In this lab, we have used VOIP technology to make voice calls. We have analyzed by implementing WFQ,CBWFQ and LLQ queuing techniques for improving the call quality.

Introduction

Quality of Service or QOS is used to increase the performance of voice application. End user can get voice call performance based on the QOS. It is a very critical implementation for voice over IP or VOIP based calls.QOS deals with reducing the delay and drop of packets compare with low priority traffic. If the delays are long, voice quality will be noisy and conversation will be very bad.QOS make sure the standard voice services by using existing resources. With this lab we have learned fragmentation with frame relay, traffic shaping techniques for improving the voice quality. In the first part of this lab, we will make voice call with FRF12 and analyze the voice quality. Then we will implement WFQ,CBWFQ and LLQ queuing techniques and will recognize suitable techniques for voice.

In our network topology, router 7 is working as frame-relay switching. Router 1 and Router2 are connected with two telephones.

Lab topology:

Figure 1 Lab 2 topology

We have configured VOIP peer between router 1 and router 2 with our lab instruction.

1. Configure voice over IP over Frame-Relay (FRF.12) and appropriate dial peers at Router 1 and Router 2 with the following information:

Committed burst size (Bc) = 12000 bits
Committed bit rate (CIR) = 64 kbps
Frame relay fragment = 1500 bytes
Voice codec: G.729

In this lab, we have used below information between router 1 and router 2:

Following table shows initial configuration between router 1 and router 2:

2. Test your configuration by making a call between the two phones. Note the voice quality.

With making a call between these phones, we have found voice quality is good.

3. Generate two ping traffic flows with 3000-byte packet size across PVC1. Make a voice call. Note that the voice quality deteriorates.

To increase the traffic flow, we have changed the packets size 3000 byte by using extended ping command. After that we make call between our phones and gets distort voice because of delay and jitter.

4. Configure the frame-relay fragmentation and traffic shaping at the serial interfaces to improve the voice quality (the fragment delay is required to be less than 10 ms).

To improve the voice quality, we have configured frame-relay fragmentation and traffic shaping between router 1 and router 2 serial interfaces:

Question 1: How do you choose appropriate fragment size and committed burst size (Bc) to implement the frame-relay fragmentation and traffic shaping? Why the voice quality is improved after the configuration?

In our lab requirements, fragment delay is less than 10 ms. So we have calculated the fragment size based on the following formula:

Fragment size (Maximum):

• Fragment_size = (0.01 sec) * CIR = (0.01 sec) * 64 kbps = 80 bytes

Parameters of Traffic Shapping:

• Burst size (Committed): Bc = 0.01 seconds * CIR = 640 bits

After these configuration, we have made voice calls and have get better voice quality. Voice quality have improved because of smaller fragmentation.

Question 2: Explain why FIFO queuing should not be used if fragmentation is configured.

Fragmentation helps to break large data traffic into smaller data traffic. For this voice traffic gets priority and have served faster. In the FIFO technology, if any large data entered into the queue then in that time if any voice traffic comes, then it needs to be wait until large data traffic finishes. There is no way to prioritize the voice traffic in FIFO techniques.

5. Set IP precedence of the voice traffic to 5. Generate two ping traffic flows with 3000-byte packet size across PVC1. Make a voice call. Note the voice quality.

In the type of service or TOS byte of Header, we have set IP precedence. IP precedence can identify class of services. Out of seven bits, left three digits are use in IP precedence. These values can be from zero to seven. Here larger number means higher priority. We have set IP precedence 5 and we make ping traffic with 3000 bytes in the PVC1.We have get voice quality good than the previous quality.

Following table shows the configuration between router 1 and router 2:

6. Configure a RTP priority queue for voice traffic. Generate two ping traffic flows with 3000-byte packet size across PVC1. Make a voice call. Note the voice quality

Following table shows the configuration between router 1 and router 2:

We have generated two ping traffic between router 1 and router 2 with 3000 bytes packet size. After that we have test voice calls between our phones. We have get voice quality is good than previous. It has happened because 27 kbps bandwidth is reserve for voice packets and voice packets has no need to wait in the queue.

Question 3: Determine the minimum bandwidth required for the RTP priority queue configuration.

We have configured voice traffic with RTP priority queue. Our size of voice packet is 66 bytes. So the minimum requirement of bandwidth is 8*66/0.02 or 26,400 bps or 26.4 kbps. We have used G729 codec and voice payload size is 20 bytes. We set our lab bandwidth is 27 kbps.

Question 4: Compare the voice qualities at Steps 4, 5, and 6, and explain the causes of quality differences.

To compare voice qualities between steps 4,5, and 6, we have found voice quality is worst in step 4.It has happened for voice call and ping is ready at a time, all packets are transfer in the same queue. So lots of packet are drop because of more queuing delay.

Get Help With Your Essay

If you need assistance with writing your essay, our professional essay writing service is here to help!
Find out more about our Essay Writing Service

Voice traffic has high priority when we use IP precedence 5 in step 5.Our voice and data traffic still use the same bandwidth. Data traffic still transfer even voice traffic arrives. So ping traffic transfers and voice traffic waits. For this, voice quality is not good because there is no bandwidth reservation for voice traffic.

In step 6,we have configured 27 kbps bandwidth in RTP priority queuing. This bandwidth is reserve for voice traffic. So voice packets always use this defined bandwidth and voice traffic has priority than ping traffic. So in this case, voice quality is better.

7. Configure three classes: VoIP, VoIP signaling, and default. Reserve bandwidth 25 kbps for VoIP class, and 8 kbps for VoIP signaling class, respectively. Do not assign priority queue to any class.

Three different classes has been created in this case. They are assign with fix bandwidth. We have used access-lists for voice traffic classification. Following table shows configuration between router 1 and router 2.

Question 5: Explain the differences and similarities between CBWFQ and WFQ.

CBWFQ can utilize bandwidth effectively compare to WFQ. It is actually the extended version of WFQ. During the congestion period, CBWFQ can guarantee the minimum bandwidth. It will switch when it gets more bandwidth again. In the CBWFQ, we can define different classes and each different classes we can assign separate bandwidth. The differences and similarities between WFQ and CBWFQ are as follows:

Similarities

CBWFQ has default traffic class but if we do not define this class then CBWFQ and WFQ has no difference in queuing techniques.

Differences

Based on the user define classes ,CBWFQ can traffic queuing but WFQ cannot queue traffic. For the traffic flow, CBWFQ can make sure specific bandwidth for it but WFQ can not guarantee that. Network administrator can use CBWFQ more flexibly. They utilize this CBWFQ with different priorities for different types of traffic.

Question 6: Based on the above configuration, what are the maximum and minimum bandwidths that are available for the default class?

The maximum bandwidth that are available for the default class was 64 kbits/s

Default class is use all available bandwidth like bandwidth = CIR = 64 kbit/s if there are no voice traffic
Because if there is no voice or voice signaling traffic then the default class is use all available bandwidth (bandwidth = CIR = 64 kbit/s).

The minimum bandwidth that are available for the default class is 31 kbit/s

But if for voice or voice signaling flow, the reserving bandwidth will be 25 kbits/s and 8 kbit/s respectively. Other will be available for default class. So, minimum bandwidth available for default class is:
- 64kbps – (25kbps + 8kbps) = 31 kbit/s

We have generated two ping traffic through PVC1 and at the same time we do voice call. Our packet size is 3000 bytes. We have found same voice quality like step 6.

8. Establish a voice call between the two phones. At the same time, generate two ping traffic flows with 3000-byte packet size between the two routers. Note the voice quality.

We have generated voice calls in our phones. At the same time between router 1 and router 2 we have generated 3000 byte ping traffic. We have found the acceptable voice quality. For voice traffic we have assign different class but there is no priority assign for this class. So, in the receiver end, both data and voice packets are receive similar way. So voice is not be clear because of delay introduce.

9. Configure a priority queue for VoIP class using LLQ.

We have configured priority queue with LLQ. Following table shows the configuration:

10. Make a voice call between the two phones. Note the voice quality.

After the configuration, we have make phone call with 3000 byte ping traffic in the same time. We found the improved voice quality.

Question 7: Explain why the voice quality is improved after the priority configuration.

Voice quality has improved for priority configuration. Previously we do not assign priority for different define classes of voice and voice signaling. So, voice quality is improved because of priority assigned. So for any ping traffic comes first it is served first. With the priority queue configuration, it has been resolved. So for any voice traffic arrival, it is use priority queue with allocated bandwidth.

Conclusion

In this lab, we have configured two routers and make phone call. We have observed the different call quality. Call quality depends on traffic flow. Traffic flows depends on different parameters like delay, jitter, loss of packets, etc. These hampers the quality of services. We have implemented three queuing discipline here like WFQ,CBWFQ,CBWFQ with LLQ. The objective of this lab is to analyze and improve the quality of voice service. We have found, the best voice quality when we have implemented different class for voice traffic and assign high priority value for voice traffic.

Page | 1