10.7 Performance Evaluation
In this section, we evaluate the performance of OppCast. The compared protocols are as follows.
- Slotted p-persistence broadcast (Wisitpongphan et al. December 2007b) (Slotted-p). Upon receiving a packet from j, a node i rebroadcasts the packet with a fixed probability q after the backoff delay Tij, if it receives the WM packet for the first time and has not received any duplicates during the delay. Otherwise, it drops the packet. The delay-distance function is slotted and linear. Slotted-p is shown to be the best among the probability-based protocols (Wisitpongphan et al. 2007b). We set τ = 5ms, NS = 5, q = 0.5 (the settings used in Wisitpongphan et al. 2007b) and the forwarding range R = CR = 250 m in the simulations.
- Contention-based dissemination (CBD) (T-Moreno 2007), a typical broadcast protocol also based on opportunistic forwarding. It does not differentiate between relay nodes, and uses WM as implicit ACKs. A node in the forwarding range will set a backoff timer upon receiving a WM for the first time; it cancels the timer only if it receives duplicate WMs during the backoff process, otherwise it rebroadcasts. The delay-distance function is continuous and linear. We set the maximum backoff delay to be 10 ms, which is below the value (50 ms) adopted by (T-Moreno 2007) (since our CR = 250 which is smaller than the one used in (T-Moreno 2007), and the channel tends to be less congested). Also, we set R = CR = 250 m.
Meanwhile, the IDEAL protocol ...