Understanding Bridges and Collision Domains
In any network design conversation, bridging inevitably appears, whether you’re looking at old‑school Ethernet or modern LANs. At its core, a bridge acts as a traffic filter that partitions a single collision domain into several smaller ones. A collision domain is simply the set of devices that share the same physical medium and therefore can interfere with each other’s transmissions. When devices are connected to a hub, the hub forwards every frame out all ports. Because the hub has no awareness of where a frame is headed, every device that is connected to the hub is in the same collision domain. As a result, any transmission can collide with another, forcing the devices to detect the collision, stop, back off, and then retransmit. These collisions degrade throughput and limit the overall bandwidth that each device can receive.
Bridges were introduced to cut that collision domain into smaller pieces. By attaching multiple ports, a bridge can forward frames only when necessary. Think of a bridge as a smart intersection: it knows which direction traffic needs to go and only opens the appropriate gates. The logic that decides whether a frame should be forwarded is based on the source and destination MAC addresses contained in the Ethernet frame. Because this logic operates at the data link layer (Layer 2), the device is classified as a bridge.
The first bridges ran the forwarding logic in software on a general‑purpose processor. Those devices were limited in the number of ports they could expose, usually no more than 16, and they processed frames at a speed that matched the 10 Mbps or 100 Mbps Ethernet links of the era. When a bridge receives a frame, it records the source MAC address along with the port the frame entered. Over time, this learning process builds a forwarding table that maps each known MAC address to a specific port. Once the table is populated, the bridge can decide, for each incoming frame, whether the destination address is in the same port, a different port, or unknown.
When a frame’s destination address is on the same port that the frame entered, the bridge discards it because the two devices are already sharing a collision domain and the frame would just cause a needless loop. If the destination is on a different port, the bridge forwards the frame only out that port. When the destination address is unknown – for example, a new device that has just powered on – the bridge floods the frame out every port except the one it arrived on. This ensures the frame reaches the intended host while keeping the forwarding table current. Broadcast frames, addressed to FF‑FF‑FF‑FF‑FF‑FF, are always flooded to every port except the source port because all devices on the LAN need to see the broadcast.
Bridges also preserve the broadcast domain. All ports on a bridge still share the same broadcast domain, meaning that a broadcast packet sent from any port will reach all other ports. The key difference is that bridges reduce the collision domain: each pair of devices that are connected to the same bridge port share a collision domain, while devices on different ports do not. By limiting the scope of collisions, bridges improve the overall reliability of the network, especially in environments where many devices need to communicate over a shared medium.
In practice, you might see a bridge deployed to isolate a departmental network, to segment a large office into logical units, or to separate legacy systems that still rely on hubs. Although bridges are largely supplanted by switches in most modern deployments, understanding their operation is essential for grasping how Layer 2 switching works today.
No comments yet. Be the first to comment!