Ethernet hub

How it works

When a device sends an Ethernet frame into a hub, the hub regenerates and transmits that same signal out of all other ports. Hubs do not parse frames, learn Media Access Control (MAC) addresses, or make forwarding decisions. They provide basic signal amplification and timing recovery so that segments can be extended or several devices can be connected in a star-shaped physical topology.

Characteristics and operational limits

Because all ports on a hub share the same medium, the available bandwidth is shared among all connected devices. Hubs normally operate in half-duplex mode: connected devices must use Carrier Sense Multiple Access with Collision Detection (CSMA/CD) to avoid and resolve collisions when two or more stations transmit simultaneously. Collisions reduce effective throughput and increase latency as devices back off and retry.

• Layer: Physical layer device (Layer 1).
• Traffic handling: No MAC learning, no frame filtering, broadcasts to all ports.
• Duplex: Typically half-duplex; full-duplex requires switches and separate collision domains.
• Indicators: LEDs commonly show link and activity state for each port.

Types of hubs

Hubs are commonly described in a few categories:

1. Passive hubs: Simple wiring panels that only join cable segments without signal amplification; they rely on attached devices to provide signals.
2. Active (repeater) hubs: Contain electronics to regenerate and retime signals, extending reach and maintaining signal quality across multiple ports.
3. Managed or intelligent hubs: Offer monitoring LEDs, port status indicators and sometimes rudimentary management features; they still forward frames to all ports and do not perform switching.

Connectors and legacy support

Early hub models provided a mixture of connectors to accommodate different Ethernet variants — for example RJ45 for twisted-pair, BNC for 10BASE2 coax, or AUI for 10BASE5 segments. Modern hubs for copper Ethernet mostly use RJ45 ports for 10BASE-T and 100BASE-TX, though hubs that support higher speeds are rare because switching devices superseded them.

History, decline, and current uses

Hubs were popular when twisted-pair Ethernet in star topologies replaced coaxial bus wiring. During the 1990s they were widely used in small networks and workgroups. Over time, Ethernet switches became affordable and offered per-port bandwidth, full-duplex operation, and reduced collisions, which led to hubs becoming largely obsolete for general-purpose networking.

Despite their decline, hubs still have niche uses: they can be useful in laboratories and training environments for demonstrating CSMA/CD and collision behavior, or in some diagnostic scenarios where receiving all traffic on every port simplifies packet capture. Passive hub-like devices are sometimes used to mirror traffic for analysis, but managed network taps and switches with port mirroring have largely replaced hubs for professional monitoring.

Distinctions and noteworthy facts

Hubs are often confused with switches; the key difference is that switches operate at Layer 2 and forward frames only to the destination port based on MAC address learning, creating separate collision domains per port. Because hubs repeat every frame to every port, they are less secure (easier to eavesdrop on traffic) and less efficient. For most modern networks, switches or routers are recommended instead of hubs.

In summary, an Ethernet hub is a basic multiport repeater that played a central role in early local-area networks. Its simplicity and low cost were advantages historically, but the shared-medium behavior and collision-prone operation make it a legacy technology except for specific educational or diagnostic applications.