Analysis of QoS parameters of VOIP calls over Wireless Local Area Networks

Analysis of QoS parameters of VOIP calls over Wireless Local Area Networks

Ayman Wazwaz, Computer Engineering Department, Palestine Polytechnic University, Hebron, Palestine,

Duaa sweity d.m.s_1989@hotmail.com

Dana Nimer

d_i_a_nimer@hotmail.com

Wafaa Nassar w.m.n.1989@hotmail.com

ABSTRACT

Voice over IP (VOIP) is one of the fastest growing Internet application protocols. In this study, we investigate the main factors that affect VOIP calls, and show the effect of mobility in different speeds on VOIP’s call quality. We show the consequences of increasing of the number of VOIP users in the network, and discuss 802.11e Standard that gives priority to VOIP calls. We assumed that the roaming service is enabled to avoid the interruption of calls during movement. Results show that without roaming, moving away from access points lead to the call drop after exiting from the access point coverage area , and by increasing the number of VOIP users, the quality of the voice decreases, and the parameters including delay and loss would increase, by increasing the speed of the mobile users, and the delay become larger.

Keywords: VOIP, Wireless LAN, MOS, OPNET, mobility, roaming.

1. INTRODUCTION

VOIP is a protocol that involves the delivery of voice and multimedia sessions over internet protocol networks. Main factors that affect the QoS parameters of VOIP call quality are:

 Movement speed during mobility.

 Number of VOIP users.

 Number of calls per hour.

 Background traffic.

We will measure the following parameters to compare these cases: delay, packet loss, the Mean Opinion Score (MOS) value, jitter, signal to noise ratio, end to end delay, and throughput.

Packet loss: is the total loss that occurs due to network congestion and late arrival.

End to end delay: defined as the total time it takes for a person, communicating another person at the other end.

Delay: mainly caused by network congestion which leads to a slow delivery of packets.

Jitter: the variation in transmission delay.

MOS value: The Mean Opinion Score (MOS) provides a numerical indication of the perceived quality from a voice codec during and after the transmission and compression of voice data [7].

Factors that can affect MOS include packet loss, jitter, and end-to-end delay.

MOS value is rated between 1 and 5, the higher value the better quality. It is preferred for a good VIOP call to have MOS greater than 3.6.

Signal to noise ratio: the ratio between the signal to noise power.

2. RELATED WORK

In [5], authors explain and measure the effect of the handover in WiFi networks for both intra and inter mobility for VoIP traffic. The study was oriented towards the assessment of the variation of the throughput and the packet delay jitter during the handover operation.

In [6], authors explain the basic concepts and issues of Wireless/Cellular network that can improve the QOS of a cellular WLAN.

in [2], this research project focused on the inherent limitations of wireless networks, especially in the areas of QOS and security, compared to wired Standards. Authors used VOIP as the multimedia benchmarking environment to explore the differences in the quality of service of a wireless and wired network and attempt to identify the main challenge areas for enhancing the QOS of VoIP in WLANs.

In [1], authors presented a survey on the voice capacity of an IEEE 802.11- based WLAN and the QOS enhancement mechanisms in the MAC layer. Only a limited number of voice connections can be supported in an 802.11 WLAN because of the overhead and the inherent inefficiency of the MAC protocol. Accurate voice capacity estimation is critical for effective and efficient admission control for Voice Over WLAN (VOWLAN).

The 13th International Arab Conference on Information Technology ACIT'2012 Dec.10-13

ISSN : 1812-0857

2

3. VOIP Quality Of Services :

3.1 QoS requirement:

QOS is a critical issue, because real-time applications are very sensitive to delay.

Therefore, QOS of VoIP is an import concern to ensure that voice packets are not delayed, lost or dropped during the transmission over the network. [3]

In wireless network, voice is digitized with the G.711 coding standard and transported at 64 Kbps, while G.711 is the main digital codec for quality voice service ,a number of more efficient codec’s are used for both cellular and voice applications.[3]

Many VoIP handset manufacturers use the G.711 codec because it provides superior quality. G.711 codec requires a 64 kilobits per second (kbps) data stream in each direction (uplink/downlink).

When encapsulation, overhead from Real time Protocol (RTP), UDP, IP and Wi-Fi headers are accounted for the bandwidth requirement. [4]

VoIP quality of service is measured based on different parameters like delay, jitter, packet loss. VoIP QOS is improved by controlling the values of these parameters to be within the acceptable range.

3.2 VoIP Over WLAN and QOS:

The IEEE 802.11 Standard specifies that a mobile device can only be associated with one Access Point (AP) at a time , so there is a risk that the communication is interrupted while performing the handover.

The duration of the period when the mobile device in unable to exchange data traffic via its old and new access points is often referred to as the handover latency or handover delay. If the mobile device experiences degraded signal quality in the communication with its AP, it will be at some point, it will be in the region of handover procedure. If the handover threshold value is configured so that a handover is triggered before connectivity with the current AP is lost, then the time to detect movement will not affect the total handover latency. To find candidate access points to re-associate with the mobile device will start to scan the different radio channels.

Because 802.11 networks were designed to carry data, not voice, 802.11 b/g standards have no QoS mechanisms built-in to tell the network to

prioritize voice packets over data, so a surge in network traffic can disrupt voice calls. With voice being a real time application, QoS control is essential and without it may lead to end-to- end delays, jitter, out of sequence errors, packet losses and contention. [4]

3.3 The Nature of Roaming in 802.11

Wireless local area network (WLAN), provides users the mobility freedom to move and roam around within the local coverage area. It simplifies the network by linking two or more computers or devices to enable communication between devices. In addition, WLAN simultaneously share resources within a broad coverage area, Using radio frequency (RF) technology, and it transmit and receive data over the air, without additional wiring. The mobility and roaming capabilities gives user a freedom to be connected everywhere and anywhere.

Voice devices, like Wi-Fi phones and PDAs, are extremely sensitive to delay and jitter. When these VOIP clients roam between buildings and floors they can experience disruptions and dropped calls. Meanwhile standard PC clients may experience slower data transfers while Web browsing may be disrupted when roaming. [7]

Wi-Fi devices are designed to go through a series of steps in order to establish a connection with a new AP whenever the current AP reaches an unacceptably low service level. This process is called 802.11 roaming.

4 NETWORK MODELING

The network in our research is infrastructure WLAN, where APs are connected using Ethernet links in a university campus; the network has five APs, workstations, router, and a server as shown in figure1. The parameters are shown in following table:

Parameter Value

Physical characteristic 802.11g

Data rate 54 Mpbs

Roaming capability Enable

Transmit power 0.005 W

power threshold -85 db

Number of scenarios 5 scenarios

Number of seeds 10 seeds

Duration 10 minutes

Types of traffic VOIP and background

traffic

Table 1: parameters used in simulation

Figure 1: Network Model using OPNET simulator

SCENARIOS

Using OPNET modeler, we made simulations for the different scenarios to measure the jitter, MOS, packet loss, and the end to end delay in the

modeled network.

4.1:

Mobility Speed and traffic received:

In this scenario, the workstation moves across five APs with different speed , so as we increase the mobility speed the probability of handover would increase and more traffic data would be lost.

As the speed of mobility increases, the delay also would increase as shown in Figure 2.

Figure 2: delay at different speeds

As the speed of mobility increases, the quality of voice call decreases (MOS), because as the power received to the mobile will vary multiple times, that causes instability of voice quality.

Figure 3: MOS value at different speeds

4.2 The effect of Background Traffic on QOS:

In this scenario, we applied different amount of traffic with different types of background traffic.

The amount of data drop increases as traffic increase according to the increase of the load, so some packet will be dropped due to congestion, as shown in figure 4.

Figure 4: Data Dropped for different background traffic types.

Figure 5, shows as small variation of the MOS value , which means that the background traffic has minimal effect on the packet loss end to end delay and the voice speech quality .

Number of seeds 10 seeds

Duration 10 minutes

Types of traffic VOIP and background

traffic

Table 1: parameters used in simulation

Figure 1: Network Model using OPNET simulator

SCENARIOS

Using OPNET modeler, we made simulations for the different scenarios to measure the jitter, MOS, packet loss, and the end to end delay in the

modeled network.

4.1:

Mobility Speed and traffic received:

In this scenario, the workstation moves across five APs with different speed , so as we increase the mobility speed the probability of handover would increase and more traffic data would be lost.

As the speed of mobility increases, the delay also would increase as shown in Figure 2.

Figure 2: delay at different speeds

As the speed of mobility increases, the quality of voice call decreases (MOS), because as the power received to the mobile will vary multiple times, that causes instability of voice quality.

Figure 3: MOS value at different speeds

4.2 The effect of Background Traffic on QOS:

In this scenario, we applied different amount of traffic with different types of background traffic.

The amount of data drop increases as traffic increase according to the increase of the load, so some packet will be dropped due to congestion, as shown in figure 4.

Figure 4: Data Dropped for different background traffic types.

Figure 5, shows as small variation of the MOS value , which means that the background traffic has minimal effect on the packet loss end to end delay and the voice speech quality .

Number of seeds 10 seeds

Duration 10 minutes

Types of traffic VOIP and background

traffic

Table 1: parameters used in simulation

Figure 1: Network Model using OPNET simulator

SCENARIOS

Using OPNET modeler, we made simulations for the different scenarios to measure the jitter, MOS, packet loss, and the end to end delay in the

modeled network.

4.1:

Mobility Speed and traffic received:

In this scenario, the workstation moves across five APs with different speed , so as we increase the mobility speed the probability of handover would increase and more traffic data would be lost.

As the speed of mobility increases, the delay also would increase as shown in Figure 2.

Figure 2: delay at different speeds

As the speed of mobility increases, the quality of voice call decreases (MOS), because as the power received to the mobile will vary multiple times, that causes instability of voice quality.

Figure 3: MOS value at different speeds

4.2 The effect of Background Traffic on QOS:

In this scenario, we applied different amount of traffic with different types of background traffic.

The amount of data drop increases as traffic increase according to the increase of the load, so some packet will be dropped due to congestion, as shown in figure 4.

Figure 4: Data Dropped for different background traffic types.

Figure 5, shows as small variation of the MOS value , which means that the background traffic has minimal effect on the packet loss end to end delay and the voice speech quality .

4 Figure 5: MOS value at different background

4.3 The effect of adding 802.11e to access points : This scenario included two access points, the first (AP1) did not support 802.11e, and the second (AP2) supported 802.11e, and the network contains 4 mobile nodes; the first 2 nodes can make VOIP calls and the receivers are moving from AP1 to AP2 and the next two mobile nodes have a best effort (no priority) traffic; and they move from AP1 to AP2 , when the receivers exist from AP1 area which did not support 802.11e the delay is equal for both data and VOIP calls , but when they move to AP2 which supported 802.11e the delay on best effort receiver is larger than the VOIP delay on the VOIP receiver;

this result came from the priority that AP2 gave to voice after t=200 the receiver pass to the coverage to AP2 , as explained is figure 6 and figure 7.

Figure 6: best effort delay

Figure7: 802.11e delay

4.4 Channels interference effect:

This scenario include 3 APs, we changed the entire channel configuration as shown in table 2:

Table 2: Access Pints channels’ numbers

When the channels are not overlapped, the signal to noise ratio will be larger than using adjacent or co-channel due to the high interference on the adjacent and co-channel as shown in fig 8.

Figure 8: signal to noise ratio

At the case of co-channels, the end to end delay is higher than the adjacent and the non overlap channels as shown in figure 9.

4 Figure 5: MOS value at different background

4.3 The effect of adding 802.11e to access points : This scenario included two access points, the first (AP1) did not support 802.11e, and the second (AP2) supported 802.11e, and the network contains 4 mobile nodes; the first 2 nodes can make VOIP calls and the receivers are moving from AP1 to AP2 and the next two mobile nodes have a best effort (no priority) traffic; and they move from AP1 to AP2 , when the receivers exist from AP1 area which did not support 802.11e the delay is equal for both data and VOIP calls , but when they move to AP2 which supported 802.11e the delay on best effort receiver is larger than the VOIP delay on the VOIP receiver;

this result came from the priority that AP2 gave to voice after t=200 the receiver pass to the coverage to AP2 , as explained is figure 6 and figure 7.

Figure 6: best effort delay

Figure7: 802.11e delay

4.4 Channels interference effect:

This scenario include 3 APs, we changed the entire channel configuration as shown in table 2:

Table 2: Access Pints channels’ numbers

When the channels are not overlapped, the signal to noise ratio will be larger than using adjacent or co-channel due to the high interference on the adjacent and co-channel as shown in fig 8.

Figure 8: signal to noise ratio

At the case of co-channels, the end to end delay is higher than the adjacent and the non overlap channels as shown in figure 9.

4 Figure 5: MOS value at different background

4.3 The effect of adding 802.11e to access points : This scenario included two access points, the first (AP1) did not support 802.11e, and the second (AP2) supported 802.11e, and the network contains 4 mobile nodes; the first 2 nodes can make VOIP calls and the receivers are moving from AP1 to AP2 and the next two mobile nodes have a best effort (no priority) traffic; and they move from AP1 to AP2 , when the receivers exist from AP1 area which did not support 802.11e the delay is equal for both data and VOIP calls , but when they move to AP2 which supported 802.11e the delay on best effort receiver is larger than the VOIP delay on the VOIP receiver;

this result came from the priority that AP2 gave to voice after t=200 the receiver pass to the coverage to AP2 , as explained is figure 6 and figure 7.

Figure 6: best effort delay

Figure7: 802.11e delay

4.4 Channels interference effect:

This scenario include 3 APs, we changed the entire channel configuration as shown in table 2:

Table 2: Access Pints channels’ numbers

When the channels are not overlapped, the signal to noise ratio will be larger than using adjacent or co-channel due to the high interference on the adjacent and co-channel as shown in fig 8.

Figure 8: signal to noise ratio

At the case of co-channels, the end to end delay is higher than the adjacent and the non overlap channels as shown in figure 9.

Figure 9: end to end delay .

The MOS at the non-overlap case is better than others, because of decreasing delay and jitter.

Figure 10: MOS value

4.5 Number of calls and VOIP QOS:

In this scenario, we increased the number of active VOIP calls gradually to show how the quality of voice will be affected , each call with duration 100 sec and the call rate is 10 calls per hour , then 50 calls per hour , then 150 calls per hour ,and finally 300 calls per hour.

Figure 11 shows that as the number of calls increases, the end to end delay increases, that refers to the amount of traffic that is transmitted using AP, which leads to more time to processing ,and as the traffic increases on the AP queuing delay increases.

Figure11: number of Calls and delay When we increase the number of calls per hour, we noticed that the call quality will decrease; so MOS value will decrease as shown in Figure 12.

Figure 12 : MOS and Number of call

5. Conclusions

5.1 Effect of mobility on QoS parameter During the movement between the APs, when the receiver and sender exist on the same AP, the quality of VOIP call is better than it when one of them move away and connect to another AP, because of roaming latency.

We can detect the result on the delay, and MOS, where the delay increases as one of them go far to another AP and the MOS value become smaller than when nodes are on the same AP.

5.2 Effect of the number of workstations on the QOS parameter on the AP

Figure 9: end to end delay .

The MOS at the non-overlap case is better than others, because of decreasing delay and jitter.

Figure 10: MOS value

4.5 Number of calls and VOIP QOS:

In this scenario, we increased the number of active VOIP calls gradually to show how the quality of voice will be affected , each call with duration 100 sec and the call rate is 10 calls per hour , then 50 calls per hour , then 150 calls per hour ,and finally 300 calls per hour.

Figure 11 shows that as the number of calls increases, the end to end delay increases, that refers to the amount of traffic that is transmitted using AP, which leads to more time to processing ,and as the traffic increases on the AP queuing delay increases.

Figure11: number of Calls and delay When we increase the number of calls per hour, we noticed that the call quality will decrease; so MOS value will decrease as shown in Figure 12.

Figure 12 : MOS and Number of call

5. Conclusions

5.1 Effect of mobility on QoS parameter During the movement between the APs, when the receiver and sender exist on the same AP, the quality of VOIP call is better than it when one of them move away and connect to another AP, because of roaming latency.

We can detect the result on the delay, and MOS, where the delay increases as one of them go far to another AP and the MOS value become smaller than when nodes are on the same AP.

5.2 Effect of the number of workstations on the QOS parameter on the AP

Figure 9: end to end delay .

The MOS at the non-overlap case is better than others, because of decreasing delay and jitter.

Figure 10: MOS value

4.5 Number of calls and VOIP QOS:

In this scenario, we increased the number of active VOIP calls gradually to show how the quality of voice will be affected , each call with duration 100 sec and the call rate is 10 calls per hour , then 50 calls per hour , then 150 calls per hour ,and finally 300 calls per hour.

Figure 11 shows that as the number of calls increases, the end to end delay increases, that refers to the amount of traffic that is transmitted using AP, which leads to more time to processing ,and as the traffic increases on the AP queuing delay increases.

Figure11: number of Calls and delay When we increase the number of calls per hour, we noticed that the call quality will decrease; so MOS value will decrease as shown in Figure 12.

Figure 12 : MOS and Number of call

5. Conclusions

5.1 Effect of mobility on QoS parameter During the movement between the APs, when the receiver and sender exist on the same AP, the quality of VOIP call is better than it when one of them move away and connect to another AP, because of roaming latency.

We can detect the result on the delay, and MOS, where the delay increases as one of them go far to another AP and the MOS value become smaller than when nodes are on the same AP.

5.2 Effect of the number of workstations on the QOS parameter on the AP

6 By increasing the number of workstations around the AP, it will affect the QoS parameters of VOIP calls, such increase means increasing the traffic received on the AP. So the load will increase but the throughput will be constant after period of time because the load is very high after that the drop will appear. By the increasing the number of workstation, the delay and jitter will increase gradually and the MOS will decrease and the quality of the VOIP call becomes worse.

5.3 Effect of increasing the number of access point:

By increasing the number of APs on a network, we will solve the dead zone problems in campuses. Also, by increasing the power coverage, mobile nodes will be connected to the same AP for longer time, less roaming, and less roaming latency.

5.4 Effect of call number on VOIP QOS:

Increasing the number of VOIP calls means increasing the load on the AP. So in period of time, the AP cannot serve all the calls, so part of them will get blocked and the loss will appear and the delay will increase. Also the MOS value will decrease.

5.5 Effect of increasing speed on VOIP QOS:

By increasing the speed of the mobile nodes, the traffic received will be less; because the mobile nodes will encounter poor coverage earlier. The delay will increase and the MOS will decrease so the quality of VOIP calls will get worse.

REFERENCES

[1] Lin Cai , yang xiao ,xuemin ,jon (" VoIP over WLAN: Voice capacity, admission control, QOS, and MAC"), International Journal Of Communication Systems, 2006; 19:491–508 [2] Mona Habib and Nirmala Bulus ("Improving QoS of VoIP over WLAN (IQ-VW)"), Project Research Paper December 2002

[3] H. Kazemitabar, A Survey on Voice over IP over Wireless LANs,(2010) World Academy of Science, Engineering and Technology (WASET),7. pp. 352-35

[4] Benjamin Miller, White Paper

Is It the Network? Solving VoIP Problems on a Wireless LAN, Global Knowledge, 2007 [5] Abderrahmane Lakas, Mohammed Boulmalf ("Experimental Analysis of VoIP over Wireless Local Area

Networks"), Journal of communications, vol. 2, no. 4, June 2007

[6] Jasmeet Singh ("Quality of Service in Wireless LAN Using OPNET MODELER"), Master thesis. University of Thapar, 2009

[7] Tin-Yu (Efficient Architecture and Handoff Strategy used for VoIP Sessions in SIP Based Wireless Networks) , Wireless Personal Communications: An International Journal , Volume 43 Issue 2, October 2007

Pages 201-214