited to specific lengths (e.g., 185m for 10BASE2) to maintain signal integrity.
- Proper termination: The backbone cable must be properly terminated at both ends with a resistor to absorb the signal. If terminators are missing or faulty, signals will bounce back and forth, causing echoes (ringing) that interfere with new transmissions and bring down the entire network.
Obsolete in the modern era
Due to these inherent limitations, bus topology has become obsolete for general-purpose Local Area Networks (LANs), having been replaced by more advanced and efficient topologies like the star topology.
- Switch-based star topology: In a star topology, each device has its own dedicated connection to a central switch. This eliminates data collisions, as the switch intelligently routes data only to the intended recipient. It also means that network speed does not degrade with additional devices, and a single device failure does not affect the rest of the network.
- Legacy vs. modern speed: The 10 Mbps maximum of a legacy Ethernet bus network is a fraction of the speed offered by modern switched Ethernet networks, which commonly operate at 1 Gbps, 10 Gbps, or even higher.
While some specialized, low-traffic industrial systems may still use simplified bus designs (like the RS-485 bus), it is no longer a viable option for a high-speed networking environment.