Balancing uplink and downlink delay of VoIP traffic in WLANs using Adaptive Priority Control (APC)

TLDR: Adaptive Priority Control (APC) is introduced to balance the downlink and uplink delay of VoIP traffic at the MAC layer, by giving to the AP a higher transmission priority, which is adaptively decided according to the uplink and downlink traffic volume.

Abstract: In IEEE 802.11 wireless networks, the downlink delay rises as the number of VoIP nodes increases while the uplink delay remains small due to the same chance of media access between nodes and the Access Point (AP). This degrades the capacity and QoS of VoIP significantly. Therefore, we introduce Adaptive Priority Control (APC) to balance the downlink and uplink delay of VoIP traffic at the MAC layer, by giving to the AP a higher transmission priority, which is adaptively decided according to the uplink and downlink traffic volume. And, we verify through theoretical analysis that APC is an optimal method to balance the uplink and downlink delay.

Figures

Fig. 1. The Delay of VoIP traffic in DCF

Fig. 8. Simulation results of APC using various packetization interval

Fig. 6. Simulation results using two priority control methods with 20 ms packetization interval 64kb/s VoIP traffic

Fig. 7. Simulation results using two priority control methods with mixed (10 ms and 20 ms) packetization interval 64kb/s VoIP traffic

Fig. 2. Packet transmission in APC

Fig. 11. Retry rate of the AP (downlink traffic) in three approaches: Controlling CW, DCF and APC

Frequently asked questions

Q1. What are the contributions in "Balancinguplink anddownlink delay of voip traf c in wlans usingadaptive priority control (apc)" ?

In this paper, an adaptive priority control ( APC ) scheme is proposed to balance the downlink and uplink delay of VoIP traf c at the MAC layer, by giving to the AP a higher transmission priority which is adaptively decided according to the uplink and downlink trafc volume. 

Q2. What is the priority of the AP?

When the priority of the AP is very high (P ≥ CWM The authorN ), CW decreases to 1, then the AP transmits packetsalmost without the backoff. 

Q3. Why is the APC algorithm adaptive to the change in the number of active wireless nodes?

This is because when more than one packetization interval is used, the traffic volume of uplink and downlink becomes more asymmetric, and APC changes the priority of the AP adaptively to the change of the uplink and downlink traffic volume, while the semi-adaptive method is adaptive only to the change of the number of active wireless nodes. 

Q4. How can the authors estimate the number of active nodes?

The authors can calculate the number of the packets generated at the wireless nodes if the authors know the number of active wireless nodes, by dividing it by the packetization interval, and the authors can estimate the number of active wireless nodes by checking the received packets from wireless nodes. 

Q5. How have the authors implemented the APC algorithm?

The authors have implemented the APC algorithm using the QualNet simulator and have shown that APC balances the uplink and downlink delay effectively in VoIP traffic with various packetization intervals. 

Q6. What is the value of the 90th percentile value of the uplink?

The authors measured the 90th percentile value 3 of the uplink and2PhysicalLayer ConvergenceProtocol 3Generally, 90th percentile value is used for measuringQoS of VoIPbecauseit indicatesthe jitter of VoIP applications. 

Q7. How long does the backbone network delay?

The authors assumed the codec delay to be about 30-40 ms at both the sender and the receiver, and the backbone network delay to be about 20 ms. 

Q8. What is the reason why the retry rate of the AP in APC is lower?

The reason why the retry rate of the AP in APC is lower than that in DCF is that contention free transmission of the AP decreases the probability of packet collision. 

Q9. How many bytes did the authors add to the t value?

The authors added an additional 12 bytes to the1The t value is calculatedwith 160B (20ms packetization interval and G.711codec)payloadin 11Mb/s transmissionratepayload reflecting the overhead incurred by RTP [15] header. 

Q10. What is the average queue size of the nodes?

The queue size of the nodes is the number of packets generated in the application layer minus the number of packets sent to the AP. 

Q11. What is the average value of the uplink and downlink delay?

The authors plotted both the 90th percentile and average value of uplink and downlink delay because the 90th percentile value is a good measure of the capacity for the VoIP traffic, and the average value is used to check the balance of the uplink and downlink delay. 

Q12. What is the retry rate of the AP?

N i j =1 ( t s P The authorj − Rj )N i where, Qi is the estimated average queue size of wireless nodes at i th sampling time, N i is the number of active wireless nodes at i th sampling time, ts is sampling interval in milliseconds, P The authorj is the packetization interval of the wireless nodes j in milliseconds, and R j is the number of packets the AP received from the wireless node j . 

Q13. What other factors affect the performance of APC?

The authors also evaluated the performance of APC with other packetization intervals, which decide the VoIP packet size and affect the capacity for VoIP traffic. 

Q14. What is the definition of the jitter in a VoIP traffic?

The authors considered VoIP traffic with silence suppression, using the conversational speech model with double talk described in ITU-T P.59 [16]. 

Q15. What is the problem of the AP in CW control?

In order to use the ratio of the queue size of the AP and a node as the priority value of the AP, the AP needs to know the queue size of all wireless nodes as well as the queue size of itself. 

Q16. How many ms of downlink delay is used for measuring QoS?

downlink delay of voice packets, and defined the capacity of VoIP as the maximum number of wireless nodes so that the average of the 90th percentile of the one-way end-to-end delay for both direction does not exceed 60 ms.1) 

Q17. Why is the CW of the AP changed to 1=P?

The reason is that even if the CW of the AP is changed to 1=P of CWM The authorN , the transmission rate of the AP is not exactly P times because the backoff time is chosen randomly within the CW size.