Evaluation and Comparison of Wired VoIP Systems to VoWLAN

ENSC 427: COMMUNICATION NETWORKS SPRING 2013 Evaluation and Comparison of Wired VoIP Systems to VoWLAN Project website: http://www.sfu.ca/~rza6/ensc427.html Group #15 Leslie Man 301048471 lpm4@sfu.ca Bill Lu 301049313 zbl@sfu.ca David Zhong 301077194 rza6@sfu.ca

Table of Content Abstract...3 Introduction...3 Background Information...3 History of VoIP...3 VoIP over Different Protocol...3 Voice over UDP...3 Voice over SCTP...4 SCTP VS UDP...4 LAN VS WLAN...4 Implementation...5 Simulation with ns-2...5 Wired VoIP System Topology...5 VoWLAN System Topology...6 Simulation Results and Discussion...6 AWK - Filtering Data in Trace File...7 Plotting with gnuplot... Error! Bookmark not defined. Throughput...7 Packet Loss...8 Latency...8 Packet Delay Variation...8 Results and Discussion... Error! Bookmark not defined. References:...14 Appendix A: NS-2-Code...14

Abstract VoWLAN is essentially a VoIP system implemented over Wi-Fi, which offers a significantly extended operational range. However, the extension presents trade-offs in stability, such as packet drop rate and additional delays. This project will simulate the performance of voice communication over wired connections to its Wi-Fi counterpart using NS2-2.35, in areas such as operational stability and operational range. Introduction This project is for better understanding of how VoIP works in underlying layers of the network. The wired VoIP and VoWLAN topologies are built by using NS2. In order to make the simulation more realistic as in the real world, the background traffic are added. The simulation is starting with UDP (User Datagram Protocol), and Stream Control Transmission Protocol (SCTP). To have better understanding of the difference between the VoIP on wired and wireless network, packet loss, throughput, delay and jitter are measured and analyzed. To better visualize the measured results, the simulation results are plotted. Background Information History of VoIP Voice over Internet Protocol (VoIP) is a communication protocol which runs over Internet Protocol (IP) networks. In 1974, IEEE published a paper titled "A Protocol for Packet Network Interconnection". In Aug 1974, Network Voice Protocol (NVP) first tested over Arpanet. The first VoIP application, Speak Freely, was released to public in 1991. Three years later, a free VoIP application for Linux, MTALK, appeared [1]. After forty years of developments, VoIP has become a very mature communication protocols. There are many popular VoIP applications in 20 th. VoIP was implemented in MSN in 2005. Google Voice service permitted VoIP connections through Gmail or Google Talk in 2009. The most recently use of VoIP is Facebook Inc. It launched free calling app for IPhone in January 2013. VoIP over Different Protocol Voice over UDP The UDP is one of important members of the Internet protocol suite. It uses a simple transmission without doing any error checks. Therefore, UDP is not suitable for applications which need error checking all the time. However, the advantage of UDP is less delay. Theoretically, packet loss must within certain range in order to deliver normal condition voice package. 3

Voice over SCTP Similar to UDP, Stream Control Transmission Protocol (SCTP) is a transport layer protocol. It contains both feature from UDP and TCP. It uses minimal message-oriented Transport Layer protocol as UDP. Also, it is in-sequence transport of messages with congestion control like TCP. Therefore, it is reliable and fast. SCTP VS UDP From message orientation point of view, SCTP is much more reliable than UDP because there is error detection in SCTP. To the ordered message service, both SCTP and UDP give unordered service. Compare to UDP s 16-bit ones-complement sum, SCTP uses 32-bit end-to-end checksum [2]. LAN VS WLAN LAN refers to a wired network while WLAN is used to refer to a wireless network. Theoretically, wireless 802.11g has 54Mbps, but wired Ethernet has 1000 Mbps. Therefore, voice data transfer in LAN is much faster than WLAN. 4

Implementation, Simulations, and Calculations Simulation with ns-2 The tests will be carried out in NS-2. AWK scripts will be used to parse the resulting simulation trace files, and evaluate the throughput, latency, jitter, and packet loss. Gnuplot will be used to graph the resulting data. The test topology will include two VoIP clients, sending two-way traffic to each other to simulate normal voice communication. The VoIP traffic will be running at the start of simulation. At a later time, background traffic will be added to evaluate its effects on the VoIP traffic. The background traffic will be sent via separated nodes/clients, while only sharing the transmission paths. Two protocols will be used for each of the wired and wireless scenarios: UDP and SCTP. Both protocols are available in NS-2 as source agents. LossMonitor sinks are attached to each node in the topology. Wired VoIP System Topology Figure 1 - Wired VoIP simulation in NAM 5

In our implementation, we decided to use G7.11 codec which is a commonly use audio codec nowadays. This codec uses a 64kbps bit rate and packet size of 160 bytes [3]. Node 0 & Node 1 will be our VOIP traffic and Node 3 & Node 4 will be our background. There are two-way traffic for both background and VoIP. The background traffic will be sending at a constant bit rate of 128kbps. The bandwidth limits for the routers (Node 4 & Node 5) is set to the same bitrate as the background traffic. Therefore once the background traffic start the links between the routers will be overloaded and we can see the impact it will have on the VOIP application. The topology for the wired setup is shown in Figure 1, with blue and red denoting two-way traffic between the VoIP clients. Red and yellow denotes two-way traffic between the background sources. The queue buildup exists at the center link, serving as a bandwidth bottleneck. VoWLAN System Topology Figure 2 - VoWLAN simulation in NAM The wireless test setup is similar to the wired topology, with the same traffic route shown in figure 2. Mobile node 3 is sending VoIP traffic (160 bytes with 0.020s interval) to node 0. The traffic arrives at the base station node, travels through node 1 and is received by the client at node 0. Node 0 will send VoIP traffic via the same route back to mobile node 3. This will ensure the performances between WLAN and wired tests can be accurately compared. The base station acts as a receiver for wireless traffic from nodes 3 and 4. The bandwidth bottleneck for the wireless setup is the duplex link between node 1 and the base-station, which will generate dropped packets upon traffic exceeding the transmission speed of 64kbps. 6

AWK - Filtering Data in Trace File AWK is a data extraction tool in Ubuntu. Trace file is an output data file from ns-2 simulation. Its file extension is.tr. We only use several types of events in trace file. They are + represent enqueue, - represent dequeue, r represents receive, and d represents drop. The normal trace file structure is shown as following Table-1: 1 2 3 4 5 6 7 8 9 10 11 12 abbr. Timestamp Source Node Destination Node Packet Type Packet Size Flags Flow ID Source Address Table-1: Trace File Normal Event Structure Destination Address Sequence Number 1 2 3 4 5 6 7 8 9 10 11 12 abbr. Time Node ID X Coordinate Y Coordinate Trace Name Reason Event Identifier Packet Type Table-2: Trace File Normal Event Structure Packet Size Time To Send Data Unique Packet ID Destination MAC Address AWK scripts will be used to parse the trace files created by the simulations, and perform calculations to obtain the performance values. The trace files under the parameter trace-all will contain all necessary information to obtain the throughput, latency, jitter, and packet loss. AWK Performance Calculations Throughput Throughput is the average rate of successful data pass over a communication channel. It is measured in bytes/sec. For example, from client Node 0 to client Node 1, both clients connect to server Node A and Node B, the throughput of VoIP refers to the total amount of voice data transfer between Node 0 and Node 1. In the trace file, r represents receive in normal and wireless event, which can be used to tack throughput. The formula is shown as following: ( ) The instantaneous throughput will generate a graph showing the amount of information received by the destination node over each second. This is useful for evaluating the immediate effects of the background traffic on the pre-existing VoIP traffic [6]. The average throughput will produce a single value showing the average throughput for the entire duration of the simulation. The formula is as following: 7

Packet Loss Packet loss happens when packets of data travelling across a computer network never reaches its destination. It occurs when packets enter a queue when it is full. This is for the case of Drop-Tail queuing method, which will drop the last packet attempting to enter the queue upon reaching queue limit. For this project, the cumulative packet loss will be evaluated. The cumulative packet loss tells the total numbers of packets dropped throughout he simulation. In the trace file, the d flag represents drop in normal and wireless events, which can be used to count and calculate the instantaneous packet loss and cumulative packet lost. The cumulative packet loss can be plotted on a graph progressively over each second, and will also give summed total of packet loss at the end of the runtime. This will allow easy comparison between various scenarios and protocols. Latency Latency is a measure of time delay experienced in a system. In our simulation, the delay is measured by taking the time difference between when a packet is sent from the source node, and when it reaches its destination. The latency (end-to-end delay) will be measured for each packet sent from a VoIP client, and reaches its partner. The packet will be tracked using its packet ID as it travels through each node in the network. The instantaneous latency formula is given as: Instantaneous Latency = Receive time of destination node Send time of source node The instantaneous latency can be presented graphically, where the end-to-end delay (in seconds, y-axis) for a specific packet is posted at the time it was received by the destination node (x-axis) The Average Latency is a single value, calculated from the formula given below: Average Latency = cumulative total of instantaneous latency / simulation runtime The average latency produces a single value which can be easily used to carry out performance comparisons between different scenarios. Packet Delay Variation Jitter is an informal name for IP packet delay variation (IPDV), but it is often use in electronics and telecommunication. Jitter is the undesired deviation from true periodicity of an assumed periodic signal in computer network. As an example, say packets are transmitted every 20 ms. If the 2nd packet is received 30 ms after the 1st packet, IPDV = 10 ms. This is referred to as dispersion. If the 2nd packet is received 10 ms after the 1st packet, IPDV = +10 ms. This is referred to as clumping [4]. Instantaneous Jitter = Current Latency Previous Latency 8

Instantaneous Jitter is graphed over the simulation period, and each data point will be mapped to its respective time in the Instantaneous Latency graph. This allows examination of the stability of the network as the traffic pattern constantly changes throughout the simulation. The average jitter can be calculated as: Instantaneous Jitter = sum(current Latency Previous Latency) / simulation runtime The produced value can be easily used to evaluate each scenario. Results and Discussion The left side is wired VoIP results, and the right side is WoWLAN simulation results. The performance between wired and WLAN VoIP networks is compared for UDP and SCTP protocols. The background traffic initiates at 2 seconds into the simulation. UDP - Throughput Figure - Throughput - Wired VoIP VS VoWLAN The resulting throughput between wired and wireless VoIP networks under UDP is similar. This is due to the mobile nodes being placed relatively close to the base station, so that the signal is strong enough to mirror a wired connection. As seen from the figures, without any background traffic, both VoWLAN and wired networks easily reach the throughput of 60kbps, guaranteeing reliable voice traffic between the clients. UDP - Packet Lost 9

Figure - Packet Lost - Wired VoIP VS VoWLAN The packet loss is significantly more for wired than for wireless under the UDP protocol. This does not match the theoretical results, in which the instability of wireless packet transfers will result in higher packet loss. This result is due to the queuing type of the duplex links, as there are three separate queues between sender and receiver, which results in undesired queue stacking characteristics resulting in significant amount of packet loss over long periods of times. UDP - Latency Figure - Latency - Wired VoIP VS VoWLAN The latency is shorter for the wireless network than the wired network under the UDP protocol. This is due to the mobile nodes being in too close proximity of the base station, which resulted in better performance for the wireless signals than for the preset delays existing in the duplex links. However, the fluctuation of the latency is significantly higher for the VoWLAN, which is accurate to real-world scenarios. UDP - Jitter 10

Figure - Jitter - Wired VoIP VS VoWLAN As mentioned in the latency test, the jitter is higher for the wireless network than for the wired. This is also theoretically correct, as wireless signals are prone to interference and signal instability. SCTP Throughput Figure - Throughput - Wired VoIP VS VoWLAN The throughput under the SCTP protocol remains the same for the wired VoIP network, as the background traffic caps the transmission speed, and the Drop-Tail queuing method becomes a race condition between VoIP and background traffic. However, the wireless network benefits from the error detection of the SCTP protocol, which in itself attributes for some of the extra throughput. 11

SCTP Latency Figure - Latency - Wired VoIP VS VoWLAN Similar to the UDP protocol tests, the SCTP results in higher overall latency, but displays the same characteristics for the VoIP latency in the presence of background traffic. The SCTP protocol gives a slightly higher latency due to its error detection functionalities. This also results in higher jitter for periods when errors occur within the traffic. SCTP Jitter Figure - Jitter - Wired VoIP VS VoWLAN The following tables summarizes the results of the simulation. The average values can be used to compare the performances between wired and wireless networks for both UDP and SCTP protocols. Wired VoIP UDP SCTP Throughput (kbps) 21.6 35.9431 Latency(s) 0.0445072 0.138427 Jitter(s) 0.000359982 0.00035492 Table 3 Average Simulation Results of Wired VoIP 12

WLAN VoIP UDP SCTP Throughput (kbps) 23.76 30.752 Latency(s) 0.0496553 0.142328 Jitter(s) 0.000107147 0.0031937 Table 4 Average Simulation Results of VoWLAN Conclusion The purpose of the project is to evaluate the performance difference between wired and wireless VoIP networks. The scenarios are created to include a variety of real-world cases, such as for both UDP and SCTP protocols, and with the addition of background traffic [5]. The results are conclusive for the cases of throughput, packet loss, latency, and jitter. The wireless nodes are affected by its proximity to the base station, acting as the receiver. In close proximity, the wireless nodes will perform better than the wired network, which had a preset delay of 5ms for each of the duplex links. However, the latency and jitter is unstable for wireless networks, especially as the distance increases between the mobile nodes and the base station. 13

References: [1] Wikipedia. Voice Over IP, Available: http://en.wikipedia.org/wiki/voice_over_ip [Accessed: April 8, 2013] [2] Computer Networks. VoIP performance over UDP and SCTP in NS2, Available: https://sites.google.com/site/networksprojectwiki/bit10/compnetworks/voip-performance-over-udp-andsctp-in-ns2 [Accessed: April 15, 2013] [3] CISCO, Voice Over IP - Per Call Bandwidth Consumption, Available: http://www.cisco.com/en/us/tech/tk652/tk698/technologies_tech_note09186a0080094ae2.shtml [Accessed: April 13, 2013] [4] Wikipedia. Packet delay variation, Available: http://en.wikipedia.org/wiki/packet_delay_variation [Accessed: April 8, 2013] [5] Hole, D.P.; Tobagi, F.A.;, "Capacity of an IEEE 802.11b Wireless Lan supporting VoIP," Communications, 2004 IEEE Internation Conference on, vol.1, no., pp. 196-201 [6]T. Issaraiyakul and E. Hossain Post processing NS2 Result using NS2 Trace Trace file format [Online] Available FTP: http://www.ns2ultimate.com/post/2496927327/post-processing-ns2-resultusing-ns2-trace-trace Appendix A: NS-2-Code Wired VoIP using UDP: # starting the simulator set ns [new Simulator] #dynamic routing #$ns rtproto DV set nf [open lanudp-out.nam w] $ns namtrace-all $nf #open tracefile set nd [open lanudp-out.tr w] $ns trace-all $nd proc finish { { 14

global ns nf $ns flush-trace close $nf exec nam lanudp-out.nam & exit 0 #declare colors $ns color 1 blue $ns color 2 red $ns color 3 purple $ns color 4 orange #create node set n0 [$ns node] set n1 [$ns node] set n4 [$ns node] set n5 [$ns node] set r1 [$ns node] set r2 [$ns node] # shapes for the routers $r1 shape box $r2 shape box #label $n0 label "0" $n1 label "1" $n4 label "4" $n5 label "5" $r1 label "router" $r2 label "router2" #color nodes $n0 color blue $n1 color purple $n4 color red $n5 color orange # how connections are established and bandwidth//delay for our lines # G7.11 codec $ns duplex-link $n0 $r1 64kb 5ms DropTail # background traffic $ns duplex-link $n1 $r1 128kb 5ms DropTail 15

#G7.11 codec $ns duplex-link $n4 $r2 64kb 5ms DropTail #background traffic $ns duplex-link $n5 $r2 128kb 5ms DropTail # link between routers $ns duplex-link $r1 $r2 128kb 5ms DropTail #orientation $ns duplex-link-op $n0 $r1 orient right-down $ns duplex-link-op $n1 $r1 orient right-up $ns duplex-link-op $r1 $r2 orient right $ns duplex-link-op $r2 $n4 orient right-up $ns duplex-link-op $r2 $n5 orient right-down #monitor queues $ns duplex-link-op $n0 $r1 queuepos 0.5 $ns duplex-link-op $r1 $r2 queuepos 0.5 $ns duplex-link-op $r2 $n4 queuepos 0.5 #$ns queue-limit $n0 $r1 3 #$ns queue-limit $n4 $r2 3 #create source and traffic ####voip #Create a UDP agent and attach it to the node set source0 [new Agent/UDP] $ns attach-agent $n0 $source0 #set flow id easier to trace afterwards $source0 set fid_ 1 #create a cbr traffic and attach to source set traffic0 [new Application/Traffic/CBR] #set packet size $traffic0 set packetsize_ 160 #packet size * 8 / speed $traffic0 set interval_.02 # Attach traffic source to the traffic generator $traffic0 attach-agent $source0 $source0 set class_ 1 #Connect the source and the sink set sink4 [new Agent/LossMonitor] 16

$ns attach-agent $n4 $sink4 $ns connect $source0 $sink4 #background #create udp agent set source1 [new Agent/UDP] $ns attach-agent $n1 $source1 #flowid $source1 set fid_ 3 #cbr traffic set traffic1 [new Application/Traffic/CBR] $traffic1 set packetsize_ 320 #packet size * 8 / speed $traffic1 set interval_.02 # Attach traffic source to the traffic generator $traffic1 attach-agent $source1 $source1 set class_ 3 #Connect the source and the sink set sink5 [new Agent/LossMonitor] $ns attach-agent $n5 $sink5 $ns connect $source1 $sink5 #2nd voip #create a udp agent set source4 [new Agent/UDP] $ns attach-agent $n4 $source4 $source4 set fid_ 2 set traffic4 [new Application/Traffic/CBR] $traffic4 set packetsize_ 160 #packet size * 8 / speed $traffic4 set interval_.02 # Attach traffic source to the traffic generator $traffic4 attach-agent $source4 $source4 set class_ 2 #Connect the source and the sink set sink0 [new Agent/LossMonitor] 17

$ns attach-agent $n0 $sink0 $ns connect $source4 $sink0 #background set source5 [new Agent/UDP] $ns attach-agent $n5 $source5 $source5 set fid_ 4 set traffic5 [new Application/Traffic/CBR] $traffic5 set packetsize_ 320 #packet size * 8 / speed $traffic5 set interval_.02 # Attach traffic source to the traffic generator $traffic5 attach-agent $source5 $source5 set class_ 4 #Connect the source and the sink set sink1 [new Agent/LossMonitor] $ns attach-agent $n1 $sink1 $ns connect $source5 $sink1 ####simulation $ns at 1 "$traffic0 start" $ns at 1 "$traffic4 start" $ns at 2 "$traffic1 start" $ns at 2 "$traffic5 start" $ns at 10.0 "$traffic1 stop" $ns at 10.0 "$traffic5 stop" $ns at 10.0 "$traffic0 stop" $ns at 10.0 "$traffic4 stop" $ns at 10.0 "finish" $ns run Wired VoIP using SCTP: # starting the simulator set ns [new Simulator] #dynamic routing #$ns rtproto DV 18

set nf [open lansctp-out.nam w] $ns namtrace-all $nf #open tracefile set nd [open lansctp-out.tr w] $ns trace-all $nd proc finish { { global ns nf $ns flush-trace close $nf # exec nam rtpout.nam & exit 0 #set the colors of the packets being sent $ns color 1 blue $ns color 2 red $ns color 3 purple $ns color 4 orange #create nodes set n0 [$ns node] set n1 [$ns node] set n4 [$ns node] set n5 [$ns node] set r1 [$ns node] set r2 [$ns node] # shapes for the routers $r1 shape box $r2 shape box #label $n0 label "0" $n1 label "1" $n4 label "4" $n5 label "5" $r1 label "access point" $r2 label "router2" #color the nodes $n0 color blue $n1 color purple 19

$n4 color red $n5 color orange # how connections are established and bandwidth//delay for our lines # G7.11 codec $ns duplex-link $n0 $r1 64kb 5ms DropTail # backgroung $ns duplex-link $n1 $r1 128kb 5ms DropTail #G7.11 codec $ns duplex-link $n4 $r2 64kb 5ms DropTail #background $ns duplex-link $n5 $r2 128kb 5ms DropTail # link between routers $ns duplex-link $r1 $r2 128kb 20ms DropTail #orientation $ns duplex-link-op $n0 $r1 orient right-down $ns duplex-link-op $n1 $r1 orient right-up $ns duplex-link-op $r1 $r2 orient right $ns duplex-link-op $r2 $n4 orient right-up $ns duplex-link-op $r2 $n5 orient right #to monitor the queues $ns duplex-link-op $n0 $r1 queuepos 0.5 $ns duplex-link-op $r1 $r2 queuepos 0.5 $ns duplex-link-op $n4 $r2 queuepos 0.5 #set queues limit so the queues will not be that long so the packet created will be sent at the same time $ns queue-limit $n0 $r1 3 $ns queue-limit $n4 $r2 3 #Create a sctp agent and attach it to the node set source0 [new Agent/SCTP] $ns attach-agent $n0 $source0 $source0 set fid_ 1 #create a traffic of cbr set traffic0 [new Application/Traffic/CBR] #set packet size $traffic0 set packetsize_ 160 #packet size * 8 / speed $traffic0 set interval_.02 20

# Attach traffic source to the traffic generator $traffic0 attach-agent $source0 $source0 set class_ 1 #create sink set sink4 [new Agent/LossMonitor] $ns attach-agent $n4 $sink4 #create 2nd voip agent set source4 [new Agent/SCTP] $ns attach-agent $n4 $source4 $source4 set fid_ 2 #create a traffic for 2nd voip agent set traffic4 [new Application/Traffic/CBR] $traffic4 set packetsize_ 160 #packet size * 8 / speed $traffic4 set interval_.02 # Attach traffic source to the traffic generator $traffic4 attach-agent $source4 $source4 set class_ 2 # set sink0 [new Agent/LossMonitor] $ns attach-agent $n0 $sink0 #Connect the source and the sink #$ns connect $source $sink #connect both sources $ns connect $source0 $source4 #return $traffic ############################ BACKGROUND AGENT AND TRAFFIC #background 1 agent set source1 [new Agent/SCTP] $ns attach-agent $n1 $source1 $source1 set fid_ 1 set traffic1 [new Application/Traffic/CBR] $traffic1 set packetsize_ 320 #packet size * 8 / speed $traffic1 set interval_.02 # Attach traffic source to the traffic generator $traffic1 attach-agent $source1 21

$source1 set class_ 3 set sink5 [new Agent/LossMonitor] $ns attach-agent $n5 $sink5 #background 2 agent set source5 [new Agent/SCTP] $ns attach-agent $n5 $source5 $source5 set fid_ 2 set traffic5 [new Application/Traffic/CBR] $traffic5 set packetsize_ 320 #packet size * 8 / speed $traffic5 set interval_.02 # Attach traffic source to the traffic generator $traffic5 attach-agent $source5 $source5 set class_ 4 set sink1 [new Agent/LossMonitor] $ns attach-agent $n1 $sink1 #Connect the source and the sink #$ns connect $source $sink #connect the both source $ns connect $source1 $source5 #run simulation $ns at 1 "$traffic0 start" $ns at 1 "$traffic1 start" $ns at 2 "$traffic4 start" $ns at 2 "$traffic5 start" $ns at 60 "$traffic0 stop" $ns at 60.0 "$traffic1 stop" $ns at 60.0 "$traffic4 stop" $ns at 60.0 "$traffic5 stop" $ns at 60.0 "finish" $ns run AWK code (for Wired VoIP) Throughput: #throughput bytes receieved in destination node in 1s 22

BEGIN { ###decare variables sim_time = 0.0; bytes_counter=0; counter =0; tracker =0; { #get current simulation time sim_time = $2; #get simulation time track = $2; #tracker = track - counter; #if receive update bytes counter if ($1=="r" && $3 == "5" && $4 == "2" ){ #get the packet size and * 8 to convert to 8 and divide by a 1000 to get the unit "k" bytes_counter+=$6*8/1000; print "bytes:" bytes_counter #use to check if the data point exceed 1s tracker = track - counter; #if it exceed 1s if (tracker >= 1){ printf("%f %f\n", sim_time, bytes_counter) > "throughput"; #increase counter counter ++; #reset the checker tracker = 0; print "counter:" counter; #use to calculate average throughput total_throughput += bytes_counter; #reset bytes counter bytes_counter = 0; END { #calculate average throughput ave_throughput = total_throughput/counter; print "ave throughput:" ave_throughput; print("done"); Packet Lost: #packetlost number of packet lost from the voip nodes 23

BEGIN { #declare var time2 = 0.0; num_packet=0; bytes_counter=0; send_flag =0; send_flag2=0; { #get the current simulation time time2 = $2; if ($1=="+" && ($3 == "0" && $4 == "4") && $8 == 1 && send_flag == 0) { send_flag = 1; if ($1=="r" && ($3 == "0" && $4 == "4") && $8 == 1 && send_flag ==1){ send_flag =2; if ($1=="+" && ($3 == "4" && $4 == "5") && $8 == 1 && send_flag ==2){ send_flag=3; if ($1=="d" && ($3 == "4" && $4 == "5") && $8 == 1 && send_flag ==3){ #counter for the numbers of packet drops num_packet++; printf("%f %f\n", time2, num_packet) > "packetlost"; print "number of packet lost : " num_packet; print "packet lost time :" time2; send_flag = 0; if ($1=="+" && ($3 == "2" && $4 == "5") && $8 == 2 && send_flag2 == 0) { send_flag2 = 1; if ($1=="r" && ($3 == "2" && $4 == "5") && $8 == 2 && send_flag2 ==1){ send_flag2 =2; if ($1=="+" && ($3 == "5" && $4 == "4") && $8 == 2 && send_flag2 ==2){ send_flag2 =3; if ($1=="d" && ($3 == "5" && $4 == "4") && $8 == 2 && send_flag2 ==3){ #counter for the numbers of packet drops num_packet++; printf("%f %f\n", time2, num_packet) > "packetlost"; print "number of packet lost 2 : " num_packet; print "packet lost time :" time2; send_flag2 = 0; if (num_packet == 0) { #if there are no packet lost 24

printf("%f %f\n", time2, num_packet) > "packetlost"; print "no packet lost"; END { print("done"); Latency: #latency difference in the time receieve from packets BEGIN {highest_packet_id = 0; sendflag =0; flowid=0; count=0; { #######--------------- # $1 = action $2 = time $3 = from node $4 = destionation node $12 = packetid ##-------------######### sim_time = $2; packet_id =$12; #cheack if there is a change in packet id and save it if (packet_id > high_packet_id && $3 == "0" ){ highest_packet_id = packet_id; if ($1 == "+" && $3 == "0" && $4 == "4" ){ # if ($1 == "-" && $3 == "0" && $4 == "2" ){ #save the send time of specific packet id send_time[highest_packet_id] = sim_time; #print "send_time" send_time[highest_packet_id]; # if ($1 == "r" && $3 == "4" && $4 == "5" && $8 == "1"){ #if ($1 == "r" && $3 == "0" && $4 == "4" && $8 == "1"){ if ($1 == "r" && $3 == "5" && $4 == "2" && $8 == "1"){ #if ($1 == "r" && $3 == "3" && $4 == "1" && $8 == "1"){ #get the packet id for that condidtion highest_packet_id =$12; #get the receieve time & start of that packet id rcv_time[highest_packet_id] = sim_time; start = send_time[highest_packet_id]; end = rcv_time[highest_packet_id]; #calculate latency duration = end - start; printf("%f %f\n", sim_time, duration) > "latency"; #some up the latency for calculation average latency count++; duration_ave += duration; 25

#else if ($1 == "d" && $3 == "4" && $4 == "5" && $8 =="1"){ #if there are any drops else if ($1 == "d" && $3 == "2" && $4 == "3" && $8 =="1"){ print"drop"; END { #calculate average latency latency_ave = duration_ave/count; print "ave latency:" latency_ave; print("done"); Jitter: #jitter difference in e2e delay BEGIN {highest_packet_id = 0; flowid=0; count=0; { #######--------------- # $1 = action $2 = time $3 = from node $4 = destionation node $12 = packetid #-------------######### #get simulation time sim_time = $2; #get packet id packet_id =$12; #if there is a change in packet id in node 0 if (packet_id > high_packet_id && $3 == "0"){ highest_packet_id = packet_id; #if node 0 have an enqueue if ($1 == "+" && $3 == "0"){ #save the current simulation time to send time of that specific packet id send_time[highest_packet_id] = sim_time; if ($1 == "r" && $3 == "5" && $4 == "2" && $8 == "1"){ #get current packet id highest_packet_id =$12; #save the receieve and send time of that packet id rcv_time[highest_packet_id] = sim_time; start = send_time[highest_packet_id]; end = rcv_time[highest_packet_id]; #calcuate latency duration = end - start; #calculate jitter (difference between the new and old latency) respone = duration - duration_old; printf("%f %f\n", sim_time, respone) > "jitter"; 26

print "packetid:" highest_packet_id; print "send time:" start; print "end time:" end; print "new duartion:" duration; print "old duration:" duration_old; print "jitter:" respone; #save the old latency duration_old = duration; ###### sum up the jitter for calculating ave jitter respone_ave += respone; count++; else if ($1 == "d" && $3 == "4" && $4 == "5" && $8 == "1"){ print"drop"; END { #to calculate average jitter ave_jitter = respone_ave/count; print "ave jitter:" ave_jitter; print("done"); 27

VoWLAN Topology in ns2: VoWLAN UDP: set opt(chan) Channel/WirelessChannel ;# channel type set opt(prop) Propagation/TwoRayGround ;# radio-propagation model set opt(netif) Phy/WirelessPhy ;# network interface type set opt(mac) Mac/802_11 ;# MAC type set opt(ifq) Queue/DropTail/PriQueue ;# interface queue type set opt(ll) LL ;# link layer type set opt(ant) Antenna/OmniAntenna ;# antenna model set opt(ifqlen) 50 ;# max packet in ifq set opt(nn) 2 ;# number of mobilenodes set opt(adhocrouting) DSDV ;# routing protocol set opt(cp) "" ;# connection pattern file set opt(sc) "" ;# node movement file. set opt(x) 100 ;# x coordinate of topology set opt(y) 100 ;# y coordinate of topology set opt(seed) 0.0 ;# seed for random number gen. set opt(stop) 10 ;# time to stop simulation #voip set opt(ftp1-start) 1.0 set opt(ftp3-start) 1.0 #background set opt(ftp2-start) 2.0 set opt(ftp4-start) 2.0 set num_wired_nodes 2 set num_bs_nodes 1 # ================================================================== # check for boundary parameters and random seed if { $opt(x) == 0 $opt(y) == 0 { puts "No X-Y boundary values given for wireless topology\n" if {$opt(seed) > 0 { puts "Seeding Random number generator with $opt(seed)\n" ns-random $opt(seed) # create simulator instance set ns_ [new Simulator] # set up for hierarchical routing $ns_ node-config -addresstype hierarchical AddrParams set domain_num_ 2 ;# number of domains 28

lappend cluster_num 2 1 ;# number of clusters in each domain AddrParams set cluster_num_ $cluster_num lappend eilastlevel 1 1 4 ;# number of nodes in each cluster AddrParams set nodes_num_ $eilastlevel ;# of each domain set tracefd [open test-udp.tr w] set namtrace [open test-udp.nam w] $ns_ trace-all $tracefd $ns_ namtrace-all-wireless $namtrace $opt(x) $opt(y) # Create topography object set topo [new Topography] # define topology $topo load_flatgrid $opt(x) $opt(y) # create God create-god [expr $opt(nn) + $num_bs_nodes] #create wired nodes set temp {0.0.0 0.1.0 ;# hierarchical addresses for wired domain for {set i 0 {$i < $num_wired_nodes {incr i { set W($i) [$ns_ node [lindex $temp $i]] # configure for base-station node $ns_ node-config -adhocrouting $opt(adhocrouting) \ -lltype $opt(ll) \ -mactype $opt(mac) \ -ifqtype $opt(ifq) \ -ifqlen $opt(ifqlen) \ -anttype $opt(ant) \ -proptype $opt(prop) \ -phytype $opt(netif) \ -channeltype $opt(chan) \ -topoinstance $topo \ -wiredrouting ON \ -agenttrace ON \ -routertrace OFF \ -mactrace OFF #create base-station node set temp {1.0.0 1.0.1 1.0.2 1.0.3 ;# hier address to be used for wireless ;# domain set BS(0) [$ns_ node [lindex $temp 0]] $BS(0) random-motion 0 ;# disable random motion #provide some co-ord (fixed) to base station node $BS(0) set X_ 50.0 29

$BS(0) set Y_ 50.0 $BS(0) set Z_ 0.0 # create mobilenodes in the same domain as BS(0) # note the position and movement of mobilenodes is as defined # in $opt(sc) #configure for mobilenodes $ns_ node-config -wiredrouting OFF for {set j 0 {$j < $opt(nn) {incr j { set node_($j) [ $ns_ node [lindex $temp \ [expr $j+1]] ] $node_($j) base-station [AddrParams addr2id \ [$BS(0) node-addr]] #coordinates for mobile nodes $node_(0) set X_ 25.0 #mobile node #node #2 $node_(0) set Y_ 25.0 $node_(0) set Z_ 0.0 $node_(0) label "VoIP" $node_(1) set X_ 75.0 #mobile node #node #2 $node_(1) set Y_ 25.0 $node_(1) set Z_ 0.0 $node_(1) label "Background" $W(0) label "VoIP" $W(1) label "Background" $BS(0) label "BaseStation" #class color $ns_ color 1 blue $ns_ color 2 red $ns_ color 3 black $ns_ color 4 yellow #create links between wired and BS nodes $ns_ duplex-link $W(0) $W(1) 64kb 5ms DropTail $ns_ duplex-link $W(1) $BS(0) 64kb 5ms DropTail $ns_ duplex-link-op $W(0) $W(1) orient down $ns_ duplex-link-op $W(1) $BS(0) orient left-down # setup TCP connections set tcp1 [new Agent/UDP] #$tcp1 set class_ 2 #$tcp1 set fid_ 1 30

$tcp1 set class_ 1 set sink1 [new Agent/LossMonitor] $ns_ attach-agent $node_(0) $tcp1 $ns_ attach-agent $W(0) $sink1 $ns_ connect $tcp1 $sink1 #set ftp1 [new Application/FTP] set ftp1 [new Application/Traffic/CBR] $ftp1 set packetsize_ 160 $ftp1 set interval_ 0.020 #$ftp1 set class_ 3 $ftp1 attach-agent $tcp1 $ns_ at $opt(ftp1-start) "$ftp1 start" #background traffic set tcp2 [new Agent/UDP] #$tcp2 set fid_ 2 $tcp2 set class_ 2 set sink2 [new Agent/LossMonitor] #$ns_ attach-agent $W(1) $tcp2 #$ns_ attach-agent $node_(1) $sink2 $ns_ attach-agent $node_(1) $tcp2 $ns_ attach-agent $W(1) $sink2 $ns_ connect $tcp2 $sink2 #set ftp2 [new Application/FTP] set ftp2 [new Application/Traffic/CBR] $ftp2 set packetsize_ 320 $ftp2 set interval_ 0.020 #$ftp2 set class_ 4 $ftp2 attach-agent $tcp2 $ns_ at $opt(ftp2-start) "$ftp2 start" ################################################## set tcp3 [new Agent/UDP] $tcp3 set class_ 3 set sink3 [new Agent/LossMonitor] $ns_ attach-agent $W(0) $tcp3 $ns_ attach-agent $node_(0) $sink3 $ns_ connect $tcp3 $sink3 set ftp3 [new Application/Traffic/CBR] $ftp3 set packetsize_ 160 $ftp3 set interval_ 0.020 $ftp3 attach-agent $tcp3 $ns_ at $opt(ftp3-start) "$ftp3 start" set tcp4 [new Agent/UDP] $tcp4 set class_ 4 set sink4 [new Agent/LossMonitor] $ns_ attach-agent $W(1) $tcp4 $ns_ attach-agent $node_(1) $sink4 31

$ns_ connect $tcp4 $sink4 set ftp4 [new Application/Traffic/CBR] $ftp4 set packetsize_ 320 $ftp4 set interval_ 0.020 $ftp4 attach-agent $tcp4 $ns_ at $opt(ftp4-start) "$ftp4 start" # source connection-pattern and node-movement scripts if { $opt(cp) == "" { puts "*** NOTE: no connection pattern specified." set opt(cp) "none" else { puts "Loading connection pattern..." source $opt(cp) if { $opt(sc) == "" { puts "*** NOTE: no scenario file specified." set opt(sc) "none" else { puts "Loading scenario file..." source $opt(sc) puts "Load complete..." # Define initial node position in nam for {set i 0 {$i < $opt(nn) {incr i { # 20 defines the node size in nam, must adjust it according to your # scenario # The function must be called after mobility model is defined $ns_ initial_node_pos $node_($i) 20 # Tell all nodes when the simulation ends for {set i {$i < $opt(nn) {incr i { $ns_ at $opt(stop).0 "$node_($i) reset"; $ns_ at $opt(stop).0 "$BS(0) reset"; $ns_ at $opt(stop).0002 "puts \"NS EXITING...\" ; $ns_ halt" $ns_ at $opt(stop).0001 "stop" proc stop { { global ns_ tracefd namtrace # $ns_ flush-trace close $tracefd close $namtrace 32

# informative headers for CMUTracefile puts $tracefd "M 0.0 nn $opt(nn) x $opt(x) y $opt(y) rp \ $opt(adhocrouting)" puts $tracefd "M 0.0 sc $opt(sc) cp $opt(cp) seed $opt(seed)" puts $tracefd "M 0.0 prop $opt(prop) ant $opt(ant)" puts "Starting Simulation..." $ns_ run VoWLAN SCTP: set opt(chan) Channel/WirelessChannel ;# channel type set opt(prop) Propagation/TwoRayGround ;# radio-propagation model set opt(netif) Phy/WirelessPhy ;# network interface type set opt(mac) Mac/802_11 ;# MAC type set opt(ifq) Queue/DropTail/PriQueue ;# interface queue type set opt(ll) LL ;# link layer type set opt(ant) Antenna/OmniAntenna ;# antenna model set opt(ifqlen) 50 ;# max packet in ifq set opt(nn) 2 ;# number of mobilenodes set opt(adhocrouting) DSDV ;# routing protocol set opt(cp) "" ;# connection pattern file set opt(sc) "" ;# node movement file. set opt(x) 100 ;# x coordinate of topology set opt(y) 100 ;# y coordinate of topology set opt(seed) 0.0 ;# seed for random number gen. set opt(stop) 10 ;# time to stop simulation #voip set opt(ftp1-start) 1.0 set opt(ftp3-start) 1.0 #background set opt(ftp2-start) 2.0 set opt(ftp4-start) 2.0 set num_wired_nodes 2 set num_bs_nodes 1 # =================================================================== # check for boundary parameters and random seed if { $opt(x) == 0 $opt(y) == 0 { puts "No X-Y boundary values given for wireless topology\n" if {$opt(seed) > 0 { puts "Seeding Random number generator with $opt(seed)\n" 33

ns-random $opt(seed) # create simulator instance set ns_ [new Simulator] # set up for hierarchical routing $ns_ node-config -addresstype hierarchical AddrParams set domain_num_ 2 ;# number of domains lappend cluster_num 2 1 ;# number of clusters in each domain AddrParams set cluster_num_ $cluster_num lappend eilastlevel 1 1 4 ;# number of nodes in each cluster AddrParams set nodes_num_ $eilastlevel ;# of each domain set tracefd [open test-sctp.tr w] set namtrace [open test-sctp.nam w] $ns_ trace-all $tracefd $ns_ namtrace-all-wireless $namtrace $opt(x) $opt(y) # Create topography object set topo [new Topography] # define topology $topo load_flatgrid $opt(x) $opt(y) # create God create-god [expr $opt(nn) + $num_bs_nodes] #create wired nodes set temp {0.0.0 0.1.0 ;# hierarchical addresses for wired domain for {set i 0 {$i < $num_wired_nodes {incr i { set W($i) [$ns_ node [lindex $temp $i]] # configure for base-station node $ns_ node-config -adhocrouting $opt(adhocrouting) \ -lltype $opt(ll) \ -mactype $opt(mac) \ -ifqtype $opt(ifq) \ -ifqlen $opt(ifqlen) \ -anttype $opt(ant) \ -proptype $opt(prop) \ -phytype $opt(netif) \ -channeltype $opt(chan) \ -topoinstance $topo \ -wiredrouting ON \ -agenttrace ON \ -routertrace OFF \ -mactrace OFF 34

#create base-station node set temp {1.0.0 1.0.1 1.0.2 1.0.3 ;# hier address to be used for wireless ;# domain set BS(0) [$ns_ node [lindex $temp 0]] $BS(0) random-motion 0 ;# disable random motion #provide some co-ord (fixed) to base station node $BS(0) set X_ 50.0 $BS(0) set Y_ 50.0 $BS(0) set Z_ 0.0 # create mobilenodes in the same domain as BS(0) # note the position and movement of mobilenodes is as defined # in $opt(sc) #configure for mobilenodes $ns_ node-config -wiredrouting OFF for {set j 0 {$j < $opt(nn) {incr j { set node_($j) [ $ns_ node [lindex $temp \ [expr $j+1]] ] $node_($j) base-station [AddrParams addr2id \ [$BS(0) node-addr]] #coordinates for mobile nodes $node_(0) set X_ 25.0 #mobile node #node #2 $node_(0) set Y_ 25.0 $node_(0) set Z_ 0.0 $node_(0) label "VoIP" $node_(1) set X_ 75.0 #mobile node #node #2 $node_(1) set Y_ 25.0 $node_(1) set Z_ 0.0 $node_(1) label "Background" $W(0) label "VoIP" $W(1) label "Background" $BS(0) label "BaseStation" #class color $ns_ color 1 blue $ns_ color 2 red $ns_ color 3 black $ns_ color 4 yellow #create links between wired and BS nodes $ns_ duplex-link $W(0) $W(1) 64kb 5ms DropTail 35

$ns_ duplex-link $W(1) $BS(0) 64kb 5ms DropTail $ns_ duplex-link-op $W(0) $W(1) orient down $ns_ duplex-link-op $W(1) $BS(0) orient left-down # setup TCP connections set tcp1 [new Agent/SCTP] $tcp1 set class_ 1 set sink1 [new Agent/LossMonitor] $ns_ attach-agent $node_(0) $tcp1 $ns_ attach-agent $W(0) $sink1 #$ns_ connect $tcp1 $sink1 set ftp1 [new Application/Traffic/CBR] $ftp1 set packetsize_ 160 $ftp1 set interval_ 0.020 $ftp1 attach-agent $tcp1 $ns_ at $opt(ftp1-start) "$ftp1 start" set tcp3 [new Agent/SCTP] $tcp3 set class_ 3 set sink3 [new Agent/LossMonitor] $ns_ attach-agent $W(0) $tcp3 $ns_ attach-agent $node_(0) $sink3 #$ns_ connect $tcp3 $sink3 set ftp3 [new Application/Traffic/CBR] $ftp3 set packetsize_ 160 $ftp3 set interval_ 0.020 $ftp3 attach-agent $tcp3 $ns_ at $opt(ftp3-start) "$ftp3 start" $ns_ connect $tcp1 $tcp3 #background traffic set tcp2 [new Agent/SCTP] $tcp2 set class_ 2 set sink2 [new Agent/LossMonitor] $ns_ attach-agent $node_(1) $tcp2 $ns_ attach-agent $W(1) $sink2 #$ns_ connect $tcp2 $sink2 set ftp2 [new Application/Traffic/CBR] $ftp2 set packetsize_ 320 $ftp2 set interval_ 0.020 $ftp2 attach-agent $tcp2 $ns_ at $opt(ftp2-start) "$ftp2 start" set tcp4 [new Agent/SCTP] $tcp4 set class_ 4 set sink4 [new Agent/LossMonitor] $ns_ attach-agent $W(1) $tcp4 36

$ns_ attach-agent $node_(1) $sink4 #$ns_ connect $tcp4 $sink4 set ftp4 [new Application/Traffic/CBR] $ftp4 set packetsize_ 320 $ftp4 set interval_ 0.020 $ftp4 attach-agent $tcp4 $ns_ at $opt(ftp4-start) "$ftp4 start" $ns_ connect $tcp2 $tcp4 # source connection-pattern and node-movement scripts if { $opt(cp) == "" { puts "*** NOTE: no connection pattern specified." set opt(cp) "none" else { puts "Loading connection pattern..." source $opt(cp) if { $opt(sc) == "" { puts "*** NOTE: no scenario file specified." set opt(sc) "none" else { puts "Loading scenario file..." source $opt(sc) puts "Load complete..." # Define initial node position in nam for {set i 0 {$i < $opt(nn) {incr i { # 20 defines the node size in nam, must adjust it according to your # scenario # The function must be called after mobility model is defined $ns_ initial_node_pos $node_($i) 20 # Tell all nodes when the simulation ends for {set i {$i < $opt(nn) {incr i { $ns_ at $opt(stop).0 "$node_($i) reset"; $ns_ at $opt(stop).0 "$BS(0) reset"; $ns_ at $opt(stop).0002 "puts \"NS EXITING...\" ; $ns_ halt" $ns_ at $opt(stop).0001 "stop" proc stop { { global ns_ tracefd namtrace # $ns_ flush-trace 37

close $tracefd close $namtrace # informative headers for CMUTracefile puts $tracefd "M 0.0 nn $opt(nn) x $opt(x) y $opt(y) rp \ $opt(adhocrouting)" puts $tracefd "M 0.0 sc $opt(sc) cp $opt(cp) seed $opt(seed)" puts $tracefd "M 0.0 prop $opt(prop) ant $opt(ant)" puts "Starting Simulation..." $ns_ run AWK file code: Throughput: #throughput BEGIN { node =1; time1 = 0.0; time2 = 0.0; { num_packet=0; count = 0; bytes_counter=0; time_counter=0; time2 = $2; #if ($3=="_3_"){ #bytes_counter+=$8*8/1000; # if ( ($3=="1" && $4=="0" && $1=="r") ($3=="1" && $4=="0" && $1=="r")){ bytes_counter+=$6*8/1000; time_counter = time2-count; if ( time_counter>=1) { #bytes_counter = bytes_counter/time_counter; printf("%f %f\n", time2, bytes_counter) > "throughput"; total_bytes += bytes_counter; time_counter=0; bytes_counter=0; count++; 38

END { average = total_bytes/count; print "average throughput:" average; print("done"); Packet Loss: BEGIN { { node =1; time1 = 0.0; time2 = 0.0; num_packet=0; packet_id = 0; bytes_counter=0; time2 = $2; if ($1=="s" && $3=="_3_") { packet_id = $6; if ( ($1=="d" $1 == "D") && $12==packet_id ) { bytes_counter += $6; num_packet++; printf("%f %f\n", time2, num_packet) > "packetlost"; time1 = $2; END { print("done"); 39

Latency: #latency BEGIN { send_time = 0.0; receive_time = 0.0; sim_time = 0.0; send_flag = 0; lat_time = 0; ave_lat = 0; count =0; { sim_time = $2; if ($1=="s" && $3=="_3_" && send_flag==0) { send_time = $2; send_flag = 1; if ($1=="r" && $3=="1" && $4=="0" && send_flag==1) { receive_time = $2 send_flag = 0; lat_time = receive_time - send_time; printf("%f %f\n", sim_time, lat_time) > "latency"; #leslie ave_lat += lat_time; count++; END { average = ave_lat/count; print "average latency:" average; print("done"); Jitter: BEGIN { send_time = 0.0; receive_time = 0.0; sim_time = 0.0; 40

{ send_flag = 0; lat_time_curr = 0; lat_time_prev = 0; jitter = 0; count =0; sim_time = $2; if ($1=="s" && $3=="_3_"&& send_flag==0) { send_time = $2; send_flag = 1; if ($1=="r" && $3=="1" && $4=="0" && send_flag==1) { receive_time = $2 send_flag = 0; lat_time_curr = receive_time - send_time; jitter = lat_time_curr - lat_time_prev; printf("%f %f\n", sim_time, jitter) > "jitter"; ave_jitter += jitter; lat_time_prev = lat_time_curr; count++; END { average = ave_jitter/count; print"average jitter:" average; print("done"); Gnuplot Code: Throughput: set title 'VoWLAN over SCTP - throughput' set grid set ylabel 'kbps' set xlabel 'time' plot 'throughput' w linespoints title 'voip throughput' Packet Loss: set title 'VoWLAN over SCTP - packetloss' set grid set ylabel 'byte' set xlabel 'time' plot 'packetlost' w linespoints title 'voip packetlost' Latency: set title 'VoWLAN over SCTP - Latency' set grid 41

set ylabel 's' set xlabel 'time' plot 'latency' w linespoints title 'voip latency' 42