ClassLan

Repeaters Hubs Switches Bridges

Repeaters

One distance limitation in LANs arises because electrical signals become weaker as they travel along a cable. Some LAN technologies allow two cables to be joined together by a device called a repeater. When a repeater detects a signal on one cable, it transmits an amplified signal on the other cable. This is illustrated in Figure 2.

Figure 2: Single repeater

The repeater connects directly to the Ethernet segments, without the use of a transceiver. The maximum size of an Ethernet segment is 500 m, so a single repeater can double the effective length of an Ethernet to 1,000 m. Unfortunately, this cannot be continued indefinitely! The Ethernet standard requires low delay for CSMA/CD to work; if the delay is too large, the scheme fails.

The inventors of Ethernet envisaged its deployment in an office building with two Ethernet segments on each floor and an additional vertical segment connecting the floors, so they put a limit of four repeaters in an Ethernet. This architecture is illustrated in Figure 3.

Figure 3: Multiple repeaters

In this Figure, no two stations are separated by more than two repeaters. Increasing the sizes of the segments by adding an additional repeater per floor would mean that no two stations are separated by more than four repeaters.

The most important disadvantage of a repeater is that it does not understand frames, it simply amplifies the electrical signal. Therefore, if a collision or electrical interference occurs on one segment, repeaters cause the same problem to occur on all other segments.

posted by ClassLan @ 1:26 PM