Tham khảo tài liệu 'quality of service and resource allocation in wimaxfig part 12', kỹ thuật - công nghệ, cơ khí - chế tạo máy phục vụ nhu cầu học tập, nghiên cứu và làm việc hiệu quả | 266 Quality of Service and Resource Allocation in WiMAX Finally the 3rd test embraces the same full load segmentation principles as exemplified in the second test but we reduce traffic on rtPS connections down to 3Mbsec while 2 more BE connections with Mb sec load were set to simulate cameras with LD video flows. In this test we add new BE connection with Mbsec load to imitate control data transmission such as GPS location or image delivery. Thus in the last scenario we admit that 3 rtPS cameras were switched to less consuming mode with rare frame second rate video or black white color transmission but the rest of the total system load was allocated for new 2 cameras with LD video traffic transported over BE connections respectively. Moreover an additional Mbs BE connection produces a slight increase in the total system load for adequate analysis. The main simulation parameters of the considered tests are provided in the Table . It should be noted that we set the same values of the total system load and system bandwidth for most of the experiments except the final scenario with a small load overcome. The total data amount is re-allocated between the varied number of transport connections of defined QoS classes to model the variations of quality-selected video streams to compare network performance for the considered test scenarios. Summary throughput comparison is illustrated in Figure . Every graph on this figure correlates to summarized throughput values of a particular test. The whole simulation was carried out with support of WiMAX software module for NS-2 simulator designed by Chen Wang Tsai and Chang and proposed in Chen et. al 2006 . Simulation results analysis With much attention to HD video streams we should note that the higher date rate of about 7 Mbsec for UGS connection corresponding to superior video transmission levels out around the same value throughout the whole experiment. This fact intensely shows that for all cameras with
