Shared Media LAN: What Shared Access Means for Ethernet

Shared media LAN refers to a local area network where all connected devices contend for the same communication channel, and in Ethernet, shared access means that only one device can transmit data at a time across the entire network segment. In practical terms, every device on a shared Ethernet segment must listen before sending, wait if the channel is busy, and back off if two devices transmit simultaneously. This article explains how shared access works, why it was the original Ethernet design, and how it compares to the switched networks used today.

Background: From ALOHA to Ethernet

Ethernet was developed in the 1970s at Xerox PARC, inspired by the ALOHA network used in Hawaii. ALOHA allowed multiple radio transmitters to share a single frequency, accepting that collisions would occur. Robert Metcalfe and his colleagues adapted this idea for wired networks, creating a system where all computers on a coaxial cable shared the same bandwidth. The original Ethernet ran at 10 Mbps over a thick coaxial cable, with every device tapping into the same physical medium. Because the cable was a single continuous conductor, a signal sent by one machine propagated to all others, making it a true shared medium.

In 1980, Digital Equipment Corporation, Intel, and Xerox published the DIX standard, which formalized the 10 Mbps Ethernet specification. The Institute of Electrical and Electronics Engineers (IEEE) later standardized this as IEEE 802.3 in 1983. For the next decade, most Ethernet networks used shared media, typically with coaxial cables or hubs. Only in the mid-1990s did switches become affordable, shifting Ethernet from shared access to dedicated connections.

How Shared Access Works

Shared Ethernet relies on a protocol called Carrier Sense Multiple Access with Collision Detection (CSMA/CD). The name describes the process:

• Carrier Sense: Each device listens to the cable before transmitting. If it detects a signal (a carrier), it waits until the channel is clear.
• Multiple Access: All devices have equal right to transmit; there is no central controller assigning turns.
• Collision Detection: If two devices transmit at nearly the same time, their signals overlap and corrupt each other. The sending devices detect this voltage change and abort transmission.

After a collision, each device waits a random amount of time (using a backoff algorithm) before retrying. The random delay reduces the chance that the same two devices collide again. The maximum cable length for a shared Ethernet segment is limited by the round-trip time: a device must be able to detect a collision before it finishes sending the smallest frame. For 10 Mbps Ethernet, that limits a segment to about 500 meters of thick coax, or 100 meters for twisted-pair cabling.

In a shared environment, the total bandwidth is fixed. If 10 devices are active on a 10 Mbps segment, each effectively gets only a fraction of that speed, and the fraction shrinks as traffic increases. Collisions also waste bandwidth, so the effective throughput is always less than the nominal rate.

Shared Media vs. Switched Ethernet

Modern Ethernet networks use switches, which create a dedicated, point-to-point connection between the switch port and the connected device. When a switch receives a frame, it forwards it only to the intended destination port, not to all ports. This eliminates collisions and allows full-duplex operation, where a device can send and receive simultaneously. In contrast, shared Ethernet is half-duplex: a device can either send or receive, but not both at once.

The table below summarizes the key differences:

• Collisions: Shared Ethernet has collisions; switched Ethernet does not.
• Bandwidth: Shared Ethernet divides bandwidth among active devices; switched Ethernet gives each device the full port speed.
• Duplex: Shared Ethernet is half-duplex; switched Ethernet supports full-duplex.
• Topology: Shared Ethernet uses a bus or star with hubs; switched Ethernet uses a star with switches.
• Cost: Shared Ethernet hubs were cheaper but less efficient; switches are more expensive but provide better performance.

For small networks with light traffic, shared Ethernet works fine. But as the number of devices or the volume of traffic grows, collisions increase and throughput degrades. Switched networks scale much better because each port is an isolated collision domain.

Common Problems with Shared Ethernet

Shared media Ethernet suffers from several inherent issues:

• Collisions and retransmissions: High traffic leads to frequent collisions, causing delays and wasted bandwidth.
• Distance limitations: The need to detect collisions restricts cable lengths, especially for coaxial segments.
• Single point of failure: A break in the coaxial cable can take down the entire segment.
• Security: Because all devices see all traffic, a device can easily capture data meant for another machine (packet sniffing).
• No prioritization: All devices are treated equally; there is no quality-of-service mechanism.

These problems were acceptable for early office networks but became severe as networks grew. The shift to switched Ethernet was driven by the need for higher performance, better security, and easier troubleshooting.

When to Use Shared Media (and When Not)

Shared media Ethernet is essentially obsolete for new installations. However, you might still encounter it in legacy systems, industrial control networks, or specialized applications where simplicity and cost are more important than speed. For example, some building automation systems use a 10 Mbps shared Ethernet bus for low-data-rate sensors. In such cases, the low bandwidth is sufficient, and the simplicity of a bus topology can be an advantage.

For any modern office, home, or data center network, switched Ethernet is the clear choice. It provides higher throughput, better reliability, and improved security. Even if you have an old hub, replacing it with a switch will immediately improve performance because each device gets a dedicated path to the switch.

How to Identify a Shared Media Network

If you are troubleshooting a network and suspect it might be shared, look for these signs:

• Hubs instead of switches: Hubs are Layer 1 devices that simply repeat signals to all ports, creating a shared collision domain.
• Coaxial cable: Thicknet (10BASE5) or Thinnet (10BASE2) coaxial cables indicate shared Ethernet.
• Half-duplex mode: Network adapters configured for half-duplex are likely connected to a hub.
• High collision counts: If a network interface reports excessive collisions, it is probably on a shared segment.

You can check the link speed and duplex mode in your network adapter settings. If it says “Half Duplex” or “10 Mbps Half”, you are on a shared medium. Modern switches negotiate full duplex automatically, so seeing half-duplex is a red flag.

Frequently Asked Questions

Why did Ethernet originally use shared media?

When Ethernet was developed, the technology to create cheap, high-performance switches did not exist. A shared coaxial cable was simple and cost-effective, requiring only a transceiver and a tap to connect a device. The CSMA/CD protocol was a clever way to manage access without a central controller, making the network robust and easy to extend.

Can I still use a hub today?

Yes, but it is not recommended. A hub will limit your network to 10 or 100 Mbps, force half-duplex operation, and create a single collision domain. For any modern application, a switch is far superior and costs only a little more. If you have a hub, replace it with a switch to improve performance and reliability.

What is a collision domain?

A collision domain is the set of devices that can collide with each other when transmitting simultaneously. In shared Ethernet, every device on the same segment is in the same collision domain. In switched Ethernet, each switch port is its own collision domain, so collisions do not occur.

Does shared Ethernet still exist in modern standards?

Yes, but only in legacy or specialized contexts. The IEEE 802.3 standard still includes clauses for 10BASE-T and 100BASE-TX half-duplex operation, and some industrial protocols use shared media. However, virtually all new Ethernet equipment supports full-duplex switched operation, and half-duplex is rarely used.

How does CSMA/CD handle collisions?

When a device detects a collision, it sends a jam signal to ensure all other devices know a collision occurred. Then it stops transmitting and waits a random backoff time before trying again. The backoff time is chosen from a range that doubles with each successive collision, up to a maximum, after which the device gives up and reports an error.

Is shared Ethernet slower than switched Ethernet?

Yes, in practice. Even if the nominal speed is the same (e.g., 10 Mbps), shared Ethernet has to handle collisions and retransmissions, which reduce effective throughput. Switched Ethernet provides full bandwidth to each device and supports full-duplex, so it is almost always faster.

Related

• Steam Families vs Epic Library Sharing: What You Can Share
• The Witcher 3 Accessibility Settings Explained
• Can You Buy Your Own Aeroplane for Less Than £1?