Wired and Wireless Network Topologies

The topology of interconnection in a DRTS was discussed in Chapter 2 for wired networks. However, with the advent of wireless networking, wireless topologies have also become very important, a typical example being wireless sensor networks. Logical topologies that can be implemented in a wired network can also be implemented over a wireless network. However, a wireless network has certain constraints arising out of reflection, refraction, and scattering of transmitted signals that arrive at the receiver following different trajectories from the transmitter to the receiver. Since these paths usually have different lengths, the copies arrive at different times (delay spread) and with different phase shifts at the receiver and overlap causing a significant reduction of received power, and this phenomenon is known as a deep fade. If the stations move relative to each other, then the number of trajectories and the resultant phase shifts vary with time. This results in a fast fluctuating signal strength at the receiver and the phenomenon is termed as fast or multipath fading. These cause a larger bit error and packet loss in a wireless network compared to a wired network. Statistical studies on packet loss and bit errors in a wireless network have been reported by contemporary researchers [1] and the results indicate the following:

• • Both bit errors and packet losses occur in bursts spread over a finite interval separated by intervening error-free periods.
• • The bit error rates depend on the modulation scheme. Schemes with higher bit rates are associated with higher error rates.

However, the most important phenomena that distinguish a wireless network from a wired network are hidden and exposed terminals. These are explained with Figures 3.3 and 3.4, respectively.

FIGURE 3.3

Hidden terminal.

FIGURE 3.4

Exposed terminal.

As seen in Figure 3.3, the transmission radii of the three stations P, Q, and R are such that transmitters P and Q are in the range of R, but P is not in the range of R and vice versa. If R starts to transmit to Q, the transmission cannot be detected by P by its carrier-sensing mechanism and it considers the medium to be free. Consequently, if P starts a packet transmission, a collision occurs at Q. The terminals P and R remain hidden to each other. In a wired local area network (LAN) this problem does not occur.

The problem of an exposed terminal also occurs because of overlapping radii of transmitters. As shown in Figure 3.4, the transmitters P, Q, R, and S are such that only communications between the pairs {P, Q}, {Q, R}, and {R, S} are possible.

In the event Q starts a transmission to P, R cannot start a simultaneous transmission to S, as it finds the medium busy. Transmitter Q thus acts as an exposed terminal for R. Topologies for DRTS applications on a wireless network should take these factors into consideration.