A bridge is a Data Link layer networking device that connects two or more LAN segments, forwarding frames selectively based on destination MAC address, rather than blindly repeating traffic like a hub (Physical layer) or routing by IP like a router (Network layer). It builds a forwarding table by learning source addresses from frames it observes, reducing collisions and filtering unnecessary cross-segment traffic.

Types of network bridges

1. Transparent Bridge
Operates invisibly to the rest of the network — end stations are completely unaware of its presence. It builds its forwarding table automatically through a learning process (observing source addresses of incoming frames) rather than requiring manual configuration. This is the most common bridge type in Ethernet LANs, and typically runs the Spanning Tree Protocol to avoid loops when multiple bridges/redundant links exist.

2. Source Routing Bridge
Used historically in Token Ring networks. Here, the sending station itself determines and includes the complete path (sequence of bridges/rings) the frame should take, embedded within the frame header. The bridge simply follows the specified route rather than making its own independent forwarding decision.

3. Translational Bridge
Connects two LAN segments using different Data Link layer technologies (e.g. Ethernet to Token Ring), translating frame formats between the two so that stations on each side can communicate despite using different framing/addressing conventions.

4. Local Bridge
Connects two LAN segments that are directly, physically close together (e.g. within the same building), typically over a single hop with minimal latency concerns.

5. Remote Bridge
Connects two LAN segments that are geographically distant, typically over a WAN link (leased line, etc.), effectively extending a LAN across a wide-area connection.

Among these, the transparent bridge is by far the most widely used in modern Ethernet networks, since it requires no manual configuration and integrates seamlessly into existing LANs.