From one link’s perspective, the advantage of FD commu- nications is clear; it basically doubles the link’s throughput. However, such gain is less obvious in the case of a network with multiple interfering links. When these links operate in the same vicinity (i.e., the same collision domain), it is not always optimal for all links to operate in FD fashion [1, 3]. To illustrate, consider the three scenarios in Figure 1. In each scenario, two links are active at the same time and over the same frequency channel. As shown in Figure 1(a), transmitting in an HD fashion enables both links a → b and c → d to operate simultaneously over the same channel, achieving a total throughput of 2R bps (for simplicity, in this example we assume that all transmissions are associated with a constant rate R). However, if link a → b switches unilaterally from HD to FD, as shown in Figure 1(b), collisions may occur at both nodes a and d, reducing the network throughput to R bps (only a → b transmission is successful). The same argument applies if link c → d switches to FD mode instead of a → b. If both links operate in FD mode, collisions will occur at all four nodes, reducing the network throughput to zero (see Figure 1(c)). Note that this is a simplified example of a small network with only two links. The situation worsens with a higher number of links, where the collision probabil- ity increases between concurrent transmissions from different links.