NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November

OPNET 23, 2010

INTRODUCTION:

LAN (local area network):

A network can be defined as a group of devices connected together such that they
can communicate each other and can share their resources. Based on the area coverage and
the number of interconnected devices the networks are classified basically into three types
like Local Area Networks, Wide Area Networks and Metropolitan Area Networks.

LAN is a computer network concentrated in a geographical area, such as in a building

or on a university campus. [ CITATION Jam10 \l 2057 ]

In the link layer there are two fundamentally different types of link-layer channels.
They are Broad cast channels and link-layer channel. Broad cast channels are common in
Local Area Networks (LAN), wireless LANs, satellite networks and hybrid fiber-coaxial cable
(HFC) access networks.[ CITATION Jam10 \l 2057 ]

Whenever a user wants to connect to the internet from the university or office, the
connection is made from the host to LAN to router to internet. The transmission rate, R, of
the most LANs is very high. 10 mbps LANs were common in the early 1980s, and nowadays
100 Mbps, 1 Gbps and even 10 Gbps are available.[ CITATION Jam10 \l 2057 ]

Two classes of LAN technologies were popular in the work space in the 1980s and
early 1990s. These are Ethernet LANs (also known as 802.3 LANs) and token passing
technologies, including Token ring (also known as 802.5) and Fiber Distributed Data
Interface (FDDI). Ethernet LANs are random access based. Due to wide development of
relentless Ethernet Token passing technologies came to an extinct.[ CITATION Jam10 \l 2057 ]

Token-Ring is defined “as a method for interconnecting stand-alone work stations in

a peer-to-peer mode that allows each work station to act as its own transaction
device.”[ CITATION Art00 \l 2057 ]

Addressing:

LAN addressing is also called as physical address or MAC address. The MAC address
size for most LANs is 6 bytes long, by which 248 MAC addresses are possible. Actually it is
not the node (host or router) that has a link-layer address but the node’s adapter that has
the link-layer address. It is possible to change the MAC address of the adapter via software,
that is, the address is variable and can also be used as fixed address. IEEE manages the MAC
address space so that no two adapters have the same address. [ CITATION Jam10 \l 2057 ]

Address Resolution Protocol (ARP) is used to translate network-layer addresses

(Internet IP addresses) into link-layer addresses (MAC addresses) and vice versa.

1037089
Page 1
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

LAN
(wired or
wireless)
[ CITATION Jam10 \l 2057 ]

Figure 1: Local Area Network

ETHERNET (802.3):
Ethernet is introduced by the Xerox Corporation in the early 1970s as a different
networking concept to token ring. “Ethernet is designed to run a single transmission line to
which all stations are connected.”[ CITATION Art00 \l 2057 ] .

Ethernet is also peer-to-peer connection but the monitoring device is not required.
The stations in the Ethernet do not require waiting for a token to be passed before
transmitting, unlike Token Ring. Stations can be connected directly to each other and can
send the data at any time required.[ CITATION Art00 \l 2057 ]

Today Ethernet is considered as the most powerful wired LAN technology. It is

expected to remain for several years in future. Ethernet is the first high speed LAN which is
widely deployed. Its hardware being cheap and less complex, Ethernet has become a great
success compared to other LANs.

Bob Metcalfe and David Boggs invented the original Ethernet LAN in the mid- 1970s.
Interconnection between the nodes in the original Ethernet LAN is made using the coaxial
bus, which is shown in the figure2 given below. “Ethernet with a bus topology is a broadcast
LAN -all transmitted frames travel to and are processed by all adapters connected to the
bus.”[ CITATION Jam10 \l 2057 ]

[ CIT
ATION Jam10 \l 2057 ]

Figure2: The original Metcalfe Ethernet Sketch

1037089
Page 2
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

In the late 1990s many universities and companies replaced bus topology with
Ethernet installations with a hub based star topology which is also a broadcast LAN, where
the hosts are directly connected to the hub via a twisted pair copper wire. “A hub is a
physical layer device that acts on individual bits rather than frames.” [ CITATION Jam10 \l
2057 ]

Later in the early 2000s the Ethernet has seen a greater evolutionary change, the
hub at the center of the star topology is replaced with a switch which is a collision-less as
well as bona-fide store-and-forward packet switch.[ CITATION Jam10 \l 2057 ]

star
[ CITATION Jam10 \l 2057 ]

Figure3: Star topology with Switch at center

Ethernet Frame Structure:

[ CI
TATION Jam10 \l 2057 ]

Figure4: Ethernet Frame Structure

Consider an example where one host with an adapter A having MAC address AA-AA-
AA-AA-AA-AA sends an IP datagram to the host with an adapter B having MAC address BB-
BB-BB-BB-BB-BB. “The sending adapter encapsulates the IP datagram within an Ethernet
frame and passes the frame to the physical layer. The receiving adapter receives the frame
from the physical layer, extracts the IP datagram, and passes the IP datagram to the
network layer.”[ CITATION Jam10 \l 2057 ]

Data field (46 to 1,500 bytes): This field carries the IP datagram. 46 is the minimum and
1,500 is the maximum transmission unit of Ethernet. If the size exceeds 1,500 bytes the
fragmentation of the datagram is to be performed by the sending host and if the size falls
below 46 bytes the data field is to be filled out to 46 bytes by stuffing. [ CITATION Jam10 \l
2057 ]

Destination address (6 bytes): This field contains the MAC address of the destination
adapter, BB-BB-BB-BB-BB-BB.

Source address (6 bytes): This field contains the MAC address of the transmitting adapter,
AA-AA-AA-AA-AA-AA.

Type field (2 bytes): This field permits Ethernet to multiplex network-layer protocols.

1037089
Page 3
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Cyclic Redundancy Check (4bytes): This field allows the receiving adapter to detect the bit
errors in the frame.

Preamble (8 bytes): Each of the first 7 bytes of the preamble has a value of 10101010 and
the last byte has a value of 10101011. The first 7bytes are used to wake up the receiving
adapters and to synchronise their clocks to that of sender’s clock. The last 2 bits of the 8 th
byte of the preamble that is two consecutive 1’s alerts the receiving adapter that is adapter
B, about the arrival of the ‘important stuff’.[ CITATION Jam10 \l 2057 ]

All Ethernet technologies are unreliable because these provide connection-less

services to the network layer. There is no handshaking or acknowledgement reports before
or after sending the frames between the nodes adapters. Due to this the sending node’s
adapter will have no idea about the frame transmission. This type of unreliable transport
makes Ethernet simple and cheap.

CSMA/CD:

Carrier Sense Multiple Access with Collision Detect is the Ethernet’s Multiple Access
Protocol. Since Ethernet LAN employs broadcast, it needs a multiple access protocol.
CSMA/CD follows the following steps[ CITATION Jam10 \l 2057 ]

1) For an adapter there is no notation of time slots; that is an adapter may begin to
transmit at any time
2) Carrier Sensing means when some other adapter is found transmitting a frame, it can
sense the state of the link whether busy or idle.
3) Collision Detect means, as soon as the sending adapter detects that the other
adapter is using the link, it aborts its transmission and avoids collision.
4) Before retransmitting the adapter waits for a random time which is small compared
to the time taken to transmit the frame.

Since an adapter sends a frame with a probability 1 whenever a busy line goes idle,
the Ethernet is said to be 1-persistent protocol. [ CITATION RNi09 \l 2057 ]

Transmission Control Protocol (TCP):

TCP is a connection oriented service at the transport layer and is specially designed
to provide a reliable end-to-end byte stream over an unreliable network. “TCP was designed
to dynamically adapt to properties of the internetwork and to be robust in the face of many
kinds of failures.” [ CITATION And03 \l 2057 ]

TCP was initially defined in RFC 793. Due to different errors and inconsistencies
were detected there is a requirement for a change in some areas. These clarifications and

1037089
Page 4
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

some bug fixes are detailed in RFC 1122. Extensions are given in RFC 1323. [ CITATION
And03 \l 2057 ].

Implementation of a TCP is done by performing a three-way handshake between the

source and destination prior to sending the data and performing a four-way handshake
before breaking down the connection.[ CITATION RNi09 \l 2057 ]

A TCP connection provides a full-duplex service. A TCP connection is always point-

to-point; connection is between a single sender and a single receiver. Multicasting in a
single send operation is not possible using TCP connection. [ CITATION Jam10 \l 2057 ]

TCP Segment Structure:

[ CITATION Jam10 \l
2057 ] Figure5: TCP Segment Structure

TCP header is 20 bytes long. The header includes Source and Destination port
numbers which are used for multiplexing/demultiplexing data from/to upper-layer
applications. The header also includes a Checksum field.[ CITATION Jam10 \l 2057 ]

The Sequence number and Acknowledgement number fields perform their usual
function that is the Sequence number is just a counter that is incremented on each RTP (Real time
transport control protocol) packet sent. It is used to detect lost packets. The latter specifies the
next byte expected, not the last byte correctly received. Both are 32 bits long because every

1037089
Page 5
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

byte of data is numbered in a TCP stream. These are used for reliable data transfer service.
[ CITATION And03 \l 2057 ]

The 16-bit receive window field is used for flow control. Flow control in TCP is
handled using a variable-sized sliding window. “The Window size field tells how many bytes
may be sent starting at the byte acknowledged. A Window size field of 0 is legal and says
that the bytes up to and including Acknowledgement number - 1 have been received, but
that the receiver is currently badly in need of a rest and would like no more data for the
moment, thank you. The receiver can later grant permission to send by transmitting a
segment with the same Acknowledgement number and a nonzero Window size field”.
[ CITATION And03 \l 2057 ]

“The 4-bit header length field specifies the length of the TCP header in 32-bit words.
Due to the TCP options field the TCP header can be of variable length. Technically, this field
really indicates the start of the data within the segment, measured in 32-bit words, but that number
is just the header length in words, so the effect is the same.”[ CITATION And03 \l 2057 ]

“The Options field provides a way to add extra facilities not covered by the regular
header. The most important option is the one that allows each host to specify the maximum
TCP payload it is willing to accept. Using large segments is more efficient than using small
ones because the 20-byte header can then be amortized over more data, but small hosts
may not be able to handle big segments. During connection setup, each side can announce
its maximum and see its partner's. If a host does not use this option, it defaults to a 536-
byte payload. All Internet hosts are required to accept TCP segments of 536 + 20 = 556
bytes. The maximum segment size in the two directions need not be the same.” [ CITATION
And03 \l 2057 ]

“The flag field contains 6 bits. The ACK bit is set to 1 to indicate that the
Acknowledgement number is valid. If ACK is 0, the segment does not contain an
acknowledgement so the Acknowledgement number field is ignored. URG is set to 1 if the
Urgent pointer is in use. The Urgent pointer is used to indicate a byte offset from the
current sequence number at which urgent data are to be found. The RST bit is used to reset
a connection that has become confused due to a host crash or some other reason. It is also
used to reject an invalid segment or refuse an attempt to open a connection. The SYN bit is
used to establish connections. The connection request has SYN = 1 and ACK = 0 to indicate
that the piggyback acknowledgement field is not in use. The connection reply does bear an
acknowledgement, so it has SYN = 1 and ACK = 1. In essence the SYN bit is used to denote
CONNECTION REQUEST and CONNECTION ACCEPTED, with the ACK bit used to distinguish
between those two possibilities. The FIN bit is used to release a connection. It specifies that
the sender has no more data to transmit. However, after closing a connection, the closing
process may continue to receive data indefinitely. Both SYN and FIN segments have
sequence numbers and are thus guaranteed to be processed in the correct order.” [ CITATION
And03 \l 2057 ]

1037089
Page 6
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

User Datagram Protocol (UDP):

UDP is a connection-less oriented service at the transport layer. UDP is described in
RFC 768. “UDP transmits segments consisting of an 8-byte header followed by the
payload.”[ CITATION And03 \l 2057 ] It is very small compared to the TCP header where the
TCP header is of 20 bytes long. Each UDP segment is handed independently of each other.
There is no handshaking between the sender and the receiver before transmission of the
frames in UDP unlike TCP which is a connection oriented service. UDP adds nothing to IP
apart from multiplexing/demultiplexing and error checking. [ CITATION Jam10 \l 2057 ]

“UDP takes messages from the application process, attaches source and a
destination port number fields for the multiplexing/demultiplexing services, add two other
small fields, and passes the resulting segment to the network layer. Then the network layer
encapsulates this segment from the transport layer into an IP datagram and then delivers
the segment to the receiving host with a best-effort.”[ CITATION Jam10 \l 2057 ]

DNS (domain name system) is an example of an application-layer protocol that uses

UDP. HTTP uses TCP rather than UDP because reliability is more important for web pages
with text. UDP is faster than TCP because there is no hand-shaking and congestion-control
mechanism before delivering the segment. And thus it does not introduce any delay to
establish a connection. Since it is connection-less there is no guarantee of the delivery of the
data segment.[ CITATION Jam10 \l 2057 ]

UDP Segment Structure:

The data filed of the UDP segment is occupied by the Application data. For example,
the data field generally consists of either a query message or a response message in DNS.
The UDP header has only four fields each having only 2 bytes of length. [ CITATION Jam10 \l
2057 ]

The source port is mainly needed when a reply is needed to be sent back to the
source. The source port field is copied from the incoming segment into the destination port
field of the outgoing segment. This process of sending the reply specifies which process on
the sending machine is to get it.[ CITATION And03 \l 2057 ]

The length field indicates the length of the UDP segment. The UDP length field
includes the 8-byte header and the data. [ CITATION And03 \l 2057 ][ CITATION Jam10 \l 2057 ]

The checksum field indicates the receiving host to check whether errors have been
introduced in the segment. The UDP checksum is optional and stored as 0 if not computed. [
CITATION And03 \l 2057 ][ CITATION Jam10 \l 2057 ]

1037089
Page 7
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

32 bits

Length, in bytes of
UDP segment,
including header [ CITATION Jam10 \l
2057 ]

Figure6: UDP Segment Structure

Task 1: Analysis of network performance with 1024 packet size:

OPNET (Optimized Network Engineering Tools) OPNET is a discrete network
simulator which contains a comprehensive development environment supporting the
modelling and performance evaluation of communication networks and distributed
systems.[ CITATION Mic02 \l 2057 ]

Performance evaluation and trade-off analysis require large volumes of simulation

results to be interpreted and OPNET includes a tool for graphical representation and
processing of simulation output. Simulation runs can be configured to automatically
generate animations of the modelled system at various levels.[ CITATION Mic02 \l 2057 ]

A model and analysis of an Ethernet LAN is performed using OPNET software in this
assignment report with 8 nodes interconnected in a bus topology via coaxial link. The data
rate of coaxial link is taken as 10 Mbps.

1037089
Page 8
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure7: Network Layout

In the above figure Ethernet LAN model is designed by using OPNET Modeler 14.5.
The following steps are followed to design an Ethernet LAN [ CITATION RNi09 \l 2057 ]

1) Initially after opening the OPNET the new project is selected from the file menu and
named as 1037089_Ethcoax and the scenario is named as Question_1. Selecting
create the empty scenario from the initial topology dialog box and choosing office
from the network scale list assigned 600 to X span and 400 to Y span. Close the
object Palette dialog box.
2) Network configuration is selected as a bus topology by choosing the rapid
configuration from the topology menu item.
3) In the rapid configuration dialog box select models button is selected and from the
model list ethcoax is chosen. The Node model, Link model and Tap model are
updated by choosing ethcoax_station and eth_coax and eth_tap respectively. The
number of nodes is selected as 8. Placement of the model is done by changing the
parameters of the Head of bus and the bus size and tap size.
4) Selecting edit attributes (advanced) from the menu the bus attributes are changed.
The model is changed to eth_coax_adv and the colour and size of the bus are
changed to the convenient colour and size. The thickness is chosen as 3.
5) Then the project is saved by choosing the save button from the menu.

To configure the traffic generated by the nodes we follow the following the steps [ CITATION
RNi09 \l 2057 ]

1) Right click on any of the nodes and click select similar nodes. Again right click on any
of the nodes and select edit attributes.
2) By expanding the traffic generation parameters hierarchy change the value of start
time to constant(1), on state time to exponential(100) and off state time to
constant(0).

1037089
Page 9
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

3) Expanding the packet generation parameters hierarchy change the value of packet
size to constant(1024), right click on the inter arrival time and select the promote
attribute to higher level. This is applied to all nodes by selecting apply changes to
selected objects.

Choosing the statistics to be collected during the simulation [ CITATION RNi09 \l 2057 ]

1) Expect on the nodes or links right click on the work shape and select choose
individual DES statistics.
2) Expand the global statistics hierarchy and then expand ethcoax hierarchy and check
delay(sec) box.
3) Expanding the traffic sink hierarchy check the box traffic received(bits/sec). On the
bottom of the right hand side tick the box generate scalar data and select time
average from the drop down list.
4) Expanding the traffic source hierarchy check the box traffic sent(bits/sec). On the
bottom of the right hand side tick the box generate scalar data and select time
average from the drop down list.
5) Expand the node statistics hierarchy and again expand the ethcoax hierarchy and
select collision count, deference time and transmission attempts.
6) Expand the link statistics hierarchy and again expand bus hierarchy and select
throughput(bits/sec) and utilization.

To examine the performance under different loads, we need to run the simulation many
times by changing the load into the network. The following steps are followed [ CITATION
RNi09 \l 2057 ]

1) Go to configure/run simulation DES button and click on it.

2) On the common tab, change the duration to 30 seconds, seed to 0 and the update
interval to 100000.
3) On the left hand side expand the inputs hierarchy and choose the object attributes
tab and the add by using add button.

The interarrival time is added for all nodes by following steps[ CITATION RNi09 \l 2057 ]

1) Choose the first attribute in the list and click on it, then click wildcard button. Now
click the node_0 and select the asterisk from the list and click ok
2) Having the asterisk a new attribute is generated and it is added by clicking under the
add? button corresponding to the cell.
3) After removing the add button from the node_0 and leaving the asterisk selected
click ok.

1037089
Page 10
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

4) Interarrival times are assigned exponential by clicking the enter the multiple values
in the Office Network.*.Traffic Generation Parameter dialog box.
5) By double clicking on the cells under the value button different exponentials values
are assigned for example exponential(2).
6) Then click ok and then click on run button.

Analyzing the results:

Throughput Vs Offered Load

By clicking on the configure/run simulation button and selecting DES Parametric

studies tab expand the scalar Statistics hierarchy. Selecting the Traffic Received
(bits/sec).average click the Set As Y-Series button and selecting the Traffic Sent
(bits/sec).average click the Set As X-Series button. The following figure results which gives
the relationship between the offered load and throughput.

Throughput is defined as the average number of bits successfully received or

transmitted by the receiver or transmitter channel per unit time, in bits per second. (Source:
OPNET Network Simulation Software, Version: 14.5)

Offered load is the actual data which is fed to the network through the bus using the
8 nodes that is the transmitted bits. (Source: OPNET Network Simulation Software, Version:
14.5)

Delay represents the end to end delay of frames accepted by all Ethernet layers in
the network. This statistic is measured from the time a frame is submitted to the Ethernet
layer for transmission to the time the frame is received by the Ethernet layer of the
destination node. (Source: OPNET Network Simulation Software, Version: 14.5)

Figure 7: Throughput vs Offered Load

1037089
Page 11
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

The above figure shows that the Traffic Source Traffic Sent (bits/sec). Average and
Traffic Sink Traffic Received (bits/sec).Average are linear only up to 8 Mbps Traffic sent
which indicates that when the Traffic Sent exceeds the 8 Mbps the nonlinearity occurs due
to delay between the Transmitted Bits and the Received bits. The throughput also limited to
9 Mbps even though the Coaxial link data rate is 10 Mbps, which is not achieved in practical.
Theoretically the linearity should extend up to 10 Mbps but in practical this does not happen
because of some limitations.

Network Utilization:

To analyse the network utilization select DES Graphs tab and expand the Object
Statistics and then Office Network hierarchies and then bus_0[0] and then bus hierarchies
and then tick the Utilization box. Select Overlaid Statistics and average from the
Presentation tab. The figure 8 is obtained.

Utilization is defined as the percentage of the consumption to date of an available

channel bandwidth, where a value of 100.0 would indicate full usage. (Source: OPNET
Network Simulation Software, Version: 14.5)

Simulation studies are generally done on a (usually computer-based) model of the

system created for the purpose of studying certain system dynamics and characteristics but
not on the real-world system. The purpose of any model is to enable its users to draw
conclusions about the real system by studying and analyzing the model. [ CITATION Moh08 \l
2057 ]

Figure 8: Network Utilization (Utilization Vs Simulation Time)

1037089
Page 12
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

The above figure explains the relationship between the Simulation Time and
Utilization for different exponential interarrival times. It is clear that for interarrival time
exponential (2) the utilization is very less compared to interarrival time exponential (0.005).
It is known that when the interarrival time between the packets is less, then the more
number of packets can be transmitted which indicates that the utilization of the bus is more.
So it is clear in the above figure that based on the interarrival times the utilization of the bus
is varied having high utilization for low interarrival times.

Collision Count on node_0

To view the collision count on the node_0 the following procedure is to be followed.
Expand the node_0 hierarchy and Collision Count and then tick the required interarrival
times. Here in this report the collision count is studied for 0.03 and 0.005 interarrival times.
In the presentation tab select Stacked Statistics and As Is and then click the show button.
Keeping this panel visualized in minimized size also select a new panel with Average instead
of As Is in the presentation tab. Select add button and click on the previous panel to see the
average value. Doing this gives the following figure 9

For
Run
5

For
Run
9

Figure 9: Collision Count on node_0

In the above figure the collision count on node_0 is studied for two interarrival
times, for 0.03 seconds and 0.005 seconds. It is clear that the average collision count for 0.3
seconds interarrival time is lower than the collision count for 0.005 seconds interarrival time
but the difference is that for the interarrival time 0.03 seconds the collision count
dramatically decreases initially and then remains constant after a few simulation time
seconds whereas for the interarrival time 0.005 the collision count increases at the initial
simulation time seconds and remains constant later.

1037089
Page 13
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

As the inter-arrival time decreases the number of packets using the bus also
increases which in turn increases the offered load leading to the collision of the packets.
This results in increase in the collision count.

Why does the throughput change while the offered load increases?

Ans: As the offered load increases the throughput decreases because, theoretically the
Coaxial link data rate is limited to 10Mbps but in practical it can allow only up to 8 Mbps, so
when the offered load increases beyond the 8Mbps the collision has to occur between the
packets, but CSMA/CD protocol prevents the collision but leads to the delay of the packets.
Due to this delay the throughput of the network is changed to non-linearity.

Task 2: Analysis with different Packet sizes:

Part 1: The formula to calculate the offered load is given by the following equation

Offered Load= No of stations × Length of Packet (in bits)

Average Packet Interarrival Time

Here in this report the number of stations considered as 8. So for different Interarrival time
and Packet Size the offered load is calculated and presented in the tabular column given
below

No: of Packets: 8 Packet Size (bytes)

Interarrival Time
(sec) 1500 1024 512 46
2 48000 bps 32768bps 16384bps 1472bps
1 96000bps 65536bps 32768bps 2944bps
0.1 960000bps 655360bps 327680bps 29440bps
0.05 1920000bps 1310720bps 655360bps 58880bps
0.03 3200000bps 2184533.3 bps 1092266.7bps 98133.3bps
0.02 4800000bps 3276800 bps 1638400bps 147200bps
0.01 9600000bps 6553600bps 3276800bps 294400bps
0.008 12000000bps 8192000bps 4096000bps 368000bps

1037089
Page 14
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

0.005 19200000bps 13107200bps 6553600bps 588800bps

Part 2 a):

Comparison of node 0’s collision count and bus throughput for three different packet sizes
between three inter-arrival times

Three different scenarios are generated for each packet size by duplicating the first
scenario generated in Question 1 and renaming as Question_2a, Question_2b and
Question_2c by changing the Packet Size (bytes) attribute in the Packet Generation
Arguments for all nodes with Constant(1500), Constant(512) and Constant(46)
respectively.

In the Configure/Run DES dialog box the inter-arrival time are changed by allowing
only three inter-arrival times, exponential(1), exponential(0.1) and exponential(0.01). And
then run the individual scenarios.

For packet size 1500 bytes

Collision Count represents the number of collisions encountered by the specified Ethernet
MAC in the node. This does not incorporate the effects of background utilization traffic.
(Source: OPNET Network Simulation Software, Version: 14.5)

Node 0’s collision count:

1037089
Page 15
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 10: node 0’s collision count for packet size 1500 bytes

The node 0’s collision count for three inter-arrival times in the above figure states
that for exponential(1) inter-arrival time the collision count is constant irrespective of the
simulation time. While the collision count for exponential(0.1) inter-arrival time appears as
6 in average at nearly 24 sec and decreases and remains constant. But the collision count for
exponential(0.01) inter-arrival time is very high compared to two other inter-arrival times
and has a great variations. It is observed that there is a sudden raise in the number of
collision count at the initial time and gradually decreases from the peak to remain constant.

Bus throughput:

Throughput represents the average number of bits successfully received or transmitted by

the receiver or transmitter channel per unit time, in bits per second. (Source: OPNET
Network Simulation Software, Version: 14.5)

1037089
Page 16
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 11: Bus throughput for packet size 1500 bytes

As the inter-arrival time increases the number of packets sharing the bus also
increases which in turn causes delay in the delivery of the packets to the receiver. This
decreases the throughput of the bus. It is clear from the above figure that for exponential(1)
inter-arrival time the bus throughput is very less compared to exponential(0.1) and
exponential(0.01) inter-arrival times. The bus throughput for exponential(0.01) is very high
which is clearly shown in the above figure. The bus throughput at the starting time remains
the same for every inter-arrival time.

For packet size 512 bytes

Node 0’s collision count

The node 0’s collision count for three inter-arrival times in the figure 12 states that
for exponential(1) inter-arrival time the collision count is constant with respect to the
simulation time. While the collision count for exponential(0.01) inter-arrival time is higher
and is varied un regularly with respect to the simulation time. It is observed that for
exponential(0.1) inter-arrival time there is no collision count in the graph given below
because the offered load for 512 bytes packet size in the exponential(0.1) inter-arrival time
is 327680 bits which is less than data rate of the Coaxial link. So the collision does not occur
and there is no delay in the delivery of the data for 512 bytes packet size in the
exponential(0.1) inter-arrival time.

1037089
Page 17
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 12: node 0’s collision count for packet size 512 bytes

Bus throughput

Figure 13: bus throughput for 512 bytes packet size

The explanation for above figure is the same as the bus throughput explanation for
the packet size 1500 bytes where the throughput of the bus decreases with increase of the
inter-arrival times.

1037089
Page 18
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

For packet size 46 bytes

Node 0’s collision count

The node 0’s collision count for three inter-arrival times in the figure 13 states that
for exponential(1) inter-arrival time the collision count is constant with respect to the
simulation time. While the collision count for exponential(0.01) inter-arrival time is higher
and is varied un regularly with respect to the simulation time. It is observed that for
exponential(0.1) inter-arrival time there is no collision count in the graph given below
because the offered load for 46 bytes packet size in the exponential(0.1) inter-arrival time is
29440 bits which is less than data rate of the Coaxial link. So the collision does not occur and
there is no delay in the delivery of the data for 46 bytes packet size in the exponential(0.1)
inter-arrival time.

Figure 13: node 0’s collision count for 46 bytes packet size

Bus throughput

The explanation for the bus throughput of the network for the packet size of 46 bytes is
same as the bus throughput explanation for the network with 1500 packet size. The
difference which can be observed that for the inter-arrival time exponential(0.001) the
throughput is starting at 60000 bits/sec at the zero simulation time.

1037089
Page 19
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 14: bus throughput for the packet size of 46 bytes

Part 2b) comparison

Comparison of node 0’s collision count between three packet sizes at 0.01 inter-arrival time

Figure 15: Comparison of node 0’s collision count between three packet sizes at 0.01 inter-
arrival times

1037089
Page 20
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

It is clear from the above figure that as the packet size increases the collision count
also increases because as the packet size increases the offered load increases which results
in collision.

Comparison of bus throughput between three packet sizes at 0.01 inter-arrival time

Figure 16: Comparison of bus throughput between three packet sizes at 0.01 inter-
arrival time

It is clear from the above figure that as the packet size increases the bus
throughput also increases because as the packet size increases the offered load
increases which in turn results in delay of the delivery of the data.

1037089
Page 21
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Task 3: Performance of TCP and UDP Protocols

A network with TCP and UDP traffic is designed similar to the Ethernet as designed in
Question_1 with a few modifications. Instead of ethcoax_station use ethcoax_wkstn in the
Create the Network tab.

From the Protocol menu, in the ‘Create Application Demands’ dialog box the
parameters are changed as followed Constant(5) for Start time, Constant(256) for the Size,
Constant(7200) for the Rate in the Request parameter section. In the Response Parameter
section Size is changed to Constant(1024).

In the Configure/Run Simulation DES tab parameters are to be changed as

followed. 30 seconds as duration, 0 as Seed and 100000 as Update interval. In the ‘Inputs-
Object Attributes’ the promoted Transport Protocol is assigned with TCP and UDP as
multiple values.

Figure 17: Network model with TCP and UDP traffic

Part 1) the Ethcoax delay

Delay is defined as the end to end delay of frames accepted by all Ethernet layers in
the network. This statistic is measured from the time a frame is submitted to the Ethernet
layer for transmission to the time the frame is received by the Ethernet layer of the
Destination node. (Source: OPNET Network Simulation Software, Version: 14.5)

1037089
Page 22
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 18: Ethcoax delay for TCP and UDP traffic

From the above figure it is known that the average Ethcoax delay for TCP is very less
compared to the average Ethcoax delay for UDP. It is known that the TCP is a connection
oriented service while UDP is a connection-less service, due to this reason in the TCP after
the establishment of the connection each packet from the given node follows the sequential
order and results in less collision and thus delay of the packets in delivering will be less
compared to UDP. While UDP is a connection-less services all packets from each node are
sent at a time and results in the delay of the delivery of the packets to the receiver.

Part 2) Queuing delay

Queuing delay is defined as the instantaneous measurements of packet waiting

times in the transmitter channel's queue. Measurements are taken from the time a packet
enters the transmitter channel queue to the time the last bit of the packet is transmitted.
(Source: OPNET Network Simulation Software, Version: 14.5)

Since TCP is a connection oriented service each node is allocated a particular time to
share the bus for the delivery of the packets, so each packet in the node has to wait for its
turn which results in the Queuing delay of the packets in the connection oriented service.
While UDP is being connection-less service every node sends the packets at a time without
any establishment of connection, but delay occurs in the Ethcoax link or channel by
CSMA/CD protocol.

1037089
Page 23
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 19: Queuing delay along with Ethcoax delay for TCP and UDP

From the above figure it is clear that Ethcoax delay is more for UDP than TCP and
Queuing delay is more for TCP than UDP. The reason why it is, is already explained above. .
Queuing delays occur when packets are sent to a transmitter which is already busy
transmitting a packet.

Part 3) the Bus throughput

Throughput (bits/sec) is defined as the average number of bits successfully received

or transmitted by the receiver or transmitter channel per unit time, in bits per second.
(Source: OPNET Network Simulation Software, Version: 14.5)

Figure 20: bus throughput for TCP and UDP Protocols

1037089
Page 24
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

The above figure explains that the bus throughput for the both TCP and UDP
connections are nearly the same, with TCP a little higher than UDP. Since TCP is a
connection oriented service there is a guarantee in the delivery of the data and the number
of bits received per second by the destination node will be better than UDP, which is a
connection-less service.

Part 4) node 7’s Application Demand- Response time

Response Time is defined as the time elapsed between the sending of the request from the
source and the reception of the response at the source. (Source: OPNET Network Simulation
Software, Version: 14.5)

Figure 21: node 7’s response time for TCP and UDP

The above figure explains that the response time for TCP is constant while for UDP it
is irregular. Since TCP is a connection oriented service, it exhibits three-way handshake
before establishing a connection and a four-way handshake during breaking of the
connection, therefore the average response for the TCP is constant for all the simulation
times. But UDP being connection less service, there is no proper response time and the
response time would regulate unevenly with respect to the simulation time.

Part 5) low-level bus BER

Bit Error Rate: (BER) is defined as the average bit error rate of the channel (Source: OPNET
Network Simulation Software, Version: 14.5)

1037089
Page 25
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 22: low-level bus BER for TCP and UDP

The above figure explains that average bus BER for UDP is higher than TCP but it remains
constant compared to average bus BER for TCP. The reason is that TCP is being connection
oriented service the channel introduces the error to the bits during their transmission
irregularly and since UDP is connection less service there is no channel to introduce un
irregular errors to the data packets and thus has constant BER for UDP. But UDP BER is very
high compared to TCP BER because it exhibits unreliable connection.

1037089
Page 26
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Task 4: the effects of introduction of errors on the Bus

To introduce the errors on the bus a new scenario is created and named as Question_4
changing the BER attribute to 0.00005.

Part 1) average bus throughput for both TCP and UDP

Figure 23: average bus throughput for both TCP and UDP

Comparison of bus throughput with question_3

Figure 24: comparison of average bus throughput with Question_3

1037089
Page 27
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

The above figure clearly explains that, the decrease of BER of bus link alters the
average bus throughput of the TCP and UDP in the network. It is also clear that TCP bus
throughput is less than UDP bus throughput with the decrease of BER. It is already studied
that the bus throughput of the TCP is more than UDP while BER is 1s that is when it is
higher. This explains that BER alters the bus throughput of both TCP and UDP.

Part 2) low-level bus average bit error rate

Figure 25: low-level bus bit error rate for both TCP and UDP for Question_4

The explanation for the above figure is same as Task 3 Part 5), but the difference
observed is that the average BER decreases greatly with the simulation time in this Task, this
is due to decrease of BER attribute value for the bus.

1037089
Page 28
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Comparison of low-level bus BER of Question_4 with Question_3

Figure 26: comparison of low-level BER between Question_3 and Question_4

The above figure describes the comparison of average low level bus bit error rate for
Question_3 and Question_4. It is clear that with the decrease of BER of bus link the average
low-level bus BER for TCP is increased while for UDP it almost remains the same but a little
less which can be neglected.

Part 3) TCP’s retransmission Count

Retransmission Count is defined as the total number of TCP retransmissions in the network.
Written when data is retransmitted from the TCP unacknowledged buffer. (Source: OPNET
Network Simulation Software, Version: 14.5)

In brief it can be said as the number of packets requested by the destination node to
the source node for retransmission when the destination node fails to get the packets.
When any error occurs during the transmission the data fails to reach the destination which
is acknowledged by the destination node. TCP is a connection oriented service which works
with a three-way handshake before establishing a connection and a four-way handshake
before breaking up the connection.

1037089
Page 29
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Figure 27: TCP’s retransmission count with BER

The above figure describes the TCP’s retransmission count along with the BER setting
as studied in the figure 26. The blue line is the retransmission count for the TCP. This figure
explains that TCP retransmission count is higher than the average low-level bus BER of both
TCP and UDP. The retransmission increases abruptly at the initial seconds and starts
decreasing gradually and varying irregularly.

It is known that since UDP is a connection less service, there is no handshaking or

acknowledgement of the reception of the packet. So if the packet fails to reach the
destination node there is no request for retransmission and hence there is no
retransmission count for UDP.

1037089
Page 30
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Conclusion:
The Ethernet network model designed with Coaxial link of data rate 10Mbps is
observed that, its performance is dependent on many parameters which includes offered
load, packet size, inter-arrival time, number of stations in the model, TCP and UDP protocols
and bit error rate (BER) and so on. As the packet size increases the offered load increases
which in turn decreases throughput of the network. As the inter-arrival time decreases the
offered load increases which in turn increases throughput of the network. It is observed that
the collision count increases with the increase of the packet size.

The TCP and UDP protocols have different effects on Ethcoax delay, Queuing delay,
bus throughput, response time and low-level bus BER. Ethcoax delay, response time and
BER are more for UDP than TCP. Queuing delay and bus throughput is more for TCP than
UDP. These are due to their nature of connection; that is TCP is connection oriented and
UDP is connection less service.

In the task 4 it is observed that due to decrease of BER of the bus link, the bus
throughput is varied both for TCP and UDP. With the decrease of bus BER, the bus
throughput of TCP has become less compared to UDP when the bus BER is higher. The
retransmission count for Question_3 is found to be blank for BER 1 sec.

1037089
Page 31
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

References:
Koziniec, M. W. (June 2002). Using OPNET to Enhance Student Learning. Perth, Australia.

Skordoulis, R. a. (2009). Network Simulation Modelling and Analysis of an Ethernet-LAN using

OPNET. London: Brunel University.

Tanenbaum, A. S. (March 17, 2003). Computer Networks, Fourth Edition. New Jersey : Prentice Hall.

Thompson, A. (2000). Understanding Local Area Networks - A Practical Approach. New Jersey:
Prentice Hall.

W.Ross, J. F. (2010). Computer Networking- A Top Down Approach Fifth Edition. B0ston: Pearson.

Zin, M. N. (2008). Emulation Network Analyzer Development for Campus Environment and
Comparison between OPNET. malaysia: EuroJournals Publishing.

1037089
Page 32
 NETWORK SIMULATION MODELLING AND ANALYSIS OF AN ETHERNET-LAN USING November
OPNET 23, 2010

Bibliography
Tanenbaum, A. S. (March 17, 2003). Computer Networks, Fourth Edition. New Jersey : Prentice Hall.

Thompson, A. (2000). Understanding Local Area Networks - A Practical Approach. New Jersey:
Prentice Hall.

W.Ross, J. F. (2010). Computer Networking- A Top Down Approach Fifth Edition. B0ston: Pearson.

1037089
Page 33