**4.1 Simulation model**

The OPNET simulation, as shown in Fig. 5, has been conducted to examine the performance of the proposed scheme. We assume that there are 135 mobile stations used in the simulation and the traffic parameters are set as in Table 1.

$$f\_{init}(v) = \begin{cases} k \frac{1}{\sqrt{2\pi\sigma}} e^{-\frac{(v-w)^2}{2\sigma^2}} \\ 0, v < 0 \end{cases}, v \ge 0 \tag{1}$$

**4.2 Simulation results**

times of 6 to 23, 32 to 35, and 55 to 57.

Fig. 7. Packet delay on conversational traffic class

Fig. 8. Packet delay on streaming traffic class

Both L3 and L2 handoff schemes are simulated to prove the superiority of the proposed L2 handoff over L3 handoff on most popular services in mobile environment, such as streaming, web browsing, and Email services. These services are categorized into streaming, interactive, and background traffic class, respectively. In addition, conversational class is also added for video conferencing environment. As mentioned earlier, the low latency handoff for Mobile

A Fast Handover Scheme for WiBro and cdma2000 Networks 61

Fig. 7 - 10 show that the proposed L2 handoff scheme outperforms the L3 handoff on all kinds of service classes. In fact, performances resulted in each scheme should be the same except when handoffs occur. Therefore, performance differences shown in Fig. 9 - 12 are due to handoff processes. Figures also show that handoff occurrence is very frequent at interval

IPv4 in Fig. 3 has been implemented for the L3 handoff scheme.


Table 1. Simulation parameters

Fig. 5. Simulation model for performance evaluation

Fig. 7 shows only a small part of the entire model for simplicity although there are a lot of cdma2000 and WiBro cells. WiBro network cells and cdma2000 network cells are attached one by another in the simulation. Picocells and microcells are only used to generate frequent handoffs of mobile stations. Since the Markov mobility model used in the simulation, as shown in Fig. 8, is designed for mobile stations at low-speed (20 60km/h), and the following probability density function is used, where *m* represents the average speed of a mobile station in a cell (Janevski, 2003).

Fig. 6. Mobility model for mobile stations
