Investigating the Quality of VoIP Traffic Over IPv6 Enabled Network

Transcription

1 A SUMMER INTERNSHIP PROJECT REPORT on Investigating the Quality of VoIP Traffic Over IPv6 Enabled Network Submitted By Borra Sahithi Admn No: 2013JE0619 B.Tech CSE Sai Sandeep Bhooshi M.S (EE) I Year, University of Kansas, U.S.A Under the Guidance of Dr. Rajib Ranjan Maiti Assistant Professor, IDRBT, Hyderabad July 2016

2 CERTIFICATE This is to certify that the summer internship project report entitled Investigating the quality of VoIP traffic over IPv6 enabled network. submitted to Institute for Development & Research in Banking Technology (IDRBT), Hyderabad is a bonafide record of work done by Borra Sahithi, Admn no. 2013JE0619, B.tech Computer Science and Engineering, Indian School of Mines, Dhanbad and Sai Sandeep Bhooshi, M.S (EE) I Year, University of Kansas, U.S.A, under my supervision from 9th May, 2016 to 9th July, 2016 and 24 th May, 2016 to 9 th July, 2016 respectively. Dr. Rajib Ranjan Maiti Assistant Professor, IDRBT, Hyderabad.

3 Acknowledgement We would like to express our sincere gratitude to the Institute for Development and Research in Banking Technology (IDRBT) and particularly to Dr. Rajib Ranjan Maiti who was our guide for this project. This opportunity of learning about VoIP over IPv6 enabled network was a boon to us as one rarely gets such exposure. We would like to add that this short period in IDRBT has added a different facet to our lives as this is a unique organisation being a combination of academics, research, technology, communication services, crucial applications etc,. We are extremely grateful to Dr. Rajib Ranjan Maiti for his advice, innovative suggestions and supervision. We thank him for introducing us to this excellent area of Voice over Internet Protocol for research. We are thankful to IDRBT for providing such an amazing platform for students, like us, to work in real application oriented research. We thank one and all who made this project successful either directly or indirectly. Borra Sahithi Project Trainee IDRBT, Hyderabad & Sai Sandeep Bhooshi Project Trainee IDRBT, Hyderabad

4 Table of Contents Abstract 1. Introduction 2. Simulation Model 2.1Network Scenario 1: Configuration details Step 1: Define the application Step 2: Construct the profiles Step 3: Assign the profiles to the nodes Step 4: Configuring source and destination Step 5: Setting the parameters to be measured Simulation Results 2.2 Network Scenario 2: Configuration details Step 1: Configuring the application and Profile Definitions Step 2: Configuring the nodes Results 3. Conclusion 4. References

5 Abstract Today Voice over IP (VOIP) is one of the fastest growing Internet application protocols. This paper presents a comparative approach to study the Quality of Service of VoIPv6 and VoIPv4 on integrated wireless LAN/WAN and wired network. The network model was simulated using OPNET Modeler 14.5 software. Different parameters that indicate the QoS like MOS, jitter, end to end delay, traffic send and traffic received are calculated and analyzed in both network scenarios.

6 Performance Measurement of VoIP over IPv6 enabled network 1. Introduction VoIP (Voice over Internet Protocol) is a protocol in TCP/IP network that enables voice communication over Internet rather than the traditional Public Switched Telephone Network (PSTN). With the deployment of the new version of the Internet Protocol on TCP/IP network, IPv6, it is important to investigate the deployability of VoIP that can make use of the services of IPv6. In this direction, the aim of this project is to investigate the performance of the VoIP on both IPv4 as well as IPv6 enabled networks. To aid our needs, a network simulator, called OPNET, is used to implement the proposed VoIP system on a synthetic network. We configure our network to have a set of routers, layer 2 switches, and a set of IP phone handsets that are available in the simulator. Both wired and wireless setup is used to connect IP phone handsets to the rest of the network whereas the other parts of the network are connected using wired connections (details setup can be found in the following of the report). Different QoS metrics like packet loss, jitter, and mean opinion score, are considered for performance analysis.

7 2. Simulation Model The simulation models of networks were setup using two scenarios. Scenario 1: consists of simulation of VoIP on wired network as shown in fig.1 (a) Scenario 2: consists of simulation of VoIP on wireless network (WLAN) as shown in fig.1 (b). Figure 1(a)

8 Figure 1(b) 2.1 Network Scenario 1: Configuration details To create a new scenario, say scen1, from File menu. Configure the scen1 as a Office network scenario, and complete the configuration which is essentially an empty scenario. To the empty scenario, the network nodes (a generic name for all the devices present in the network) are added from Object Palette. If the Object Palette doesn t appear automatically, then select them from the toolbar which is with the icon. The nodes added to this scenario are the switches (selected from ethernet model) and IP phones, router and cloud (selected from the sip model, used for VoIP). The 100BaseT link is used to link IP phones to switches, and switches to routers. This link represents an Ethernet connection operating at 100Mbps. The link PPP_DS1_int is used for the connection between the routers. This link is used to connect two nodes running Internet Protocol (IP). The nodes in the topology are auto assigned ipv6 addresses as follows: Go to Protocols icon above the toolbar and select IPv6. First enable IPv6 on all interfaces and then Auto Assign the IPv6 addresses. Same process was followed for assigning IPv4 addresses.

9 Configuring Applications and Profiles in Scenario1: To model network traffic on the constructed network, one technique is to model application traffic by setting up various application attributes. To configure a workstation or LAN to model the behaviour of a user or group of users, we need to describe their behaviour. A user s behaviour or profile can be described by the applications he or she uses and how long and often the applications are used throughout the day. An application can be described in terms of its actions, which are referred to as tasks in OPNET. OPNET provides global objects for defining profiles and applications. The advantage of using a global object is that once you have defined the profiles and applications, the global objects can be re-use across the entire topology. OPNET comes with pre-defined profiles and applications that may suit the behaviour you wish to describe. You may, however, wish to modify the existing definitions to suit your needs or even create new application and profile definitions. Here, the definitions were modified according to the need. Setting up nodes to use applications for traffic generation is a multi-step process. Since each step depends on definitions from the previous step in the procedure, it is important that the steps are completed in the order specified Step 1: Define the application. The first step for using applications for generating traffic at a node is to create and configure the application definitions. 1. Right click on the Application Definition Object in your project and select Edit Attributes from the pop-up menu. Edit the value of the Application Definitions attribute. 2. In the Applications Definitions Table, add a new row by changing the Rows value to 1. Give the application you will create a descriptive name such as VoIP. 3. To describe the behaviour of the application definition you just created, double-click in the Description field. For standard applications such as Voice, double-click on the attribute and a (Voice) Table pops up.

10 4. One important parameter is the Encoder Scheme which is set to G.711 as it is the codec which gives better performance than others. We also set the Type of Service to Interactive Voice (6). Another important attribute is the Voice Frames per Packet. In OPNET terminology, a voice frame is a collection of 32 voice samples of which each sample is 8 bits; i.e., each voice frame is 32 bytes in size. However, in the standard we are adopting, each VoIP packet has a payload of 160 bytes. Hence we set Voice Frames per Packet attribute to Step 2: Construct the profiles. Once the applications have been defined, the second step is to include the application definitions in profile definitions that are later deployed to your scenario. This step uses the application definitions that were configured in Step1. 1. Right click on the Profile Definition Object and select Edit Attributes. Set the Rows to 1 and enter a profile name. 2. You can use any name for the profile (eg.profile1). For the profile you created, use the application VoIP by setting number of rows to You can change the Start time offset and repeatability options as needed. Here we set the attributes as follows.

11 The start time offset [seconds] was set to constant(10). The start time defines when the applications start to run in the beginning of the simulation. The Duration [seconds] was set to end of simulation. This setting allows for the profiles to continue operating until the end of the simulation. Number of repetitions is set to 0 since we need only one voip call to be setup between source and destination Step3: Assign the profiles to the nodes. Once you have configured your applications and profiles, you are ready to deploy these profiles on individual nodes. Here the configurations of the routers and the switches are left default. 1. Select all the IP phones by using Ctrl key and select Edit Attributes ; In Applications, select Application: Supported Profiles and add the profile created(profile1). 2. And then select Application: Supported Services and set it to All. 3. Select the server and do the same steps as done to the IP phones Step4: Configuring source and destination Now, we can set the source phone and destination phone through which traffic should flow as follows: Right click the source node say node_0, go to edit attributes. Now, the destination is chosen in the Application: Destination Preferences. Edit the number of rows to 1 and use the application VoIP. Now, the destination node is selected by setting the Actual Name attribute to our destination node s name say Office Network.node_2. Now, by the above settings, we can ensure that the traffic flow for our scenario can take place. Next step is to measure the performance of our setup. For this, we have to first select the parameters to be measured. This can be as follows: Step5: Setting the parameters to be measured 1. Select the DES (Discrete Event Simulation) icon present above the tool bar and select Choose Individual Statistics.

12 2. Now, select Node Statistics (selecting this gives statistics of each node separately) and select the Voice Application, Voice Called Party and Voice Calling Party and in that, select the required parameters to be measured. Here we selected Jitter, Packet Delay Variation, Packet End-to-End Delay, Traffic sent and Traffic received. After selecting the parameters, run the simulation to produce the results by clicking the icon or by going into DES and selecting Configure/Run Discrete Event Simulation and set the time for which simulation has to run. The default options are usually sufficient except you may need to change the duration and the seed for some scenarios. The duration is the simulation time: for example you want to simulate two nodes sending packets to each other for 5 minutes. In real-time it may only take several seconds. The seed is the random seed used in the simulation. If you change the seed a different sequence of events may occur. For example, if you repeat the simulation of one scenario using the same seed, the results should be exactly the same. But if you change the seed, the results will most likely be (slightly) different, even with the same scenario. Here we set the simulation time for 20 minutes and seed value default. After Simulation completes, close the tab and right click on the plain area and select View Results to view the results Simulation Results: By default, all statistics are shown against time. For example, if the simulation runs for 5 minutes (300 seconds), the Traffic Sent statistic shows the measured traffic sent (in bits per second) at each time point (the time points are part of the simulation configuration). Note that the statistics are absolute values versus time. Often you only want to know the average for the entire simulation. In most cases you can estimate the average from the plot. However for more accurate results, you may try displaying an average (instead of absolute) value. You can do so by selecting the display type in the preview window on the right. Select As Is, and change it to average or another type. The following are the figures displaying different QoS parameters like MOS, Voice jitter, Voice traffic sent (packet/sec) and Voice traffic received (packet/sec). Here we setup traffic between node_0 and node_2.

13 Jitter is defined as a variation in the delay of received packets. At the sending side, packets are sent in a continuous stream with the packets spaced evenly apart. Due to network congestion, improper queuing, or configuration errors, this steady stream can become lumpy, or the delay between each packet can vary instead of remaining constant. The following graph shows the details. IPv6 IPv4 Traffic Sent and Received: The average voice traffic sent and received is presented in the following graph. The blue line shows the traffic sent by node_0 and the red line shows the traffic received by node_2. From the graph, we observe that as time increases, the packet loss is increasing more for IPv4 than for IPv6.

14 IPv6 (20 minutes) IPv4 (20 minutes) IPv6 (30 minutes) IPv4 (30 minutes) Mean Opinion Score (MOS): MOS is the test use in telephony networks to obtain the human user's view of the quality of the network. The MOS is expressed as a single number in the range 1 to 5, where 1 is lowest perceived audio quality, and 5 is the highest perceived audio quality measurement. Here, the MOS value in our simulation is slightly above 3 for both IPv4 and IPv6 networks.

15 IPv6/IPv4 2.2 Network Scenario 2: Configuration details Now, to create a new scenario, continuing the same project, just click on the Scenario and create a new scenario for the wireless network. The nodes to be selected for this scenario are: wlan_wkstn_adv (Fixed Node) from the wireless_lan_adv model. wlan_ethernet_router_adv (Fixed Node) from the same wireless_lan_adv model. This node acts as the access point helping to connect to the workstations wirelessly. Switches from the ethernet model. Routers, cloud and links from the sip model. All the links used are 100_Base_T except for connecting routers which used PPP_DS1. Using the above nodes and links, the model is constructed as shown in fig 1(b) Step 1: Configuring the application and Profile Definitions: Right click on the Application Definition Object, select Edit Attributes. In Application Definitions, set rows to 1 and select the pre-defined application for voip the Voice over IP Call(GSM Quality) and in the Description, set the attributes of Voice just the same as explained in the above scenario.

16 Now, in Profile Definition Object, go to Profile Configuration and set all the values as done in scenario Step 2: Configuring the nodes: Select all the workstations using Ctrl key and go to Edit Attributes. There, give the Profile name in Application: Supported Profiles and set Application: Supported Services to All. Now, each access point is taken and BSS Identifier is set to say 1, 2, 3 and 4 respectively for four of them. BSS Identifier helps identify the individual access point and associated clients. The BSSID is set as follows: Right click on the access point, go to Edit Attributes and select Wireless LAN and then, in Wireless LAN Parameters, set the BSS Identifier value. Now, each access point is associated with its own clients. This can be achieved by setting the BSSID of each client to the BSSID value of the associated access point. This ensures that the client with BSSID matching to any of the access points comes under the range of that particular access point. Now, Source and Destination nodes through which traffic should flow are set as explained in the above scenario Results Traffic sent and received: From the figure we conclude that for wireless network, IPv4 has more packet loss when compared to IPv6. And also we see that as the time increases, the packet loss increases as time increases.

17 IPv6 IPv4 MOS value: We observe that MOS value which indicates the overall voice quality is same for both IPv4 and IPv6 networks. IPv6 IPv4 Jitter Value:

18 From the above graphs, we can say that absolute value of jitter is more for IPv4 than IPv6 in wireless network. IPv6 IPv4 3. Conclusion In this study, we investigate the quality of VoIP traffic sent over a network that use either IPv4 or IPv6 with the help of network simulation using OpNet modeler 14.5 simulator. Our objective is to investigate the feasibility of availing the services of VoIP over IPv6, and the performance gain in IPv6 for the same. We build a network scenario where a set of handsets (both calling and called) are connected to a wired/wireless switch which in turn connected to a router, and the routers in calling and called sides are connected through Internet with default configuration present in the simulator. We measure a set of performance metrics like, end-to-end delay, voice jitter, packet delay variation, mean opinion score, traffic delivery ratio available in the simulator. Our analysis on traffic quality in terms of mean opinion score over IPv6 shows that it is no better than the quality observed in IPv4 based network. For relatively shorter duration (e.g. less than half an hour) of VoIP calls, packet loss is less in IPv6 network compared to IPv4 network when handsets are connected through wired networks. In general, packet loss slightly increases when handsets are connected through a wireless access point compared to that of wired connection as one can expect from the system. Interestingly, in wireless settings, it is observed that the packet loss is significantly less in IPv6 based networks

19 compared to that of IPv4. Further, jitter seems to be more inconsistent in IPv6 in wired settings, however, it is less in wireless settings. References: Quality of Service Analysis for VoIP over IPv4 and IPv6, Nusiba Osman Mohamed Alkhalefa, Amin Babiker A/Nabi Mustafa, International Journal of Science and Research (IJSR) ISSN (Online): Index Copernicus Value (2013): 6.14 Impact Factor (2013): Voip Performance Analysis over Ipv4 And Ipv6, Monjur Ahmed, Alan T Litchfield, Shakil Ahmed, Adnan Mahmood, Md. Emran Hossain Meazi, I.J. Computer Network and Information Security, 2014, 11, Published Online October 2014 in MECS ( DOI: /ijcnis Performance Analysis Of Voip Traffic Over Integrating Wireless Lan And Wan Using Different Codecs, Ali M. Alsahlany, International Journal of Wireless & Mobile Networks (IJWMN) Vol. 6, No. 3, June 2014 Performance Analysis Of Voip Over Wired And Wireless Networks : Network Implementation In Aden University, Mohsen Hussein Mohammed, Wafa a Nasser Abdullah, IJRET: International Journal of Research in Engineering and Technology eissn: pissn:

20 Measuring The Performance Of Voip Over Wireless LAN, Keshav Neupane, Victor Kulgachev, Andrew Elam, Sri Harsha Vasireddy, Hetal Jasani