An Enhanced EDD QoS Scheduler for IEEE 802.11e WLAN - Advances in Computational Science and Engineering - page 57

Fig. 8. Throughput in scenario 2 (VBR trac)
Fig. 9. Packet loss ratio in scenario 2 (VBR trac)
began to drop packets when the number of STAs was 6, while our scheduler
dropped packet when the number of STAs was 8. The proposed scheduler has
better performance than SETT-EDD, because SETT-EDD uses fixed mSI and
MSI and does not transmit all the packet when burst period is long. From the
simulation, we can see that our algorithm can support more VBR video trac
than the reference scheduler or SETT-EDD, because even small percentage of
packet drop in video trac results in sever degradation in video quality.
Figure 10 shows the average delay in scenario 2. The delay of the proposed
algorithm is the smallest among the three. This is because our scheduler aims
to empty the queue after the current transmission finishes.
Figure 11 shows the jitter performance. In the figure, the proposed schedul-
ing algorithm shows better performance than the other two schedulers. This is
because our scheduler aims to empty the queue when the current transmission
completes. If the queue is not empty, our scheduler begins next SI earlier to
reduce the delay.