simulation and results - shodhgangashodhganga.inflibnet.ac.in › bitstream › 10603 › 7955 ›...

97

CHAPTER 6

SIMULATION AND RESULTS

The protocol proposed in the previous chapter was simulated on the QualNet-5.0

software after making the requisite changes through programming. The QualNet-5.0

provides the base protocols, mobility patterns and energy models etc. The software can

be customized to implement a desired protocol by making suitable modifications in the

different modules. To identify these modules and to implement the changes was a

uphill task but thanks to the technical support received from Eigen Technologies, New

Delhi we were able to simulate the protocol and collected the results.

The QualNet-5.0 provides a comprehensive set of tools with all the components for

custom network modeling and simulation projects. QualNet's unparalleled speed,

scalability, and fidelity make it easy for modelers to optimize existing networks

through quick model setup and in-depth analysis tools. Models in source form provide

developers with a solid library on which to build and experiment with new network

functionality. The end result is accurate prediction of network performance for a

diverse set of application requirements and uses. From wired LANs and WANs, to

cellular, satellite, WLANs and mobile ad hoc networks, QualNet's library is extensive.

Because of its efficient kernel, QualNet models large scale networks with heavy traffic

and mobility in reasonable simulation times.

We now start with our discussion on simulation setup.

6.1 SIMULATION SETUP

For implementing the proposed protocol, an environment had to be created with certain

fixed and variable parameters. These parameters were suitably chosen for carrying out

the simulation process and the proposed protocol was implemented to check the

outcome. A detailed list of parameters chosen is shown in Table 6.1.

98

Table 6.1

Set up Parameters

Examined Protocols AODV, AODV-n

Simulation Period 30-2000 sec

Simulation Area 1500 X 1500 sq. mt

Number of Nodes 50, 60

Traffic Type CBR (UDP)

Energy Model Mica- Motes

Communication Model IEEE 802.11

Battery Model Linear

Default Battery 1200mah

Data Rate 4096 bps

Pay load size 512 byte

Trust update interval 1 sec

Number of malicious nodes 0%-50% step 10%

Number of selfish nodes

0% to 100% step 10% for studying Network Lifetime

0% to 50% step 10% for other metrics

The snapshots of the simulation process have been shown in Fig 6.1(a) and Fig. 6.1 (b).

99

Fig 6.1(a) Snapshot of Simulation process (Nodes = 50)

Fig 6.1(b) Snapshot of Simulation process (Nodes = 60)

100

6.2 DESCRIPTION OF PARAMETERS CHOSEN

Examined Protocols: As described in the previous chapter, we chose the

standard AODV protocol from the QualNet-5.0 software and simulated it to

collect the data about various parameters. Thereafter, we simulated our own

protocol AODV-n and compared its performance with normal AODV.

Simulation period: Depending upon the requirement of the metric to be

studied, different simulation periods were chosen. The maximum simulation

period chosen was 2000 sec while studying the metric network lifetime. The

normally simulation period was 30 sec.

Simulation Area: For performing the simulation, a standard rectangular area of

the dimension 1500 X 1500 mt2 was taken.

Number of nodes: For performing the simulation two sets of readings were

taken one with 50 nodes and other with 60 nodes.

Traffic Type: Going by the convention the traffic type taken was Constant Bit

Rate (CBR).

Energy Model: To take into the account the energy dissipation due to the

various activities and the idle case scenario the energy model chosen was Mica-

Motes as suggested by Eigen technologies and as conveyed by the various

papers.

Communications Model: The IEEE standard 802.11 Distributed Coordination

Function (DCF) [131] was used as the MAC layer routing protocol. All Route

Request and Query packets were broadcasted using the un-slotted Carrier Sense

Multiple Access protocol with Collision Avoidance (CSMA/CA) wherein each

broadcasting node waits for a vacant channel by sensing the medium. If the

channel is vacant, it transmits. In case of a collision, the colliding stations wait

using the Ethernet binary exponential back off algorithm. To unicast packets,

the node first reserves the channel by transmitting a short Ready-to-Send (RTS)

101

frame. The intended recipient node, in response, sends a Clear-to-Send (CTS)

frame to the RTS sender. All nodes overhearing the RTS or CTS frames desist

from transmitting for the Network Allocation Vector (NAV) interval. Upon

receipt of the CTS, the packet is transmitted which is acknowledged by the

recipient.

Battery model: The power dissipation model of the battery was taken to be

linear to consider the uniform decay of power with time when the environment

is not changing.

The default battery, data rate and payload size were taken as per the standards

given in the literature and available on the QualNet-5.0 software.

Trust update interval: This parameter was of critical importance and had to be

optimised. A very short trust update interval will lead to very high overhead of

energy and bandwidth usage and a long trust update interval may not provide

the requisite information in time. Therefore an optimal time of 1 sec was chosen

keeping in similarity with hello interval as used by AODV.

Number of Malicious nodes: The malicious nodes strength was taken upto

50% of the total nodes strength.

Number of Selfish nodes: The selfish nodes strength was taken upto 50% of

the total nodes strength in most of the cases. However, while studying the

network lifetime metric it was the selfish node strength was raised to the level

of 100% to study the impact of such a situation.

6.3 ATTACK PATTERN TAKEN

Selfish Node Attack: In this attack, a node tries to utilize the network resources for its

own profit but is reluctant to spend its own for others since its residual battery power is

low/ very low. As the time passes, the nodes in an Ad-hoc network loose their battery

power and their chances of becoming selfish get increased.

102

Malicious Node Attack: In this attack, a malicious node dumps all the data / Control

packet which it is supposed to forward. It receives all the packets meant for it but at the

same time doesn’t forward the packets that are intended for others.

6.4 ASSUMPTIONS

The following assumptions were made for simulation purpose:

Each node declares its residual battery status correctly.

The malicious node work in individual manner and there is no such group.

The participating nodes are not in a position to modify the contents of control

packets.

6.5 METRICS USED

To evaluate the efficacy of the proposed protocol following metrics were used:

Successful Route Formation: Percentage of route successfully created to the

number of route requests generated by the source.

Average Hop Count: Average number of hops for all successful route

formation.

Throughput: How fast data can pass through a network. In our simulation

scenario it is the number of bits passing through the network in one second.

End-to-End delay: Time taken for a packet to travel from the CBR (Constant

Bit Rate) source to the destination.

Jitter: Occurs when in a transmission scenario different packets take different

amount of time in reaching from source to destination. Jitter can be measured by

using the standard deviation of packet delay. If a communication system has

large amount of jitter then the signal quality is very poor.

Packet delivery ratio (PDR): Ratio of number of data packets successfully

received at the destinations to the total number of data packets sent by various

sources.

Network Lifetime: Time at which first node of network gets dead.

103

Or

Time at which all the node of the network gets dead.

Probability of Reachability: Defined as every fraction of possible reachable

routes to the all possible routes between all different sources to all different

destinations.

6.6 RESULTS AND DISCUSSIONS

6.6.1 SUCCESSFUL ROUTE FORMATION

Fig. 6.2(a) and Fig. 6.2(b) show the results of successful route formation in the

presence of malicious nodes varying from 0 to 50% of the total node strength (60/50). It

was observed that percentage of successful route formation was almost same in case of

plain AODV irrespective of the malicious node concentration. There was a significant

drop in the successful route formation with the increase in malicious node

concentration in case of AODV-1, 2, 3. In the case of AODV-3, the successful route

formation reduces from 70% to 15% when the concentration of malicious nodes

increases from 0 to 50%. While in the case of AODV-2, the reduction was from 70% to

25% and in case of AODV-1 the reduction was found to be from 70% to 28%. The

results conveyed that the choice of a higher trust index threshold is going to have a

large impact on the percentage of successful route formation. Though there was a

decrease in the case of AODV-0, but this drop was small. This showed that though a

passage was offered to the malicious nodes yet their blacklisting had an impact on the

successful route formation leading to the non formation of certain routes.

In case of selfish node scenario, the successful route formation was close to 100% in

case of plain AODV irrespective of selfish node concentration as shown in Fig. 6.2(c)

and Fig. 6.2(d). A small drop was observed in case of AODV-0 with the increase in

selfish node concentration. Though there was a decrease in the percentage of route

formation in case of AODV-1, 2, 3 with the increase in selfish node concentration but

this decrease was much less than in case of malicious node scenario. The underlying

104

reason for this being the allowance to form the route for selfish nodes upto w number of

times (5 in our case) .

Successful Route Formation

0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50

Malicious Node Percentage

Perc

en

tag

e R

ou

te F

orm

ati

on

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.2(a) Successful Route Formation with malicious nodes (Nodes = 60)


0

20

40

60

80

100

120

0 10 20 30 40 50

Selfish Node Percentage

Perc

en

tag

e o

f S

uccessfu

l R

ou

tes

AODV

AODV- 0

AODV-1

AODV-2

AODV-3

Fig 6.2(b) Successful Route Formation with selfish nodes (Nodes =60)

105


0

10

20

30

40

50

60

70

80

90

100

0 10 20 30 40 50


Perc

en

tag

e R

ou

te F

orm

ati

on

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.2(c) Successful Route Formation with malicious nodes (Nodes =50)


0

20

40

60

80

100

120

0 10 20 30 40 50


Perc

en

tag

e o

f S

uccessfu

l R

ou

tes

AODV

AODV- 0

AODV-1

AODV-2

AODV-3

Fig 6.2(d) Successful Route Formation with selfish nodes (Nodes = 50)

106

6.6.2 AVERAGE HOP COUNT

Fig. 6.3(a) and Fig 6.3(b) show the results for average hop count in case of malicious

node environment. It was observed that the average hop count showed no change for

plain AODV with the increase in malicious node concentration. There was marginal

increase in hop count in case of AODV-0 with the increase in malicious node

concentration. The trend in case of AODV-1, 2, 3 showed an increase initially but a

decrease later on as the malicious node concentration increased. A careful scrutiny of

the scenario showed that when the malicious node concentration was very high the

successful route formation was limited to immediate neighbors or their next neighbors

in most of the cases. Similar results were observed in case of selfish node scenario as

shown in Fig. 6.3(c) and Fig. 6.3(d).

Average Hop Count for Successful Routes

0

0.5

1

1.5

2

2.5

3

3.5

4

0 10 20 30 40 50

Malicious node Percentage

Avera

ge H

op

Co

un

t

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.3(a) Average Hop Count for malicious nodes (Nodes = 60)

107


0

1

2

3

4

0 10 20 30 40 50

Selfish node Percentage

Avera

ge H

op

Co

un

tAODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.3(b) Average Hop Count for malicious nodes (Nodes = 60)


0

1

2

3

4

0 10 20 30 40 50

Malicious node Percentage

Avera

ge H

op

Co

un

t

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.3(c) Average Hop Count for malicious nodes (Nodes = 50)

108


0

1

2

3

4

0 10 20 30 40 50

Selfish node Percentage

Avera

ge H

op

Co

un

tAODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.3(d) Average Hop Count for selfish nodes (Nodes = 50)

6.6.3 THROUGHPUT

To measure the throughput the data was sent at the rate of 4096 bits/sec. The results for

the throughput are shown in Fig. 6.4(a) to Fig. 6.4(d). The plain AODV works well

when there is no malicious or selfish node. As the malicious node concentration

increases the throughput of plain AODV decreases very fast and reduces to 55% as the

malicious node concentration node reaches to 50%. However, the AODV-1, 2, 3

protocols manage to resist the impact of malicious nodes and the throughput drops to

the level of nearly 70%. Also it is worth noticing that when the malicious node

concentration is to the level of 30 to 40%, the AODV-1, 2, 3 protocols have the

throughput to the level of nearly 85 to 90%. In the selfish environment the throughput

of the AODV-1, 2, 3 protocols is nearly 95-100%. This is due to the fact that the nodes

are obligated to speak truth about their current power levels in these protocols. There is

a significant decrease in the performance of plain AODV with the increase in selfish

node concentration. The performance of AODV-0 was fluctuating in case of selfish and

malicious nodes environment.

109

Throughput

0

500

1000

1500

2000

2500

3000

3500

4000

4500

0 10 20 30 40 50


Th

rou

gh

pu

t o

f R

eceiv

er

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.4(a) Average Throughput for malicious nodes (Nodes = 60)

Throughput

3300

3400

3500

3600

3700

3800

3900

4000

4100

4200

0 10 20 30 40 50


Perc

en

tag

e T

hro

ug

hp

ut

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.4(b) Average Throughput for selfish nodes (Nodes= 60)

110

Throughput

0

500

1000

1500

2000

2500

3000

3500

4000

4500

0 10 20 30 40 50


Th

rou

gh

pu

t o

f R

eceiv

er

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.4(c) Average Throughput for malicious nodes (Nodes= 50)

Throughput

3300

3400

3500

3600

3700

3800

3900

4000

4100

4200

0 10 20 30 40 50


Perc

en

tag

e T

hro

ug

hp

ut

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.4(d) Average Throughput for selfish nodes (Nodes= 50)

6.6.4 END TO END DELAY

It indicates the time taken for a packet to travel from the CBR (Constant Bit Rate)

source to the destination. It represents the average data delay that an application

111

experiences while transmitting data. At low concentration level of malicious nodes, the

end to end delay is much lower in case of plain AODV and as the concentration of

malicious nodes increases the average delay is much lower in case of our protocol

AODV-n as shown in Fig. 6.5(a) to Fig. 6.5(d).

Average End to End Delay

0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50

Malicious Nodes

Dela

y in

seco

nd

s AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.5(a) Average End to End delay for malicious nodes (Nodes = 60)


0

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0.04

0 10 20 30 40 50

Selfish Nodes

Dela

y in

seco

nd

s AODV

AODV- 0

AODV-1

AODV-2

AODV -3

Fig. 6.5(b) Average End to End delay for selfish nodes (Nodes = 60)

112


0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50

Percentage of Malicious Nodes

Dela

y in

seco

nd

s AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.5(c) Average End to End delay for malicious nodes (Nodes = 50)


0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50

Percentage of Selfish Nodes

Dela

y in

seco

nd

s AODV

AODV- 0

AODV-1

AODV-2

AODV -3

Fig. 6.5(d) Average End to End delay for selfish nodes (Nodes = 50)

The performance of plain AODV and AODV-n in the selfish environment was found to

be fluctuating with no consistent change in one direction, on analysis it was found that

as the selfish node concentration increases most of the routes formed were of low hop

count of the order of 1 or 2.

113

6.6.5 JITTER

In a communication scenario stream line flow of data packets is necessary where in all

the data packets follow their preceding packets with the same speed. In such a scenario

the output will be smooth without any turbulence. If such a situation doesn’t exist then

the output is jerky and the jerks can be felt in the video/audio output. Flow of data

packets will be streamline if each data packet takes equal time for traveling from source

to destination. If there is a variation between the traveling times of different packets

then it will cause a jitter. The jitter can be computed by measuring the average traveling

time of each packet on a particular path and applying the standard deviation on the

traveling time. Larger the standard deviation more prominent is the jitter effect [133-

135]. Fig. 6.6 (a) to Fig 6.6 (d) show the jitter encountered by the control/data packets

in different protocols.

Jitter

0

0.002

0.004

0.006

0.008

0.01

0.012

0.014

0.016

0.018

0.02

0 10 20 30 40 50

Malicious Nodes

Jit

ter

AODV

AODV w ith Priority level =

0


1


2


3

Fig. 6.6(a) Average Jitter for malicious nodes (Nodes = 60)

114

Jitter

0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50

Selfish Nodes Percentage

Sta

nd

ard

devia

tio

n

AODV

AODV-0

AODV-1

AODV- 2

AODV-3

Fig. 6.6(b) Average Jitter for selfish nodes (Nodes = 60)

Jitter

0

0.005

0.01

0.015

0.02

0.025

0 10 20 30 40 50

Malicious Nodes Percentage

Sta

nd

ard

devia

tio

n

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.6(a) Average Jitter for malicious nodes (Nodes = 50)

115

Jitter

0

0.01

0.02

0.03

0.04

0.05

0.06

0 10 20 30 40 50

Selfish Nodes Percentage

Sta

nd

ard

devia

tio

n

AODV

AODV-0

AODV-1

AODV- 2

AODV-3

Fig. 6.6(d) Average Jitter for selfish nodes

(Nodes = 50)

6.6.6 PACKET DELIVERY RATIO (PDR)

It is the ratio of number of data packets successfully received at the destinations to the

total number of data packets sent by various sources. At low concentration level of

selfish and malicious nodes, the PDR of plain AODV is nearly same as that of AODV-

1, 2, 3 but as the concentration of selfish and malicious nodes increases the PDR for

plain AODV was found to be much lower in comparison to our protocol AODV-n as

shown in Fig. 6.7. On analysis it was found that as the concentration of selfish and

malicious nodes increases the nodes begin to drop data packets due to low battery

power or due to their rogue behavior. The performance of AODV-0 is the poorest of all

the protocol where time is taken for identification of malicious and selfish nodes but the

information remains unused. This type of the situation is bad in both the cases as the

time is wasted in identification and both malicious and selfish are inserted in the route.

116

Packet Delivery Ratio

0

20

40

60

80

100

120

0 10 20 30 40 50


Perc

en

tag

e P

acket

Delivery

Rati

o

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.7(a) Average Packet Delivery Ratio for malicious nodes (Nodes = 60)

Packet Deliver Ratio (PDR)

0

20

40

60

80

100

120

0 10 20 30 40 50


Perc

en

tag

e P

acket

Delivery

Rati

o

AODV

AODV- 0

AODV-1

AODV-2

AODV-3

Fig. 6.7(b) Average Packet Delivery Ratio for selfish nodes (Nodes = 60)

117

Packet Delivery Ratio

0

20

40

60

80

100

120

0 10 20 30 40 50


Perc

en

tag

e P

acket

Delivery

Rati

o

AODV

AODV-0

AODV-1

AODV-2

AODV-3

Fig. 6.7(c) Average Packet Delivery Ratio for malicious nodes (Nodes = 50)

Packet Deliver Ratio (PDR)

0

20

40

60

80

100

120

0 10 20 30 40 50


Perc

en

tag

e P

acket

Delivery

Rati

o

AODV

AODV- 0

AODV-1

AODV-2

AODV-3

Fig. 6.7(d) Average Packet Delivery Ratio for selfish nodes (Nodes = 50)

6.6.7 NETWORK LIFETIME

The result shows the impact of selfish node concentration on network lifetime as the

percentage of selfish nodes increases. Fig 6.8(a) shows the network lifetime on the

basis of getting down of first node. Fig. 6.8(b) show the network lifetime on the basis

118

of getting all nodes dead. Table 6.2 shows the simulation parameters used in the

scenario. The battery was taken to be 2 mah for the normal node and for the selfish

node was taken to be 0.012 mah.

Table 6.2

Set Up Parameters for Network Lifetime

Default battery Power 2mah

Selfish node Battery Power 0.012mah

Energy model Mica- Motes

Battery Dissipation Model Linear

6.6.8 PROBABILITY OF REACHABILITY (PoR)

The results show the impact of selfish nodes concentration on the PoR value as the

percentage of selfish nodes increases. It can be easily seen from the graph of Fig. 6.9

(a) and Fig. 6.9(b) that as the percentage of selfish and malicious nodes increases the

number of reachable paths becomes quite low for both the routing protocols (AODV,

AODV-n). It is important to note here that even at 100% of selfish and malicious node

concentration there is still communication between nodes. The underlying reason for

the same is malicious and selfish behavior of nodes is for others and not for themselves.

Network Lifetime

0

200

400

600

800

1000

1200

1400

1600

0 20 40 60 80 100

Selfish Nodes

Tim

e a

t w

hic

h F

irst

no

de g

ets

Dead

(Seco

nd

s)

AODV w ith hello enabled

AODV w ith Priority Level

= 0


1


= 2


3

119

Fig. 6.8(a) Network Lifetime for first node

Network Lifetime

0

200

400

600

800

1000

1200

1400

1600

0 20 40 60 80 100

Selfish Nodes

Tim

e a

t w

hic

h a

ll t

he n

od

es g

et

Dead

(S

eco

nd

s)

AODV w ith hello enabled


= 0


1


= 2


3

Fig. 6.8(b) Network Lifetime for all nodes

Probability of Reachability

0

20

40

60

80

100

120

0 10 20 30 40 50 60 70 80 90 100

Percentage of Selfish Nodes

Percen

tag

e P

ro

bab

ilit

y o

f

Reach

ab

ilit

y

AODV with Hello Disabled

AODV with Priority level = 3

AODV with Priority Level = 2


AODV With Priority Level = 0

Fig. 6.9 (a) Probability of Reachability for selfish nodes

120

Probability of Reachability

0

20

40

60

80

100

120

0 10 20 30 40 50 60 70 80 90 100

Percentage of Malicious Nodes

Percen

tag

e P

ro

bab

ilit

y o

f

Reach

ab

ilit

y

AODV with Hello Disabled

AODV with Priority level = 3



AODV With Priority Level = 0

Fig. 6.9 (b) Probability of Reachability for malicious nodes

6.7 ANALYSIS OF SIMULATION AND RESULTS

The analysis of simulation results indicate that in an environment devoid of malicious

and selfish nodes plain AODV performs much better than AODV-n. As the

concentration of malicious/ selfish nodes increases AODV-1, 2, 3 outperform the plain

AODV. Many times, the performance of the AODV-1, 2, 3 protocols are better in an

environment with large concentration of malicious/selfish nodes (40-50%) than the

performance at the lower concentration of malicious/selfish nodes (10-30%). A careful

scrutiny of the data showed that at high concentration of malicious/selfish nodes most

of the routes formed were of the low hop count which allowed for easy passage of data

packets leading to comparatively small amount of end to end delay and jitter. The

performance of AODV-0 was the poor most where in the resources were used to collect

data and compute the trust class but this information was not used to select the trust

worthy intermediate nodes. Another note worthy feature of the simulation result was

that there was not much difference between the performance of AODV-1, 2, 3

indicating that the choice of lower trust class works in the equally effective manner as

in case of higher trust class. The network life time is very high when no node in the

121

network is selfish but as the concentration of selfish nodes it decreases very fast as

shown in Fig 6.7(a), (b). On analysis it was observed that the energy dissipation in the

nodes was mainly due to hello packets and not because of data packets sent from source

to destination.

simulation and results - shodhgangashodhganga.inflibnet.ac.in › bitstream › 10603 › 7955 ›...

Documents