Wimax standards and Security The Wimax

Simulation of IEEE 802.16 QoS Operation

1. 
   Admission Control
   

The policy of admission control is similar to the Erlang B model that has been used by the voice networks for many years [11]. We apply the same concept of the Erlang B model and use the IEEE 802.16 QoS parameters to describe the

model behaviors. Since the UL and DL use separate channels, the admission control is applied to DL or UL separately. The procedure of admission control is required for UGS, rtPS, and nrtPS. BE requests are always granted as there is no committed bandwidth. In the simulation model of admission control, we have the following parameters:

1. 
   Number of SSs: this is a fixed parameter during the simulation.
   
2. 
   Connection requests per minute per SS (λ): This is the arrival rate to or from each SS and it follows the Poisson distribution. During the simulation, we use the interarrival rate (µ) to determine the time interval of the next request
   

Interarrival rate (µ) = 1/λ.
This parameter (µ) follows the exponential distribution in the simulation mode.

3. 
   Bandwidth request (in multiples of DS0): This is either MSTR for UGS or MRTR for rtPS/nrtPS. It follows the exponential distribu- tion. The bandwidth request shall be at least one DS0 (64 Kbps); otherwise, no data can be transmitted. The average bandwidth request is set at 4 × 64 = 256 Kbps.
   
4. 
   
   ÷ × =
   Data size (S) in bytes: This parameter also follows the exponential distribution. The data size and bandwidth request determine the service time of a request after its admission. For example, if the data size is 2 Mb and the requested bandwidth is 1 Mbps, the duration of the service will be 2M 1.0M 8 16 s. This is based on the store- and-forward scheme used in most network devices. The average data size is set at 2 MB.
   
5. 
   Total bandwidth (T) in bps: This is the total bandwidth allocated to either a DL or UL channel. As discussed earlier, bandwidth request could be provisioned manually or dynamically allocated. If it is statistically provisioned, the subscriber does not need to request for admission. Therefore, the simulation model of admission control is for dynamic service requests only. The total bandwidth is fixed at 27 Mbps for the simulation.
   

=

= =

=

=

=
The first simulation is to study the blocking probability (the percentage of rejected requests) and its relation with the number of SSs and the request rate (λ). Each simulation run lasts for 10–20 min. The results of blocking probability versus SSs with three different arrival rates (λ 1, 2, and 4) are illustrated in Figure 4.7. The second simulation is to perform a sensitivity analysis of blocking probability and requested bandwidth with λ 2 and SS 50, and the result is illustrated in Figure 4.8. The third simulation is to measure the channel utilization versus SSs (SS 20, 50, 100, and 200) with fixed parameters of bandwidth 2 and λ 1 for a 20-min simulation run. The results are illustrated in Figure 4.9.

100

90
80

70

Blocking (%)
60

50
40

30
20

10
0

40

60 80 100 120

140 160 180 200

Subscriber stations