ELEC 6851: Introduction to Telecommunication

Networks QualNet Project Phase 2

Date of submission –December 19th, 2022

SUBMITTED BY

Arun Kumar Ganesan Karthini Rajasekharan

(Student ID: 40235409) (Student ID: 40238556)

SUBMITTED TO

Prof. Mustafa Mehmet Ali

 Overview:

The project phase II report consists of the simulation and design of a wireless
telecommunication network as per the given particulars. The design in this report consists
of 4 wireless sub-networks such as A, B, C and D and each individual network consists of
5 nodes and 3 routers to transfer the packets across the networks. The report also consists
of the results of the simulation in the form of bar charts representing the performance of
the wireless network.

The whole project is again sub-divided into two domains: Scenario A and Scenario B that
is depicted along with Fading (Rayleigh) and without Fading. From the following
simulation results, the necessary differences are picturized which can be further used to
assimilate the effect of the domains on the performance of a wireless network. Scenario
A consists of four wireless sub-networks A, B, C and D with two connections: one UDP
and other as TCP traffic sources generating at least 500 packets.

Channel Frequency [0] = 2.4 GHz

Channel Frequency [1] = 2.5 GHz
Channel Frequency [2] = 2.6 GHz
Channel Frequency [3] = 2.7 GHz

Components Subnet A Subnet B Subnet C Subnet D

190.0.1.2 190.0.2.0 190.0.3.0 190.0.4.0
Nodes 1-190.0.1.1 5-190.0.2.1 13-190.0.3.1 9-190.0.4.1
2-190.0.1.2 6-190.0.2.2 14-190.0.3.2 10-190.0.4.2
3-190.0.1.3 7-190.0.2.3 15-190.0.3.3 11-190.0.4.3
4-190.0.1.4 8-190.0.2.4 16-190.0.3.4 12-190.0.4.4
20-190.0.1.6 19-190.0.2.6 21-190.0.3.6 22-190.0.4.6
Routing Nodes 17-190.0.1.5 17-190.0.2.5 23-190.0.4.7 18-190.0.3.5
23-190.0.2.7 18-190.0.4.5

IP Address for Nodes, Subnet and Routing Nodes

Scenario A:
In this scenario two connections are considered UDP (between node 20 and 8) and TCP
(between node 21 and 22) and each connection generates at least 500 packets.
Design of the Network Scenario A:

Fig 1: Network design of Scenario A

Connection Parameters:

Connection Packet Mean Transmission Packet Simulation time

length Start Time Duration Interval
UDP 2500 1 650 1.0 700

Connection End time Item to send Item size

TCP 620 610 512

Traffic generators parameters are set for no fading case and the bar graphs have been
generated with the required frequency for corresponding channel
Application Layer:

UDP

Fig 2: Traffic Client: Data units sent (Messages)

Fig 3: UDP-Traffic Server: Data units received (Messages)

TCP:
 Fig 5: Unicast Offered Load (TCP)

Fig 6: Unicast Received Throughput (TCP)

Results

 UDP

Traffic Client Traffic Server

Source (Node 20) Destination (Node 8)
Total Data Units Sent 640 Total Data Units Received 640

Data lost in UDP connection: 640-640 = 0.

Tab 1: Data Table of UDP connection

 TCP
 Traffic Client Traffic Server
Source (Node 21) Destination (Node 22)
Total Data Units Sent 610 Total Data Units Received 610
Throughput Sent 938145 Throughput Received 893218

Tab 2: Data Table of TCP connection

Transport Layer:
UDP - Packets from Application and Packets to Application.

Fig 7: Unicast Data Segments sent from the transport layer (UDP)

Fig 8: Unicast Data Segments Received from the transport layer

(UDP)
 TCP - Data Packets Sent, Received and Retransmitted

Fig 9: Unicast Data Segments sent from the transport layer (UDP)

Fig 10: Unicast Data Segment received at the transport layer (UDP)

Fig 11: Data Packets Retransmitted (UDP)

MAC Layer:
CTS packets sent

Fig 12: CTS packets sent

RTS packets sent

Fig 13: RTS packets sent

ACK packets sent

Fig 14: ACK packets sent

Packets drop due to transmission limit

Fig 15: Packets drop due to transmission limit

Unicast sent

Fig 16: Unicast sent

Unicast received

Fig 17: Unicast Received

 Results and Discussion
 UDP

 Node 20 is in subnetwork A and node 8 is in subnetwork B and they are

connected by UDP connection and a router (17) between them.
 For UDP, total unicast packet transmitted is 1147 and as we can see from
the below table same number of packets is received at the receiver end also
same number of packets are transmitted from router 18 so it is proved that
packets are transmitted via router not directly.
 TCP
o Node 21 is in subnetwork C and node 22 is in subnetwork D and they are
connected by UDP connection and a router (18) between them.

o Node 21 is going to send 600 packets which is received by node 22 via router
which we can see in router entry which is also 600 then 600 packets as
acknowledgement are sent to node 21 by node 22.

MAC LAYER
TCP Connection UDP Connection
Node Node Node Node
21 22 20 8
Unicast Sent 600 301 Unicast Sent 1147 0
Unicast Received 301 600 Unicast Received 0 1147
CTS Packet Sent 301 600 CTS Packet Sent 0 1147
RTS Packet Sent 612 315 RTS Packet Sent 1147 0
ACK Packet Sent 301 600 ACK Packet Sent 0 1147
RTS
RTS Retransmission due to
Retransmission due to 12 14 timeout
0 0
timeout
Packets sent due to 0 0 0 0
Packets sent due to
retransmission
retransmission

Tab 3: Data Table of UDP and TCP connection

Packets retransmission due to Timeout=0

Total Unicast sent= RTS Packets Sent-RTS Retransmission due to Timeout-Packets

Retransmissions due to Timeout.

For node 21: -612-12-0=600(unicast Sent)

For node: 22: -315-14-0=301(Unicast Sent)
For node: 20: -1147-0-0=1147(Unicast Sent)
Router Interface:

Router
Interface 0
Name of RTS CTS ACK Unicast Unicast Packet dropped due to
routers(node) Sent Received retransmission limit
17(Between 0 1147 1147 0 1147 0
Channel A and B)
18 (Between 313 600 600 301 600 0
Channel B and C)

Router
Interface 1
Name of RTS CTS ACK Unicast Unicast Packet dropped due to
routers(node) Sent Received retransmission limit
17(Between 1148 0 0 1147 0 0
Channel A and B)
18 (Between 615 301 301 600 301 0
Channel B and C)

Conclusion:
From the scenario A UDP connection, the number of packets transmitted and received is
same and there is no packet loss in network, which is the same case with TCP connection
as well. In conclusion, UDP is faster network (average end-end to delay 0.00673038) than
TCP and from the above table, it’s clear that in both UDP and TCP protocols, there is no
packet drop in the network and the packets have taken path towards router.

Scenario B:
Here, 4 wireless sub networks are connected A, B, C and D are connected to each other
through three routers. Each sub network contains 5 nodes and IEEE 802.11 as MAC
protocol and IPv4 as network protocol. Here, the connections are established between two
random node pairs in the network. The two connections are configured to UDP service
(TRAF GEN) and Rayleigh Fading has been taken into consideration for calculating the
average packet delay and throughput with fading. The parameters settings for our
calculations are as shown below:
 Mean start time: Exponential and 1 Sec
 Duration: Deterministic and 650 Sec
 Packet size distribution: Exponential and 2500 bytes
- Mean packet interval: Exponential - 0.25, 0.35, 0.45, 0.55, 0.65, 0.75, 0.85 and 0.95
 Simulation time: 700 Sec
Here, we have considered UDP connection between source node 7 and destination node
10 in the network that goes through the bottleneck router.
Design of the Network Scenario B:

Fig 18: Network design of Scenario B

Average UDP packet delay with and without fading

Average UDP Packet Delay (Sec)
Node #7
Interval
With Fading Without Fading
0.25 7.39 e-07 7.6 e-07
0.35 7.59 e-07 7.6 e-07
0.45 7.39 e-07 7.4 e-07
0.55 7.39 e-07 7.4 e-07
0.65 7.27 e-07 7.33 e-07
0.75 7.22 e-07 7.23 e-07
0.85 7.19 e-07 7.2 e-07
0.95 7.15 e-07 7.16 e-07

Tab 4: Average UDP Packet Delay

The table above shows the average UDP packet delays with fading and without fading for the
different time intervals. The table is followed by the line graph shown in figure 19 depicting the same
values. It can be observed that fading (Rayleigh fading in this scenario) has a greaterimpact on
the packet delay as the delay increases.
Graph:

Fig 19: Average UDP Packet delay with fading

Fig 20: Average UDP Packet delay without fading

UDP Client throughput with and without fading

UDP Client Throughput (bits/Sec)
Node #7
Interval
With Fading Without Fading
0.25 43186.4 75498.4
0.45 43186.4 43186.4
 0.55 43186.4 43186.4
0.65 30747.6 35226.4
0.75 27577.3 27577.3
0.85 24851.5 24851.5
0.95 22616.8 22616.8

Tab 5: UDP Client Throughput

The UDP client throughput is exactly same in both fading and without fading case.
Table above summarizes the values followed by the line graph shown below.

Graph:

UDP Client throughput with and without fading

140000
UDP Client throughput

120000
100000
80000
60000
40000
20000
0
1 2 3 4 5 6 7
Packet Interval

with without

Fig 20: UDP Client throughput with and without fading

From the above line graph and table, it is evident that client throughput is the same.

UDP Server throughput with and without fading

The table below shows that UDP server throughput has marginal difference when
simulated with fading and without fading. The effect of Rayleigh fading can be clearly
observed on the server performance.

UDP Server Throughput(bits/Sec)

Node #7
Interval
With Fading Without Fading
0.25 43185.5 75495.6
0.45 43185.5 43185.6
0.55 43185.5 43185.6
 0.65 30748.3 35227.1
0.75 27575.2 27575.9
0.85 24851.5 24851.5
0.95 22617.2 22617.2

Tab 6: UDP Server Throughput

UDP Server throughput with and without fading

140000
UDP Server Throughput

120000
100000
80000
60000
40000
20000
0
1 2 3 4 5 6 7
Packet Interval

with without

Fig 21: UDP Server throughput with and without fading

The table and curve show that UDP server throughput has marginal difference when
simulated with fading and without fading. The effect of Rayleigh can be clearly
observed on the server performance. It is evident that server throughput is slightly
more in without-fading case when compared to with-fading case. When time interval
between the packets decreases, the throughput increases (as shown in line graph).

Conclusion:

Scenario A, the total number of data units sent by the client is equal to the number of units
received by the server. The results have been discussed above for the line charts
respectively. CTS, RTS, and ACK packets are also taken in consideration. In scenario B,
the plots for server throughput proves that the difference between the throughput of with-
fading and without-fading domains is negligible for smaller distance between the nodes
and this difference increases as the distance increases.