Impact of Mobility Models on Routing Protocols for Various Traffic Classes in Mobile Ad Hoc Networks

IMPACT OF MOBILITY MODELS ON ROUTING PROTOCOLS FOR VARIOUS TRAFFIC CLASSES IN MOBILE AD HOC NETWORKS

A thesis submitted

to Kent State University in partial

fulfillment of the requirements for the

degree of Master of Science

Hayder Majid Abdulhameed Alash

May 2016

Except for previously published materials

Thesis written by

Hayder Majid Abdulhameed Alash

B.S., University of Baghdad, 2009

M.S., Kent State University, 2016

Approved by

Hassan Peyravi, Adviser

Javed Khan, Chair, Department of Computer Science

James L. Blank, Dean, College of Arts and Sciences

TABLE OF CONTENTS

LIST OF FIGURES………………………………………………………………..….VII

LIST OF TABLES…………………………………………………………………...…IX

DEDICTION………………………………………………………………………………

ACKNOWLEDGMENT…………………………………………………………………

1 INTRODUCTION...... 1

1.1 Mobile Ad Hoc Network (MANET) ...... 2

1.2 Issues in Mobile Ad Hoc Networks (MANETs) ...... 4

1.2.1 Security……………….………………………………………...…………….4

1.2.2 Routing……………………………………………………………………….5

1.2.3 Scalability…….……………………………………………………….……...5

1.2.4 Quality of Service……………………...………………………………..…...6

1.3 Routing Protocols for MANETs ...... 6

1.3.1 Proactive (Table driven) ...... …...6

1.3.2 Reactive (On demand) ...... …...... 7

1.3.3 Hybrid………………………...... 7

1.4 Other Issues Related to Quality of Service in MANETs...... 8

1.4.1 Unpredictable Link Properties ...... …...8

1.4.2 Node Mobility...... 8

1.4.3 Hidden and Exposed Terminal...... 9

1.4.4 Limited Battery Life………………...... 9

1.4.5 Route Maintenance…………...... 9

1.5 Thesis Organization ...... 10

2 SURVEY OF PERVIOUS WORK ...... 11

2.1 Mobility Models ...... 11

2.1.1 Random Waypoint Mobility Model ...... 13

2.1.2 Group Mobility Model ...... 15

2.1.2.1 Reference Point Group Mobility Model……………………………… ...16

2.2 Routing Protocols...... 17

2.2.1 Proactive Routing Protocols ...... 18

2.2.1.1 Bellman-Ford…………………………………………………………….19

2.2.1.2 Fisheyes State Protocol (FSR)………………………………………...…19

2.2.1.3 Optimized Link State Routing Protocol (OLSR)………………………...21

2.2.1.4 Source Tree Adaptive Routing Protocol (STAR)…………………….….23

2.2.2 Reactive Routing Protocols ...... ……..24

2.2.2.1 Ad Hoc On Demand Distance Vector Routing Protocol (ADOV)...... 25

2.2.2.2 Dynamic Source Routing Protocol (DSR)…………………….………....29

2.2.2.3 Dynamic MANT On Demand Routing Protocol (DYMO) …………..…33

2.2.2.4 Location Aided Routing Protocol (LAR) ……………………………….35

2.2.3 Hybrid Routing Protocols ...... 39

2.2.3.1 Zone Routing Protocol (ZRP)…………………….……………… ……..39

3 MOBILE AD HOC NETWORKS……………………………………….…………..42

3.1 Characteristics of Mobile Ad Hoc Networks (MANET) ...... 44

3.2 Applications ...... 45

4 SIMULATION MODEL ………………………………………………………...…. 48

4.1 Simulator ...... 48

4.2 Features ...... 49

4.3 Traffic Model ...... 50

4.3.1 Constant Bit Rate (CBR)……………………………………………………...51

4.3.2 Variable Bit Rate (VBR)……………………………………………………...52

4.3.3 Random Traffic………………………………...……………………………..52

4.3.4 File Transfer Protocol (FTP)…………………………………...…...………...53

4.4 Simulation Setup ...... 53

4.4.1 Simulation Time………………………………………………..……………..54

4.4.2 Mobility………………………………...………………...…………………...55

4.4.3 Node Placement and Network Shape…..………………...…………………...55

4.4.4 Physical Layer………………………………………………………….……..56

4.4.5 MAC Layer…………………………………………………………………...56

4.4.6 Network Layer………….…………………………………………….….…...56

4.4.7 Routing Protocols……………………………………………………….……56

4.5 Metrics ...... 57

4.5.1 Throughput …….……………………………………………………………..57

4.5.2 End-to-End Delay.…………………………………………………………....57

4.5.3 Average Jitter ….……………………………………………………………..58

5 SIMULATION RESULT…….………………………………………………………59

5.1 Throughput ...... 59

5.2 Average End-to-End Delay ...... 63

5.3 Average Jitter ...... 66

5.4 Modifying AODV Parameters ...... 68

6 CONCLUSION AND FUTURE WORK…………………………………………....70

6.1 Conclusion ...... 70

6.2 Future work ...... 71

GLOSSARY……………………………………………………………………………..72

REFERENCES………………..………………………………………………………...75

LIST OF FIGURES

Figure 1.1. Mobile Ad Hoc Network.……………………………………………….…. 3

Figure 2.1 Mobility Model Types………….…………………………………………... 13

Figure 2.2 Random Waypoint Mobility Model………………………………………... 14

Figure 2.3 Group Mobility Model……………………………………..……………..... 16

Figure 2.4 Classification of Routing Protocols……………………………………….... 18

Figure 2.5. Scope of Fisheye…….……………………………………………………….21

Figure 2.6. Multipoint Relays.………………………………………………………….. 22

Figure 2.7. Source Node Discovery Process…………………………………………… 25

Figure 2.8. A route Reply RREP Process ...…………………………………………… 26

Figure 2.9. Route Maintences …………..……………………………………………... 27

Figure 2.10. Build Record Route ….……...………………………………………...... 30

Figure 2.11. DSR Route Reply…………..…………………………………………….. 31

Figure 2.12. DSR Route Maintences…………………………………………….……... 32

Figure 2.13. Route Discovery in DYMO and AODV……………………..………….... 34

Figure 2.14. LAR Expected Zone………………….……………………………….…. 36

Figure 2.15. LAR Scheme 1- Request Zone... ………………………………………… 37

Figure 2.16. LAR Scheme 2 ……………….………………………………………….. 38

Figure 2.17. ZRP Zone …………………….………………………………………….. 41

Figure 3.1 Wireless Network …………….…………………………………………..... 43

Figure 5.1 Throughput in Random Waypoint Model…………….……………………. 60

Figure 5.2 Throughput in Group Mobility Model ……………….……………………. 62

vii

Figure 5.3 End-to-End Delay in Random Waypoint Model ………………………….. 64

Figure 5.4 End-to-End Delay in Group Mobility Model ……………………………... 65

Figure 5.5 Jitter in Random Waypoint Model ………………….…………………….. 66

Figure 5.6 Jitter in Group Mobility Model …………………………………………… 67

Figure 5.7 AODV Comparison ………………………………..……………………… 69

viii

LIST OF TABLES

Table 3.1 Simulation Parameters …………………………………………………..… 54

Table 5.1 Comparsion Results ……………………………………………………...... 68

DEDICATION

My Mother, who is always pray for me

My Father, who is always trust me

My Brother and Sister, I miss you more and more

My Love, I Love you forever…

To my Friends, who always support me

To my country, thank you for this gift

ACKNOWLEDGEMENTS

I would like to thank my advisor Dr. Hassan Peyravi for all his support and guidance throughout this thesis. He has developed my skills in many areas and assisted me in its writing. In addition, I would like to thank the other members of my committee,

Dr. Feodor F. Dragan, and Dr. Gokarna Sharma for the comments and revisions they have offered.

Thanks from the bottom of my heart to my beloved parents, my brother, and my sister who have always prayed for me, which became a guiding force in completing my studies successfully. Special thanks to my mother who is my everything for always being there for me. I would also like to thank my love and future wife Zainab who withstood pressure in my absence and hung on for a long period.

I am also so thankful to Marcy Curtiss, the graduate secretary. I am sincerely thankful to my friends who helped and supported me. Finally, yet importantly, I would like to extend special thanks to my country for allowing me to complete graduate studies at Kent State University.

Hayder Majid Abdulhameed Alash

April 1, 2016

Kent, Ohio

CHAPTER 1

Introduction

Mobile ad hoc networks (MANETs) are widely used in wireless networks consisting of mobile devices that communicate in the absence of any centralized support.

Examples of such networks include networks for trucks on interstate highways, wireless military battlefield networks that connect troops, aircraft, satellites, and sensors on land and in water, and interplanetary networks. Mobile devices in these networks will act as routers that generate user’s traffic and carry out network control and routing tasks. The mobility of devices in MANET dynamically changes the network topology, which makes routing between devices more complicated. When devices move, the impact could be very significant in terms of connectivity and Quality of Service (QoS).

Main challenges in QoS provisioning in MANET include dynamic bandwidth management to guarantee the end-to-end delay and throughput performance to satisfy the requirements for diverse applications. There are several factors that affect the QoS including mobility, routing algorithms, and traffic patterns. Recently, many researchers have developed several theoretical models to describe mobility, traffic patterns, and routing algorithms. In this thesis, we intend to conduct a comparative analysis of several routing algorithms under a few popular mobility models and diverse traffic patterns.

Specifically, we have developed simulation models that incorporate various mobility models, routing algorithms, and traffic sources to measure applications’ performance in

terms of end-to-end throughput (bit rate), latency, and jitter. Three classes of MANET routing algorithms (Proactive, Reactive, and Hybrid), two mobility models (Random

Waypoint and Group), and three classes of traffic patterns (constant bit rate, variable bit rate, and random) have been investigated. We have designed network topology based on randomly placed devices over a communication area. While we studied various simulation tools, we found that QualNet [1] offers many important analytical details to better assess the trade-offs across layers. This work provides network designers and network operators with significant insight about the relationship between mobility and routing, on one hand, and users and their applications, on the other hand, to effectively manage their network.

The rest of this chapter covers preliminaries and architectural issues in MANETs.

Specifically, we describe major mobility models, routing algorithms, routing maintenance, security, scalability, efficiency in terms of energy consumption, etc.

1.1 Mobile Ad Hoc Networks (MANETs)

Mobile Ad Hoc Networks are becoming more popular in recent years as an alternative to traditional wired networks. Wireless networks can be classified into three types: infrastructure networks, ad hoc network (infrastructure less), and hybrid networks that combine the two types. MANETs are independent systems for mobile devices connected by wireless links (see Fig. 1.1). They are similar to multi-hop wireless networks (MWNs), but they differ in terms of topology, architecture, and mobility. Each node in a MANET is free to move independently or as a group member in any direction. 2

A MANET has limited available resources (i.e., bandwidth) due to its dynamic environment. Throughput, delay, and delay variation are the most important metrics in

QoS assessment. Generally, applications may have additional requirements such as peak rate or sustainable peak rate requirements. Other requirements to support QoS include bandwidth, link delay, and error rate. It is hard to get this information because of dynamic change (node mobility). MANETs links are more exposed to higher bit-error rates than their wireless networks counterparts, and that causes fluctuations in link capacity.

Figure 1.1: Mobile Ad Hoc Networks

MANETs have different traffic types than fixed wireless networks. When we design a MANET, many problems can arise (such as routing, security, power consumption, quality of services, and reliability) due to shared wireless channel, limited transmission power of wireless devices, battery constrains, and mobility. Therefore, providing Quality of Services (QoS) in MANETs is more complex than in fixed wired and wireless networks.

1.2 Issues in Mobile Ad hoc Networks (MANETs)

Generally, MANETs were first proposed for military battlefield and disaster recovery communications. However, recent evolution in several application areas such as remote sensing, smart highways, remote environmental and animal movement outposts are based on ad hoc networks concepts. These applications require different QoS requirements. The bandwidth requirements vary from a few Kb/s to several Gb/s. Some are delay-sensitive, while others are loss-sensitive. Also, some are highly mobile and others may have limited mobility.

There are several issues in MANETs that are very difficult to integrate with internet. We will address some of them below.

1.2.1 Security

Security is an important issue in MANETs. In wireless networks, the link is more vulnerable to nose, error, and eavesdropping than a wired link. Providing security in the presence of mobility and wireless links is more challenging. Therefore, security is often

performed through encryption and/or physical layer spread spectrum modulation (direct sequence or frequency hopping). It is a difficult problem to find a trust channel.

1.2.2 Routing

Routing is one of the most difficult problems to implement in MANETs. Routing is the process of finding the best path to send data packets from a source to a destination.

Since every device acts as a router, the network becomes more complicated to manage.

This is because each node can move randomly in any direction within the network.

When a node moves, new paths need to be discovered and selected, as the optimal route in specific time might not work after a few seconds. Also, the environment can be changed from indoor to outdoor scenarios that cause a path to fail.

1.2.3 Scalability

The operation of MANETs strongly depends on network size and packet size.

Routing and finding feasible paths become more complicated with size. Similarly, packet size has major impact on forwarding. Scalability measures the ability of the network to provide an acceptable level of services as network grows in size and traffic. Routing protocols add more limitation for the scalability of MANETs. The dynamic topology of a

MANET creates a big challenge to provide the huge amount of broadcast message in a dynamic environment.

1.2.4 Quality of Services

Quality of Services (QoS) is a very challenging issue for the developers. It is harder to achieve high performance in MANET due to highly dynamic topology. The network should be able to provide the required quality of service for user’s demand. The performance can be characterized by delay, jitter, and bandwidth. It is difficult to maintain the quality of these parameters under mobility. In a MANET, cross-layer optimization is needed to achieve quality of service.

1.3 Routing protocols for MANETs

Routing protocols are used to find a path for transmission of packets from a source to a destination node. They should deal with the limitation in a MANET such as low bandwidth, high power consumption, and high error rates. Several routing protocols for MANETs have been developed. These routing protocols can act differently depending on the number of nodes, mobility, and type of traffic sources. The routing protocols can be categorized as: Proactive, Reactive, and Hybrid.

1.3.1 Proactive (Table Driven)

Every node in proactive routing maintains a table that contains information of the routes to any other node in the network. These tables are periodically updated due to network topology changes. Proactive routing continuously broadcasts the control message and updates tables. This operation consumes a portion of bandwidth that can be used for applications otherwise and can be considered as a waste of bandwidth.

Therefore, when the number of nodes in mobile ad hoc network increases, the size of data stored in the tables will increase. This is a major drawback of proactive routing. Major proactive routing algorithms include Bellman-Ford Routing protocol [2], Fisheye State

Routing protocol (FSR) [3], Optimized Link State Routing protocol (OLSR) [4], and

Destination Sequenced Distance Vector Routing protocol (DSDV) [5].

1.3.2 Reactive (On Demand)

Reactive routing protocols reduce the need for maintaining updated table information because a route is established only when the source node desires to send a packet to a destination. These protocols flood a message into the network to discover a route to a destination. The discovered route is maintained by route maintenance until either the route is no longer desired or the destination node becomes unreachable. The reactive routing protocols use bandwidth more efficient than proactive protocols, but they suffer from delay due to the route discovery procedure. There are several reactive routing protocols such as: Ad hoc On Demand Distance Vector routing protocol (AODV) [6],

Dynamic Source Routing protocol (DSR) [7], Dynamic MANET On Demand Routing protocol [8], and Location Aided Routing protocol (LAR) [9].

1.3.3 Hybrid

Hybrid protocols combine the best features in both proactive and reactive routing protocols. These protocols are developed to increase scalability and reduce route discovery delay. The main idea is to use the proactive routing for maintaining routes to

neighbored nodes and determining routes to far away nodes using a route discovery packet. These protocols partition the network into a number of zones such as Zone

Routing Protocol (ZRP) [10].

1.4 Other Issues Related to Quality of Service (QoS) in MANETs

Quality of services in MANET is limited due to the lack of resources and continuous topology changes, which make QoS provisioning a very complicated process.

The QoS should be provided in all network layers such as application layer, transport layer, etc. There are other equally important issues that are briefly described below.

1.4.1 Unpredictable Link Properties

Wireless links are unpredictable and change their conditions with time. Signal quality fluctuates due to several factors such as fading, interference, and multipath cancellation. These properties will influence the bandwidth and delay measurements.

1.4.2 Node Mobility

Mobility of devices (nodes) changes the network topology frequently, which changes routes dynamically. Mobility affects the transmission range between two devices. When a device moves, it may cause link failure that increases packet loss rate and retransmission. Mobility influences many factors including channel access, routing, and applications.

1.4.3 Hidden and Exposed Terminal Problems

Media access control (MAC) uses a traditional Carrier Sense Multiple Access protocol (CSMA), which introduces the hidden and exposed terminal problems. The hidden terminal problem happens when two nodes (A and B) are hidden from each other when they are colliding at receiver node C. An exposed terminal problem will result from a scenario where node B and C attempt to transmit data to node A and D respectively.

Node B is exposed to the signal range of node C, which postpones its transmission.

Nodes B and C hear each other; therefore, they will not transmit.

1.4.4 Limited Battery Life

Battery life is one of the important issues in MANETs. Mobile devices use batteries that have a limited capacity of power to supply devices. If the power of the device is consumed, it will affect itself and the entire network. QoS should be power aware and power efficient.

1.4.5 Route Maintenance

Given that the topology in MANET is a dynamic, this changes the behavior of communication medium making the accurate maintenance of network state information very difficult. Therefore, the routing algorithms in MANET must deal with inaccurate information. Nodes in a MANET can enter and leave an environment continuously, which may cause broken path during the data transfer. Thus, the need of a route with

minimal delay and overhead emerges. end-to-end QoS requires a bandwidth reservation at intermediate nodes, which may become cumbersome due to dynamic topology.

1.5 Thesis Organization

The remaining parts of this thesis are organized as follows. Chapter 2 presents a survey about related work in the literature. Chapter 3 covers a general overview about mobile ad hoc networks, their characteristics, and their applications. Chapter 4 presents simulation environment, while result and analysis are presented in Chapter 5. Finally, in

Chapter 6, we conclude the thesis with a short discussion on possible future work.

CHAPTER 2

Survey of Pervious Works

This chapter covers a brief survey of significant mobility models and routing protocols. Mobility and routing protocols have been studied extensively during the past decade for various network applications ranging from small sensor networks to interplanetary and deep space communication systems.

2.1 Mobility Models

Mobility is an important factor in wireless networks. It represents the movement of mobile nodes (MNs) and how their speed and direction are changed over time.

Mobility models represent or predict user's or wireless device’s movements. These models are often used to simulate or emulate the actual movement of the devices in terms of geometry, speed, etc., in a geographic area. A significant body of literature [11, 12] has shown that the mobility directly affects communication performance in terms of throughput and delay. Mobility models can be simulated in two ways: using traces obtained through real experiments, or generating synthetic data using the statistical characteristics. Traces are real mobility patterns that exist in life. Synthetic is trying to realistically represent the movement of users in the absence of traces availability. There are many different ways to classify synthetic mobility models such as individual and group mobility models. Figure 2.1 illustrates a hierarchical classification of mobility

models. The individual mobility model represents the individual movement of mobile nodes (MNs). It emulates the behavior of the user in real life.

We can classify individual mobility into seven different models [11]:

1. Random Waypoint Mobility Model [11]

2. Random Walk Mobility Model [11]

3. Random Direction Mobility Model [11]

4. Gauss-Markov Mobility Model [11]

5. A Boundless Simulation Area Mobility Model [11]

6. City Section Mobility Model [11]

7. A probabilistic Version of the Random Walk Mobility Model [11]

Group mobility model represents the group of MNs movements. Each mobile node’s movement is independent from other nodes’ movements that are outside its group, but its movement is highly correlated with the movement of the nodes within its group.

We can classify group mobility into five group mobility models [11]:

1. Reference Point Group Mobility Model [12]

2. Exponential Correlated Random Mobility Model [11]

3. Column Mobility Model [12]

4. Nomadic Community Mobility Model [12]

5. Pursue Mobility Model [12]

Mobility

Models

Entity Group

Mobility Mobility Models Models

Random Random Random Reference Nomadic Column waypoint Walk Direction Point Mobility Mobility

Group Model Model

Bondless Gauss- City Exponential Pursue Simulatio Markov Section Correlated Mobility n area Random Model

Figure 2.1: Mobility Model Types

2.1.1 Random Waypoint Mobility Model

Random waypoint model has been used in many mobility studies to compare performance of ad hoc network routing protocols. It was the first model created by

Johnson and Maltz [12, 14] in which the nodes are randomly placed in the simulation area. As shown in Figure 2.2, each node randomly chooses a destination and a direction.

After waiting for a period of time (pause time), each node chooses the new destination in the simulation area. The speed has been chosen from a uniform distribution [Vmin, Vmax], 13

where Vmin represents the minimum speed and Vmax represents the maximum speed. A new speed and a new destination direction are chosen independently from the previous movement. Pause time and Vmax are two important parameters in the Random Waypoint model. They affect the behavior of mobile nodes. When Vmax is high and the pause time is long, this produces a more stable network than if Vmax is low and the pause time is small.

600

500

400

300

200

100

Figure 2.2: Random Waypoint Mobility Model

This model assumes that the average speed is maintained during the simulation time. In [14], it shows during the simulation time that the average speed is decreasing until Vmin = 0. The nodes become more stuck moving long distances at a low average

speed. A simple solution has been proposed in [13] that sets a positive minimum speed

(e.g. Vmin = 1 or more). This solution enables the nodes to reach a constant speed and stabilizes their mobility as well. The model in [15] includes a new parameter Pstat that refers to the probability that the node remains stationary over the simulation process. It selects the pause time between probabilities [Pmin, Pmax], which does not change with time.

In [13, 16, 17], the models have failed to reach a steady state. The authors in [18] tried to solve the problem existed in the previous approaches. The approach in [18] provides new distributions for speed, pause time, and location in the simulation area.

They derived stationary distributions for these parameters in rectangular area to begin a simulation in the steady state distribution.

2.1.2 Group Mobility Models

Group mobility models represent a group of nodes, which move together (see Fig.

2.3) [19]. They represent the random movement of a group of mobile nodes as well as the random movement of each individual mobile node within the group [11]. There are many examples in ad hoc networks, which represent the behavior of mobile nodes as a group that moves together. For example, in military battlefield communications, a group of soldiers are working together to capture an enemy or provide protection. Other examples include rescue missions and vehicular networks.

2.1.2.1 Reference Point Group Mobility Model [12]

In reference point group mobility model [12, 19], nodes move according to the logical center path of the group. The movements of the logical center of each group and the random motion of mobile node are based on random waypoint mobility model. This logical center is used to determine group movement by a group movement vector (GM).

The motion of group center identifies the shape of the movement for all mobile nodes inside the groups.

RM Vg GM RP(T+1 1 RP(T Vg

Figure 2.3: Group Mobility Model

Each mobile node in the group mobility randomly moves around the logical center (group leader) whose movements rely on the group movement. When the leader node moves from time t to time t+1, members’ locations are changed based on the group logical center. After the location of logical center changes, it is combined with random motion vector (RM) to represent the new movement of each mobile node (MN) around its logical center.

2.2 Routing Protocols

Routing protocol represents the way that nodes communicate with each other. It provides information about routes between nodes. An efficient route must have minimum overhead and efficient bandwidth utilization. There are two fundamental routing protocols widely used in networks and distributed systems: distance vector and link state routing [20]. In distance vector, every node keeps a route to every other node. They create a route table and keep it up to data. In link state, nodes periodically deploy link state cost to all nodes in the network.

These two routing protocols are not suitable in wireless network because of the limited bandwidth, and control overhead. So, there are many other protocols that have been proposed in the literature, which adapt to changes in network conditions including changes in topology, traffic, bandwidth, etc. These routing protocols can behave differently depending on the number of nodes, node mobility, and the type of traffic source. They can be classified based on the characteristics illustrated in Figure 2.4 [21, -

22]. 17

Routing Protocols

Proactive Hybrid Reactive

Bellman FSR ZRP AODV DSR Ford

STAR OLSR ZHLS DYMO LAR

Figure 2.4: Classification of Routing Protocols

2.2.1 Proactive Routing Protocols

Proactive protocols are table driven routing protocols. Each node stores route information about neighboring nodes. Nodes build a table of routing information and keep it up to date. The route information is kept in a number of different tables. Also, if the network changes its topology, nodes can update the route information. Proactive protocols provide routes immediately because they save all routing information when they start up. These protocols have a significant problem in mobile networks because of increased overhead. Proactive routing protocols operate differently in terms of the number of tables they use, and how routing information are stored and accessed. In this

thesis, we study these routing protocols in the context of mobility in general and group mobility in specific.

2.2.1.1 Bellman-Ford Protocol [2]

Bellman-Ford routing algorithm is based on the distance vector routing [2]. Each node maintains a routing table, which contains information about the estimated time or distance to reach the destination. This protocol has a disadvantage that it is not loop free.

Loops waste time and network bandwidth. In [2], a solution was proposed to avoid this problem, which is maintaining only loop free paths and find the shortest path. In [22],

Destination Sequenced Distance Vector (DSDV) is a development to Bellman-Ford routing algorithm that creates a loop free path. Nodes use a sequenced number to distinguish an old route from a new one. The routing information is updated in the routing table in two ways: full dump and incremental update. The full dump updates the table(s) periodically by sending the full routing table, while incremental update is event driven that just sends the updated information.

2.2.1.2 Fisheye State Routing Protocol (FSR) [3]

A Fisheye State Routing protocol [3] is a proactive routing protocol (table driven) that was proposed by Kleinrock and Stevens [3]. The Fisheye is an implicit hierarchical routing protocol. It is built on a Link State Routing protocol used in wired networks. This routing protocol has been used to reduce the routing information needed to represent networks. Fisheye can maintain accurate path quality information about immediate

neighbors that are the nearest local node. This accuracy decreases as the distance increases from this node. This means that nodes maintain routing information near other neighbors.

Fisheye is functionally similar to link state routing, which maintains a topology map at each node. The key difference is the way in which information is distributed. In link state routing, packets are flooded into the network whenever a node detects a topology change. In Fisheye, nodes periodically exchange information between other nodes (not flooding like link state in wired networks). Thus, Fisheye is suitable for large networks because it consumes less bandwidth and messages that are being exchanged than the link state routing and this makes nodes keep less information [23].

Figure 2.5 shows the way by which Fisheye works in a wireless network. It was used in circles to represent different Fisheye scopes. Each scope contains a set of nodes with different colors. They can reach each other by traversing a number of hops. Nodes with small scopes contain more information than those with large scopes that are far away from the center. The scope radius of the Fisheye is important to balance between routing accuracy and the control overhead. In [24], the authors present that FSR has a bad performance for packet delivery ratio when using different pause times for different scenarios.

3 8 5 1 9

9 6 4 1 Hop=1 7 0 12 13 19 Hop=2 11 18 21 Hop>2 15 22 14 23 36 16 20 17 29 20 25 27 24 26 34 28 30 32 31

Figure 2.5: Scope of Fisheye

2.2.1.3 Optimized Link State Routing Protocol (OLSR) [4]

Optimized link state routing protocol [4] is also a proactive routing protocol (table driven). It periodically exchanges information between nodes in the network to maintain the topology up to data. The protocol is based on a link state algorithm and it provides a hop-by-hop routing. Each node should choose a set of nodes from neighbors to become a relay node called “multipoint relays” (MPR).

Figure 2.6 shows how a node selects MPR nodes. The multipoint relay nodes

(MPR) are responsible to send control traffic information through the network. OLSR minimizes the amount of information exchanged between nodes because it reduces the 21

number of duplicated retransmissions. Each node chooses a set of MPRs as one hop neighboring nodes would be counted as two hops nodes far from the local node [25]. The control messages (called Hello messages) are used to guarantee a bidirectional link with other neighbor nodes. Nodes deploy the message called Topology control (TC) to identify multipoint relay selection. Nodes do not select MPR, which can perform operation packets (read, process), and they cannot retransmit information.

Retransmitting nodes or multipoint relays

Figure 2.6: Multiple Relays

This protocol does not require reliability when it transmits control traffic because it regularly exchanges control messages. Also, it uses a sequence number to ensure that receivers know the order delivery of the control message. The protocol updates

information when nodes are moving from one location to another. Therefore, it works well with node mobility in that it can acquire any change through the control message. It is scalable for large and dense network areas [4].

In [26], they re-examine the performance of three protocols in the presence of a self-similar (maintain bursty characteristics) traffic model. OLSR showed poor performance with self-similar traffic and high mobility. They displayed the highest delay packet delivery ratio and overhead. In [27] OLSR achieves higher end-to-end reliability and overhead than AODV and SBR (Statistic Based Routing) by increasing the impact of mobility. In [28], the authors show that OLSR has the best end-to-end delay and data delivery ratio for CBR (Constant Bit Rate) traffic even though the routing load is higher as opposed to what authors found in [27]. In [29], the authors show that OLSR has the lowest routing overhead in the network but it is not a suitable routing protocol in

Vehicular Ad Hoc Networks (VANET).

2.2.1.4 Source Tree Adaptive Routing Protocol (STAR) [30]

Source tree routing protocol [30] is based on a link state algorithm. Each node has a set of links containing the paths to the destination. It protects a source tree between all links. This protocol uses two ways to update routing information: least overhead routing approach (LORA) or optimum routing approach (ORA). The LORA tries to provide viable routes that may not be optimal, but it uses less amount of routing overhead. The ORA provides the optimal paths and it uses a conditional update rather than periodic updates used in other protocols. 23

Under LORA [30], a router running STAR sends updates to its neighbors when it loses all routes to one or more destination. In [19], the amount of control overhead has been decreased by keeping the path and this can be done as long as the path information is valid and maintained. Also, it decreases the update rate by using clustering and periodic updates. STAR is suited for large networks because it minimizes bandwidth consumption for updating routing information. However, this protocol may not perform well under high mobility. Therefore, it may need more memory or processing because nodes change their neighbors when they move.

The authors in [31] try to compare the performances of three routing protocols in terms of data delivered, control overhead, and average latency. They found that STAR performed well in a small network with low connection between nodes. But in a dense connection, STAR (even with lower latency and control overhead) does not change much when nodes change the number of data flows.

2.2.2 Reactive Routing Protocols

Reactive protocols are on demand routing protocols. These protocols create the route when nodes want to send data to a particular destination. They use the route discovery to find routes to destinations by flooding a route request through the network.

There are two types of reactive protocols: source routing and hop-by-hop routing. In source routing, each packet carries the whole information from a source to a destination.

In hop-by-hop routing, which is also called (point to point routing), each packet carries information about the destination and the next hop address. The advantage of these 24

protocols is that they reduce the overhead in table driven. They reduce bandwidth consumption, but they take high routing delay because paths are established when nodes demand them [22].

2.2.2.1 Ad Hoc On Demand Distance Vector Routing Protocol (AODV) [6]

Ad Hoc On Demand Distance Vector [6, 32] is a reactive routing protocol. It is based on Destination Sequence Distance Vector (DSDV) routing protocol. AODV uses two mechanisms for route discovery process and route maintenance process. A route discovery is used to find the route only when a node has a data packet to send to a specific destination node that it does not know it. A source node broadcasts a route request message (RREQ) to its neighbor nodes (see Fig. 2.7) [33].

S D

Figure 2.7: Source Node Discovery Process

Neighbor nodes rebroadcast a route request message (RREQ) to their neighbors and they continuously broadcast until they reach a destination. Sometimes, one of the neighbor nodes knows the path to a destination node.

Finally, after a RREQ reaches a destination or neighbor node that knows a fresh path to destination, it will reply by transmitting a message back to the source node called route reply message (RREP) by creating a reverse path (see Fig. 2.8). A fresh path is the intermediate nodes that have a sequence number equal or greater than the one in RREQ.

The intermediate nodes between source and destination set up a route to a destination node. The forward route is setup for a period of time then the intermediate nodes will delete it if a route is not used. RREP will be deleted after a time out of 3000 milliseconds

[33].

S RREP D

Figure 2.8: A Route Reply RREP Process

Path maintenance is used to maintain the route between source and destination node (see Fig. 2.9). If a source node moves, it can resend a new RREQ message to discover a new route to a destination. If any node along the path or destination is moved, its upstream neighbor produces a link failure message to each of its active neighbors to inform them. AODV includes a loop free by using two counter sequence numbers and broadcast id. The sequence numbers are used to determine the freshness of information about the route to a source node. The broadcast id is unique in that it is incremented for every RREQ.

S D Link Failure notification Data transmission

Figure 2.9: Route Maintenance

Nodes may use Hello message to determine local connectivity and detect link breaks. Hello message is a control message that creates or refreshes the routing table entry. It locally broadcasts control message with a specified interval. There are two variables that control the broadcasting Hello messages: Hello Interval and Maximum

Allow Hello Loss. Hello Interval is the maximum time interval between the transmissions of Hello message. Maximum Allow Hello Loss is the maximum number of periods of

Hello interval to wait without receiving a Hello message before detecting a loss of connectivity to a neighbor.

The advantage of Ad hoc on demand routing protocol [22] is the reduction of the number of broadcasts messages because it discovers a path on demand. It is adaptable to the changes of networks. Sometimes, the delay increases because the node may rediscover the route that incurred delays to establish a route.

Papers [13, 26, 28, 34] compare the performance of different routing protocols.

They obtained, from the simulation comparative result that AODV performs well at all mobility speeds than other protocols. In [35], the authors test the behavior of AODV and other protocols in large scale mobile networks. AODV suffers more in the term of packet delivery fraction (PDF) because it increases the route discovery process for large scale networks. But it still performs well in terms of end-to-end delay.

2.2.2.2 Dynamic Source Routing Protocol (DSR) [7]

Dynamic Source Routing protocol (DSR) [7, 22, 36, 37] is a reactive routing protocol based on the source routing concept. Each date packet includes a full address

(list of nodes), which the packet must pass from source to destination. The source builds a source route in the packet header. A node in the network maintains a route cache that stores source routes. When new routes are discovered, it updates entries in route cache.

When a source node has a packet to send to another node, it checks its route cache for a route to a destination. If the route cache does not contain any information, then the source node broadcasts a route request across the network. One advantage of DSR is that allows a source node to store more than one path toward a destination. It reduces the route discovery overhead by using route cache and it reduces the route maintenance overhead.

The disadvantage of DSR is that the packet header size grows as the number of intermediate nodes increases. There are two major mechanisms that work together in

DSR:

1. Route discovery: a source node will initiate a route discovery by broadcasting a route

request packet, which may be received by other nodes within its wireless transmission

range (see Fig. 2.10). The route request contains the destination address, a request id,

a route record field, and order intermediate node address. If the route discovery finds

the route to a destination, a route reply packet will be generated, which contains a

sequence of intermediate nodes. The route record field accumulates the sequence of

hops taken during route discovery.

<1,2> 5 2 <1,2,5> <1>

<1> Source 1 <1,3,> 7 Destination 3

<1,3> 6 <1> <1,4,6> 4

<1,4>

Figure 2.10: Build Record Route

Each route request packet contains a unique request id. It represents a counter, which is increased when a new route request is sent by the source. Each node should maintain a list of the initiator’s address and request id. There are four steps that are used to process the route request packet at any nodes receivers. a) If the initiator address and request id are found in the list of recent route requests, then the route request packet is discarded. b) If the host’s address is already existed in the route record, then the route request packet is discarded.

c) If the destination address of the request matches the host’s address, then the route record contains the route by which the request reaches host’s address starting from the source. It sends a route reply packet to the source node that contains a copy of this route. d) Otherwise, it attaches this host’s address to the route record in the route request packet and rebroadcasts the packet.

<1, 2, 5>

5 2 <1, 2, 5> <1, 2, 5>

Source 1 7 Destination 3

Figure 2.11: DSR Route Reply

A source node broadcasts route requests and continue to do so till the time it reaches a destination. A route reply is sent back to the source after attaching the list with all intermediate nodes either upon the point where the request packet reaches the target node or an intermediate node, which has a route to the destination (see Fig. 2.11). There are two ways to send a reply packet: by using a reverse of route record if it supports

symmetric link, or by initiating route discovery on the part of the destination if it is an asymmetric link. The route record indicates which sequence of hops was taken [33].

2. Route maintenance: Each node transmitting a packet is responsible for verifying that

the packet has been received by the next hop during travel to a destination (see Fig.

2.12). If the node decides on a fatal transmission error at its data link layer, a route

error packet is directed to the source. The route error packet contains the address of

the node detecting the error and the initiator address. Upon receiving a route error

packet by the node, it takes out the hop that contains the error from the route cache,

and all other routes containing this hop are condensed at that point. Also, the route

maintenance uses an acknowledgement packet to confirm the status of the route from

the source to the other node.

In [26, 34], the authors show that DSR performs very well in terms of end-to-end delay, throughput, and lowest control overhead. In large scale wireless networks [28, 35], the DSR scale is well in term of the packet delivery fraction, but suffers from increase of end-to-end delay because of its route discovery process.

A B C D E

Figure 2.12: DSR Route Maintenance

2.2.2.3 Dynamic MANET On Demand Routing Protocol (DYMO) [8]

Dynamic MANET On demand [8, 38, 39] routing protocol is a reactive protocol, which was developed by the Internet Engineering Task Force (IETF) group. It is used to find a unicast route between nodes that want to communicate with each other. DYMO is built based on AODV routing protocol that it is modified to use path accumulation. There are two operations of the DYMO: Route Discovery and Route Maintenance [8]. In route discovery process, the source node welling to transmit data to a specific destination broadcasts a route request packet (RREQ) into the network. The intermediate nodes record a route to a destination. RREQ contains the source address, destination address, sequence number, and hop limit. The source node may resend a RREQ again if it does not find any route to the destination.

After the route request packet reaches the destination node, it responds back to the source node by using a Route Reply (RREP) packet. The intermediate nodes add their own address to RREP packet when they receive RREP. After that, the route between the source and destination node is established in both directions when the source node receives the route reply packet. The path accumulation reduces the routing overhead because it reduces the number of RREQ packet in route discovery [40]. Fig. 2.13 shows how path accumulations work. Besides route information of a requested target, a node will also receive information about all intermediate nodes of a newly discovered path.

This is a major difference between DYMO and AODV, which only generates route table entries for the destination node and the next hop.

A A A

AODV A B C D

D D D

A A, B A, B, C

DYMO A B D C

D, C, B D, C D

Figure 2.13: Route Discovery in DYMO and AODV

The Route Maintenance is used to monitor the route from source to destination.

Each intermediate node maintains a route. It will inform the source node that the current route is not available if the next node along the route from source to destination is broken

[41]. It sends a Route Error (RERP) packet that includes a list of addresses and sequence numbers. The source node will start a route discovery process again if it has data to transmit to a destination.

In [41], the authors compare performance between three different routing protocols DYMO, AODV, and DSR. Each protocol utilizes a random waypoint mobility model in different pause times. The DYMO shows a good packet delivery ratio of all

pause times than other protocols because it uses path accumulation that reduces the number of route request packets. Paper [42] shows that DYMO has a higher throughput than other protocols, but it has the worst performance for average jitter.

2.2.2.4 Location Aided Routing Protocol (LAR) [9]

Location Aided Routing Protocol [9, 43] is a reactive routing protocol that is based on a flooding algorithm. LAR is an improvement over AODV and DSR in terms of route request packet flooding. The purpose of using LAR is to decrease the routing overhead by using location information. This protocol uses the Global Positioning

System (GPS) to obtain the location information about nodes. LAR knows the physical location of any node needed. GPS information has small error that cannot determine the exact node position. LAR uses the location information to flood a route request packet for destination in the forwarding zone instead of an entire network space.

In Expected Zone [43], a node sends a data packet to a particular node in an expected zone. Suppose node S knows that node D is at location L and the current time is t1. Then node S is able to determine the expected zone of D by using the location information. For instance, if node D traveled with an average speed v, then node S can expect node D in the circular region of radius v (t1 - t0), centered at location L (see Fig.

2.14). The expected zone is only estimated by node S to determine all the possible locations of D. If a node moves with higher speed than the average, then the destination node may be outside the expected zone at time t1 [44].

V(t1-t0) L

Figure 2.14: LAR Expected Zone

If node S does not know the information location about node D at time t0, then node S is not able to determine the expected zone. Therefore, the entire network region is chosen to be the expected zone. Thus, it is reduced to a simple flooding routing algorithm. In Request Zone, when a source node sends data packet again, an intermediate node will forward a route request packet only, if it belongs to the request zone. The request zone comprises the expected zone and other regions adjacent to the request zone.

There are two different types of LAR request zone: LAR Scheme 1 (LAR1) and LAR

Scheme 2 (LAR1).

LAR1 Request Zone [9] uses a rectangular shape (see Fig. 2.15). Assume that source node S recognizes the old location of destination node D at (Xd, Yd) at time t0. It also recognizes its average speed v, then the expected zone at time t1 is defined as a

circle with radius R = v (t1 – t0) centered at location (Xd, Yd). In LAR1 algorithm [9], the request zone is defined as the smallest rectangle that includes current source node and expected zone. The sides of the rectangle are parallel to the axes of X and Y. The source node defines the four corners of the rectangular request zone.

A (Xs, Yd+R) P (Xd, Yd+ B (Xd + R, Yd+ R) R)

R Q (Xd+R,Yd) R (Xd,Yd) (Xd, Yd) Expected Zone Expected Zone

J (Xj, Yj) I (Xi, Yi)

S C (Xd+R,Ys) (Xs,Ys)

Network Space Request Zone

Figure 2.15: LAR Scheme 1- Request Zone

The route request packet includes the four coordinates when initiating the route discovery process. The node has discarded the route request when a node is outside the rectangle (expected zone). If the destination node receives the route request packet, it replies back with a route reply packet (as in the flooding algorithm). It includes the

current location and the actual time in the route reply packet. The source node records and uses this information for a route discovery in the future.

(Xd, Yd)

DISTs

DISTn

DISTi

DISTk N

S (Xs, Ys) K

Figure 2.16: LAR Scheme 2

The LAR Scheme 2 (LAR2) [9] explicitly estimates the requested zone in its route request packet. Suppose source node S knows the location (Xd, Yd) of destination node D at time t0. The source node calculates its distance from location (Xd, Yd) (see

Fig. 2.16). It forwards the distance with the route request packet. The node can only forward the route request packet if the distance is closer or limited to the maximum

distance. The disadvantage of this protocol is that every node needs to carry special equipment (GPS) [22].

2.2.3 Hybrid Routing Protocols

Hybrid routing protocols [22] have both features of proactive and reactive routing protocols. These protocols are used to reduce the route discovery overhead by allowing nodes that are close to each other to work together forming some sort of a backbone. This is accomplished by features of proactive and reactive routing protocols that maintain routes for nearby nodes and find routes for far away nodes by using route discovery.

Hybrid protocols can divide the network based on zones, clusters, or trees. Most of these protocols are zone based, which partitions the network space based on the number of zones (regions) in each node. There are many different hybrid routing protocols proposed in wireless networks like Zone Routing Protocol (ZRP) [10], Zone

Hierarchical Link State (ZHLS) [41], Scalable Location Updates Routing Protocol

(SLURP) [41], Distributed Spanning Trees Based Routing Protocol (DST) [41], and

Distributed Dynamic Routing Protocol (DDR) [41].

2.2.3.1 Zone Routing Protocol (ZRP)

Zone Routing Protocol [10, 45] is a hybrid routing protocol, which includes the best of both proactive and reactive routing protocols. ZRP divides the network area into overlapping zones. It determines the zone of node by using proactive routing protocols that have route information to all neighbors. It uses the reactive routing protocols for

routing between multiple zones. The route zone of each node defines the minimum distance in hops from the source node to the zone radius. It has a radius, which is evaluated by the number of hops and not as a physical distance. Each node determines its own zone size. In Figure 2.17, the routing zone of S consists of the nodes A–K, but not L.

Each node of each zone is divided into two types: peripheral nodes and interior nodes. Peripheral nodes are nodes placed at the boundary of the zone (the zone radius).

The interior nodes are nodes located inside the zone radius expect boundary. ZRP has various routing protocols used to supply routing like Intrazone Routing Protocol (IAR)

[45], Interzone Routing Protocol (IERP) [45], and Bordercast Resolution Protocol (BRP)

[45]. The Intrazone Routing Protocol (IARP) is a proactive routing that is used within the zone to identify the route to peripheral nodes. It is limited to the radius of the node zone.

IARP needs to periodically update the route information inside the radius of the node’s routing zone because it is a table driven protocol. The Interzone Routing Protocol (IERP) is the reactive routing protocol that is exploited to connect between nodes of different zones. It is only initiated when it is needed to send data to nodes outside the zone (on demand). It reduces the amount of delay by using the Bordercast Resolution Protocol

(BRP). IERP takes the advantage of routing information that IARP provides. It does not submit the query to all local nodes, but only to its peripheral nodes that will reduce delays in global route discovery [49].

The Bordercast Resolution Protocol (BRP) is used to send a route request created by IERP directly to peripheral nodes. BRP uses the local map from IARP and generates a

broadcast tree of it. It uses a query control mechanism to direct route request away from areas of the network that have already been included by another query.

J C B

G D S

E A

H F

Figure 2.17: ZRP Zone

The radius of the zone plays a key role for performance of ZRP. If it uses a large radius, then the proactive routing protocol dominant. Also, if it is uses a small radius of one hop then the reactive routing increases, which increases route discovery and, that increases the delay. In [45], authors provide a flexible solution for performance of ZRP.

They introduce many query control schemes (like Query Detection (QD1/QD2), Early

Termination (ET), and Random Query Processing Delay (RQPD)) to overcome a redundant query. 41

CHAPTER 3

Mobile Ad Hoc Networks

Wireless networks can be divided into two types: infrastructure networks and ad hoc networks (infrastructure less) (see Fig. 3.1). A fixed infrastructure wireless network is a set of wireless devices, which are connected to fixed base stations (BS). In other words, a fixed infrastructure has a central access point (AP), which is responsible for all operations such as routing and security. The base stations are used to make a connection with devices and are responsible for routing between them. Cellular networks are an example of infrastructure wireless networks. A mobile device is needed to find the nearest base station to communicate with it (and called handoff or handover).

Ad hoc network is a network without any fixed infrastructure, which has two types: static and mobile ad hoc networks. A static ad hoc network has fixed nodes, which communicate with each other through predefined links. These nodes act like a router that can receive and transmit data without the need for any access point (AP).

Mobile ad hoc networks are a set of wireless network devices, which have temporary network communication with each other without relying on a fixed infrastructure or central administration (such as router or access point AP). MANETs are peer-to-peer or multi-hop mobile wireless networks that store and forward packets from a source to a destination. MANET has the capability for easy and rapid deployment

anywhere because there is no base station needed. MANET is a self-configuring and self- organizing mobile wireless network. These networks are mainly used in military applications, business, emergency services, and conferences. All devices in MANET are moving freely and randomly, which connect between each other dynamically. The devices act like a router that is able to discover and maintain routes to other devices in the network. One of the biggest challenges in the MANET is the routing protocol. Traditional wired routing protocols cannot work well for MANET that has no infrastructure and dynamic topology changes.

Figure 3.1: Wireless Network

There are several advantages of MANETs. It can be established and removed very fast without any previous infrastructure. Also, MANET can support connection failures

(fault tolerance) due to routing protocols, which are designed to manage these problems.

Mobility is another advantage, which MANET nodes can move randomly at the same time in different ways. MANET can have less cost due to the absence of its infrastructure. Finally, it allows reusing spectrum due to short communication links, and it reduces radio transmission.

MANETs have some problems such as bandwidth constraints because the capacity of the wireless connection is less than the wired networks. Also, the battery has a power constraint, which reduces the number of operations. Therefore, MANETs need efficient algorithms to keep battery power for the longest time. Error rate in MANETs is higher than wired networks because of the increased transmission error and interferences

[46]. Security is more complex to maintain in MANETs than in wired networks because of the lack of secure boundary and centralized management. This lack may cause various links attacks.

3.1 Characteristics of Mobile Ad hoc Networks (MANETs)

MANETs are characterized by:

1. Dynamic Network Topology

Devices move in any direction, thus, the network topology changes randomly and

unpredictably. Mobile nodes establish routing between each other dynamically as

they move inside network area.

2. Bandwidth and Capacity

Wireless links have lower capacity than infrastructure networks. Throughput is less

than the maximum transmission rate after effects of noise, fading, and interference.

3. Power

Devices in MANET have different types of batteries or other exhaustible means

for their energy. The energy conservation plays an important factor when system

designs are optimized.

4. Security

MANETs are more effective to physical security threats than wired networks.

There are many security techniques applied in MANETs. The decentralization of

MANETs allows more robustness.

3.2 Applications

MANETs are widely spread in many areas. MANETs can be easily used in any environment or anytime without communication infrastructure. Ad hoc networks have been developed for military applications without any stationary infrastructure or centralized management such as battlefields [47]. There are many applications as follows:

1. Defense application: Many defense applications have been used where

communication infrastructure is impossible to locate. It is used to maintain an

information network among soldiers on the ground, vehicles, or planes in the air.

2. Personal area network application: A personal area network is a short transmission

range between devices that is used for communication between these devices (e.g.

Bluetooth, IrDA).

3. Crisis management applications: When a natural disaster happens (e.g. earthquake,

fire, flood), it is very difficult to establish a wired connection between devices.

MANET provides quick communication setup in a few hours instead of few days,

which is required for wire-line communication.

4. Industrial applications: MANET is widely used in commercial applications (e.g.

manufacture). There are many electronic devices that are interconnected. The wiring

connection leads to the crowding of space. Ad hoc networks allow for easily moved

and reconfigured networks based on need.

5. Health care applications: MANET is very helpful in critical and emergency situations.

It allows exchange information (data, video, audio) between a patient and health care

center. For example, the video information may act as an aid to determine the level of

injuries.

It is worthy to mention that MANETs are widely used in recent applications.

Some of its features include: flexibility, lack of infrastructure, auto configuration, ease of deployment, and low cost among others. In addition, the use applications make it a fundamental part of the future. As a result, MANETs are implemented on military communication, airplane, and natural disaster. There are many challenges on MANETs to

be solved, including mobility, protocols, and services. In our work, we show the performance of routing protocols under different mobility with various traffic patterns.

CHAPTER 4

Simulation Model

In this chapter, we describe the simulator software, models, and the parameters used in our simulation. We used QualNet [1] simulator from Scalable Network

Technologies to create experiments and conduct performance analysis. The software is accurate, fast, tested, and scalable.

4.1 Simulator

QualNet [1] is a planning, testing, and training tool, which tries to mimic the behavior of a communication network. QualNet has layered modules for Physical, MAC,

Network, Transport, and Application layers. We have utilized all these modules to conduct an extensive set of experiments. In these experiments, we used various mobility models and routing algorithms to support various traffic models. We acquired our results from QualNet simulation that can create a custom scenario model to predict wireless, wired, and mixed platform networks. It allows users to evaluate the behavior of a network, and examine the network features. QualNet provides a comprehensive environment for designing protocols, creating and animating network scenarios, and analyzing their performance. It is composed of the following components:

1. Architecture: the architecture consists of a graphical scenario design and

visualization tool. There are two modes in the architecture component. First, design

mode for designing experiments. Second, visualize mode for running and visualizing

experiments.

2. Analyzer: A graphical statistic analyzing tool that displays results collected during

simulation time.

3. Packets Tracer: It is a graphical tool to display and analyze packets.

4. File Editor: A text editing tool.

5. Command Line Interface: Command line access to the simulator.

4.2 Features

There are several features of the simulator that enable creating a virtual network environment and those are:

1. Speed

It can support real-time speed to enable software-in-the-loop network emulation

and human-in-the-loop modeling. Faster speed allows the designer and developer

to run several analyses by varying traffic parameters, network, and model in a

short time.

2. Scalability

It can model thousands of nodes, which benefit from the latest hardware and

parallel computing techniques. It can run on cluster, multi-core, and multi-

processor systems to model large networks with high accuracy. 49

3. Model Fidelity

It uses extremely detailed standards based implementation of protocol model and

also provides advanced models for the wireless environment to enable more

accurate modeling of real world communication networks.

4. Portability

QualNet and its library of models can run on several platforms, such as Windows

and Linux operating systems’ distributed and cluster parallel architectures, and

both 32 and 64-bit computing platforms. QualNet allows users to develop a

portable model or design a network model on their own computers and then

transfer it to a powerful multi-processor Linux server to run capacity,

performance, and scalability analyses.

5. Extensibility

It can be connected to other hardware and software applications, such as real

networks, and third party visualization software to greatly enhance the value of

the network model.

4.3 Traffic Model

There are many types of traffic sources that can be generated as a stochastic model of traffic or data source. We can classify data traffic into four types:

4.3.1 Constant Bit Rate (CBR)

CBR generator generates a fixed data rate (deterministic rate) by transmitting packets of a fixed size at a fixed rate, which is used for measuring the data rate in the network. CBR is used to simulate applications between end systems, which require expected response time and fixed amount of bandwidth to be continuously available during the connection time. It is useful for streaming multimedia content including applications services such as video and voice services on a limited capacity channel because it uses maximum bit rate, not the average. Therefore, CBR is used to take advantage of all capacity. It is not the optimal choice for storage due to the fact that it does not allocate enough data for complex sections because it wastes data for simple sections.

To solve the problem of lack of enough data for complex sections, it can choose a high bit rate to guarantee that there will be enough for the whole encoding process [48]. It is difficult to achieve a perfect CBR that deals with other coding schemes such as

Huffman coding or run length encoding to produce variable length codes. This problem can be partly solved by changing the quality or completely solved by padding. When the stream video uses a CBR, the sender could be under the CBR rate. Therefore, it is necessary to add stuffing packets in the stream to complete the data rate required. These packets do not have any effect on the stream.

4.3.2 Variable Bit Rate (VBR)

VBR is opposed to constant bit rate (CBR) where VBR files change the amount of output data per time segment. VBR allows a higher bit rate that requires more storage space to be allocated to the more complex segments of media files, while less space is allocated to less complex segments. VBR uses an average bit rate, which calculates the average of these rates. This feature of VBR produces a better space management compared to a CBR file of the same data. It allows more flexibility to use bits available to encode the sound or video data more precisely. It uses fewer bits in small encode demand and more bits in high encode demand. There are several disadvantages that are shown on VBR, which may take additional time to encode data. Therefore, the process becomes more complex. VBR may show problems during streaming when the bit rate exceeds the data rate of the communication path. We can avoid this problem by limiting the bit rate during encoding through increasing the playout buffer.

4.3.3 Random Traffic

Random traffic is a stochastic model of the traffic flows (a random distribution based traffic generator) such as a cellular network and computer network. These random distributions are applicable to both session property and traffic property. A packet generation model is a traffic of packet flows such as web traffic, and the data of which can be sent and received by a user’s web browser. It can generate different traffic models:

Exponential, Pareto, and Uniform. Exponential is an ON/OFF mode that the holding time follows in an exponential distribution. During the ON period, packets are generated at a 52

burst rate while the OFF period does not generate any traffic. Pareto is also ON/OFF traffic with burst times follows Pareto distribution. These models are used to analyze the performance of different protocols, algorithms, and network topologies.

4.3.4 File Transfer Protocol (FTP)

FTP is a standard network protocol used to transfer files from one device (client) to another (server) over TCP network. FTP uses a separate channel for control and data connections between client and server. It also uses an authentication system (username and password) to ensure that only authorized users are allowed to access a server, but, sometimes, anonymous users can connect to the server if it is set up to provide files to any user requesting them. FTP uses encryption content for secure transmission that keeps the username and password secure such as Secure Sockets Layer (SSL)/ Transport Layer

Security (TLS), and Secure Shell (SSH) File Transfer Protocol (SFTP). When the connection is established and authentication is complete, there are two basic commands used to send or received files. The main goal of FTP is to make file transfer simple and easy. FTP can be used with other applications to move files from one place to another.

4.4 Simulation Setup

We experimented our routing protocols with different mobility models and various traffic classes in MANET. We created simulations to investigate performance and evaluation of routing protocols. The simulation parameters are included in Table 3.1.

There are several parameters effects on simulation such as mobility, network shape, node placement, and other factors.

Dimension 1500×1500m Number of Nodes 50 Mobility Model Random Waypoint, Group Mobility Simulation Time 1000 sec Minimum Speed 0 m/s Maximum Speed 10 m/s Pause Time 5 sec Traffic Model CBR, VBR, Random Node Placement Strategy Random Routing Protocol FSR, OLSR, AODV, DYMO, ZRP Item Size 512 bytes MAC Protocol 802.11 Data Rate 2 Mbps Antenna Model Omni Directional

Table 3.1 Simulation Parameters

4.4.1 Simulation Time

We ran the simulation for 1000 seconds to avoid generating statistical anomalies.

Application generates traffic during this time, which starts at 1 second.

4.4.2 Mobility

Nodes move randomly in network space due to dynamic environments. There are several types of mobility models used in different studies. For example, random waypoint model is used more widely than other models. In this model, nodes move and select destination in a random way. After that, nodes pause for a period of time (nodes stop for some time after they reach the destination), then select the new destination randomly and move. A new speed and a new destination direction are chosen independently from the previous movement. Pause time is an important parameter in the Random Waypoint model that affect the behavior of mobile nodes. Random waypoint and group mobility models are used in our simulation. The speed has been selected from a uniform distribution between [Vmin, Vmax]. We selected the minimum speed 0 m/s and the maximum speed 10 m/s. Nodes paused for 5 seconds, and then moved to a new destination.

4.4.3 Node placement and Network shape

Shape of space in the network is affected when the nodes moving in space, which will influence the simulation results. Different shapes create different mobility pattern of nodes. Different shapes of networks affect the work of routing protocols. In our simulation, the network consists of 50 mobile nodes randomly placed in a 1500×1500m square space. When nodes are reached the simulation boundary, they bounce back to the simulation area. There is no obstacle inside the simulation area.

4.4.4 Physical Layer

The Physical layer is a simple radio model, which can support either Signal-to-

Noise Ratio (SNR) or Bit Error Rate (BER). In our experiments, each scenario uses

802.11b radio on physical layer with Omni direction antenna model. The Omni antenna is a basic antenna, which yields the same antenna gain irrespective of the direction of the transmitted or received signal. We set the noise factor to 10. Also, we set the radio data rate to 2 Mbps.

4.4.5 MAC Layer

There are several MAC types such as 802.11, 802.11e, 802.11s, TDMA, and

CSMA. In our simulations, we selected 802.11 for all scenarios.

4.4.6 Network Layer

In this layer, we selected IPv4 network protocols. This protocol supports mobility, which allows transparent routing of IP datagrams to mobile nodes during moving from one domain to another. The IPv4 uses 2048 bytes IP fragmentation unit to keep the whole packets.

4.4.7 Routing Protocols

In our simulation, we selected five types of routing protocols (FSR, OLSR,

AODV, DYMO, and ZRP), which are more popular in wireless networks. These protocols have different ability to react to dynamic topology changes such as node movement, links breaks, and different traffics.

4.5 Metrics

We compare the performance of Proactive (FSR, OLSR) Routing protocols,

Reactive (ADOV, DYMO) Routing protocols, and Hybrid (ZRP) Routing protocol. We have used average end-to-end delay, throughput, and Jitter to analyze and compare performances of these protocols.

4.5.1 Throughput

Throughput is a ratio of total successful data, which reaches the receiver from the sender in the time needed to receive it. Throughput is represented in bits per second

(bits/sec) or in packet per second. We analyze the throughput result of routing protocols in terms of (bits/sec). MANETs need a high throughput due to unreliable connection, limited bandwidth, dynamic topology, and limited battery. We can calculate throughput from the equation below.

Throughput (bits / sec) = total delivered data / total simulation time

4.5.2 End-to-End Delay

End-to-End Delay is the average time that a data packet needs to reach a destination. This is the time when the source starts transmitting the first packet to its receiver. It calculates all delays caused by transmission time, queuing, MAC control, and transfer time. Transmission delay is the time taken to transmit all the packets on the link while propagation delay is the time taken to transmit first bit to reach the destination.

Processing delay is the time taken by all operation between source and destination. It also

needs to calculate a delay for route discovery when using reactive routing protocols.

Applications demand is different; therefore, they need different delay levels. Most applications in network (such as video) require a low average delay. In a MANET, the delay is higher than wireless network because of the limited signal power, mobility, established routes, and failed connections. end-to-end delay is used to measure the impact of different mobility models and various traffics on different routing protocols. We can calculate end-to-end delay from the equation below.

End-to-End Delay = transmission delay + propagation delay + process delay

4.5.3 Average Jitter

Jitter is the variation in the time between different data packets arriving. The source node sends packets in a continuous flow, but the delay between packets arriving varies due to route changes, queuing, and congestion. Jitter causes serious problems at the receiving end for audio and video applications. Applications require small jitter for better performance.

CHAPTER 5

Simulation Result

In this chapter, we presented results for simulation models that show the effect of factors (mobility and traffic) together on routing algorithms and measure application performance in terms of end-to-end delay, throughput (bit rate), and jitter. We ran simulations for a 1000 second with different traffic loads. We acquired the results from separate files by QualNet then put the data into spreadsheets and created charts.

5.1 Throughput

Throughput will increase when the number of nodes increases. Thus, it will increase performance. Also, it will increase when the speed increases, but it depends on the type of routing and what is the optimal maximum speed you should select in random waypoint mobility model. The results of throughput for routing protocols are shown in

Fig 5.1. In CBR traffic, AODV protocol outperforms other routing protocols with all loads because it can respond to dynamic change in the network topology and can maintain routes better than others. OLSR routing protocol shows high throughput performance among proactive routing protocols. DYMO performs well under 50 percent loads. But when the load increases, DYMO shows a few fluctuations due to the mobility.

FSR shows lower throughputs than OLSR due to the characteristic of proactive routing

protocols. ZRP presents the worst throughput among the protocols with all loads. It cannot react fast because it needs more time to find a route.

Figure 5.1: Throughput in Random Waypoint Model (higher is better)

In random traffic (Fig. 5.1), like CBR traffic, AODV outperforms other routing protocols. DYMO and ZRP show the worst throughput among the routing protocols when the load is higher than 50 percent. OLSR shows better throughput with the load higher than 50 percent. FSR shows the worst throughput when the load is under 40 percent.

In VBR traffic (Fig. 5.1), OLSR routing protocol shows better throughput than other protocols when the load is higher than 80 percent. Also like CBR traffic, AODV outperforms other protocols while FSR shows less throughput among other routing protocols. We can see that ZRP performs almost the same as the DYMO after the load exceeds 50 percent.

Under group mobility model with CBR and VBR traffic, our results are shown in

Figure 5.2. AODV outperforms other protocols in different traffic. OLSR shows the worst throughput among other protocols. ZRP performs better than OLSR when the load is higher than 50 percent, whereas DYMO shows good throughput when the load is under

50 percent. FSR shows better throughput when the load is higher than 80 percent.

In random traffic, DYMO shows the worst throughput when the loads are higher than 50 percent, while ZRP shows the worst throughput when the loads are under 50 percent. OLSR shows better throughput than proactive protocols.

Figure 5.2: Throughput in Group Mobility Model (higher is better)

5.2 Average End-to-End Delay

According to our simulation results in random waypoint mobility model using

CBR traffic (see Fig. 5.3), delay for AODV is less than the average delay of other protocols because the route discovery process is very fast. OLSR performs almost the same AODV routing protocol. It does not need to do route discovery due to the fact that its characteristic is table driven. FSR shows less average delay when the load is 30 and 40 percent. DYMO shows the worst delay among all routing protocols, while ZRP shows higher average delay when the load is less than 50 percent.

In random traffic, AODV also outperforms other protocols with different loads, followed by OLSR, FSR, and DYMO. ZRP shows the worst average delay due to mobility influence on the ability to find routes fast. In VBR traffic, AODV also performs very well compare to other routing protocols. DYMO shows higher delay when the load exceeds 50 percent.

In group mobility models, according to our simulation, results are shown in

Figure 5.4. In CBR and VBR traffic, ZRP routing protocol outperforms other protocols

(less end-to-end delay). DYMO and AODV protocol show the worst average delay. In random traffic, OLSR outperforms other routing protocols, while DYMO shows the worst average delay.

Figure 5.3: End-to-End Delay in Random Waypoint Model (lower is better)

Figure 5.4: End-to-End Delay in Group Mobility Model (lower is better)

5.3 Average Jitter

According to our results in random waypoint mobility model (see Fig. 5.5),

AODV presents the best performance, while ZRP presents the worst performance in terms of average jitter using CBR, Random, and VBR traffic. In random and VBR traffic source, OLSR, FSR, and DYMO protocols show almost the same average jitter when the load exceeds 50 percent while in CBR traffic, DYMO shows high average jitter and follows ZRP.

Figure 5.5: Jitter in Random Waypoint Model (lower is better)

In group mobility model (see Fig. 5.6) with CBR traffic, ZRP has the best performance, while DYMO and AODV have the worst performance when compared to other protocols. In random and VBR traffic, AODV shows lower jitter than other protocols. Like CBR traffic, DYMO shows the worst performance in random traffic, while OLSR has the highest jitter in VBR traffic.

Figure 5.6: Jitter in Group Mobility Model (lower is better)

The routing protocol, which has higher throughput, will give best performance.

Our results show that AODV has a higher throughput than other routing protocols. Table

5.1 shows the comparison of routing protocols under random waypoint and group mobility models with CBR, VBR, and Random traffics.

Mobility CBR VBR Random

model Throughput Delay Jitter Throughput Delay Jitter Throughput Delay Jitter

Random

Waypoint AODV AODV AODV AODV AODV/FSR AODV AODV AODV AODV

Group mobility AODV ZRP ZRP AODV ZRP AODV AODV OLSR AODV

Table 5.1: Comparison Results

5.4 Modifying AODV Parameters

There is a set of parameters in the AODV routing protocol that affects the behavior of the network in terms Quality of Services (QoS). This leads us to think about using Hello message (control message) that is used to detect and monitor links between nodes (connectivity). The results show that end-to-end Delay in AODV decreases and throughput increases under a random waypoint mobility model with CBR, VBR, and random traffics. In Figure 5.7, we can see that ratio throughput and delay in AODV with

Hello message is better than AODV without it.

Figure 5.7: AODV Comparison (higher is better)

CHAPTER 6

Conclusion & Future Work

6.1 Conclusion

Mobile ad hoc networks are a dynamic and unpredictable network topology.

There are several routing protocols that have been proposed for MANETs. Most pervious research focused on improving the existing routing protocols or designing new routing algorithms. In our work, we investigated and compared the impact of mobility models on routing protocols for various traffic classes in MANETs. There were many factors that affected the performance of routing protocols such as mobility and traffic patterns. We designed several simulation models that brought these factors together and measured the application performance in terms of end-to-end throughput (bit rate), latency, and jitter.

Three classes of MANET routing algorithms (Proactive, Reactive, and Hybrid), two mobility models (Random Waypoint and Group), and Three classes of traffic patterns

(constant bit rate, variable bit rate, and random) have been used.

We evaluated and compared the performance of five routing protocols (FSR,

AODV, OLSR, DYMO, and ZRP) in MANETs. As a result of our simulation, it indicated that AODV has the best performance with different mobility models for various traffic patterns. It showed high throughput, less end-to-end delay, and less jitter. We found that mobility is a key factor, which affects the performance of routing protocols. AODV routing protocol performed better at all mobility speeds among others protocols.

6.2 Future Work

MANETs have received more attention in recent years. We studied the impact of mobility on routing protocols with various traffic patterns. For future work, we have several goals. The first goal is to improve the performance of AODV routing protocol.

While the second goal would be planning to test AODV routing protocol with other mobility models. Finally, we would like to suggest the development of a new real life simulation that will allow analyzing the performance of routing protocols under realistic mobility and traffic.

Glossary

MANETs Mobile Ad Hoc Networks 1

QoS Quality of Services 1, 6, 8

MWNs Multi-hop Wireless Networks 2

FSR Fisheye State Routing protocol 7, 19, 56

OLSR Optimized Link State Routing protocol 7, 21, 22, 56

DSDV Destination Sequenced Distance Vector 7, 19, 26

AODV Ad hoc On Demand Distance Vector routing protocol 7, 25, 33, 56

DSR Dynamic Source Routing protocol 7, 29, 33

DYMO Dynamic MANET On Demand routing protocol 7, 33, 56

LAR Location Aided Routing protocol 7, 35

ZRP Zone Routing Protocol 8, 39, 56

MAC Media Access Control 9, 56

CSMA Carrier Sense Multiple Access protocol 9

MNs Mobile Nodes 11, 12

MPR Multipoint Relays 21

TC Topology Control 22

SBR Statistic Based Routing 23

VANETs Vehicular Ad Hoc Networks 23

STAR Source Tree Adaptive Routing protocol 23 72

LORA Least Overhead Routing Approach 23

ORA Optimum Routing Approach 23

RREQ Route Request message 25, 26, 33

RREP Route Reply message 25, 33

IETF Internet Engineering Task Force 33

GPS Global Positioning System 35, 39

ZHLS Zone Hierarchical Link State 39

SLURP Scalable Location Updates Routing Protocol 39

DST Distributed Spanning Trees Based Routing Protocol 39

DDR Distributed Dynamic Routing protocol 39

IAR Intrazone Routing Protocol 40

IERP Interzone Routing Protocol 40

BRP Bordercast Resolution Protocol 40

QD1/QD2 Query Detection 41

ET Early Termination 41

RQPD Random Query Processing Delay 41

BS Base Stations 42

AP Access Point 42

CBR Constant Bit Rate 23, 51

VBR Variable Bit Rate 52

FTP File Transfer Protocol 53

SSL Secure Sockets Layer 53

TLS Transport Layer Security 53

SFTP Secure Shell (SSH) File Transfer Protocol 53

SNR Signal-to-Noise Ratio 56

BER Bit Error Rate 56

TDMA Time Division Multiple Access 56

REFERENCES

[1] http://web.scalable-networks.com/content/qualnet.

[2] Cheng, Chunhsiang, Ralph Riley, Srikanta PR Kumar, and Jose J. Garcia-Luna-

Aceves. "A loop-free extended Bellman-Ford routing protocol without bouncing

effect." In ACM SIGCOMM Computer Communication Review, vol. 19, no. 4, pp.

224-236. ACM, 1989.

[3] Guangyu Pei, M. Gerla and Tsu-Wei Chen, "Fisheye state routing: a routing scheme

for ad hoc wireless networks," Communications, 2000. ICC 2000. 2000 IEEE

International Conference on, New Orleans, LA, 2000, pp. 70-74 vol.1.

[4] P. Jacquet, P. Muhlethaler, T. Clausen, A. Laouiti, A. Qayyum and L. Viennot,

"Optimized link state routing protocol for ad hoc networks," Multi Topic

Conference, 2001. IEEE INMIC 2001. Technology for the 21st Century.

Proceedings. IEEE International, 2001, pp. 62-68.

[5] Perkins, Charles E., and Pravin Bhagwat. "Highly dynamic destination-sequenced

distance-vector routing (DSDV) for mobile computers." In ACM SIGCOMM

computer communication review, vol. 24, no. 4, pp. 234-244. ACM, 1994.

[6] C. E. Perkins and E. M. Royer, "Ad-hoc on-demand distance vector routing," Mobile

Computing Systems and Applications, 1999. Proceedings. WMCSA '99. Second

IEEE Workshop on, New Orleans, LA, 1999, pp. 90-100.

[7] Johnson, David B., and David A. Maltz. "Dynamic source routing in ad hoc wireless

networks." In Mobile computing, pp. 153-181. Springer US, 1996.

[8] Chakeres, Ian, and Charles Perkins. "Dynamic MANET on-demand (DYMO)

routing." draft-ietf-manet-dymo-9.txt, May (2007). [Online] Available:

http://tools.ietf.org/wg/manet/draft-ietf-manet-dymo/draft-ietf-manet-dymo-09.txt

[9] Y.-B. KO, N.H. Vaidya, Location-aided routing (LAR) in mobile ad hoc networks, in:

Proceedings of the Fourth Annual ACM/IEEE International Conference on

Mobile Computing and Networking (Mobicom 98), Dallas, TX, 1998.

[10] Z.J. Hass, R. Pearlman, Zone routing protocol for ad-hoc networks, Internet Draft,

draft-ietf-manet-zrp-02.txt, work in progress, 1999.

[11] Camp, Tracy, Jeff Boleng, and Vanessa Davies. "A survey of mobility models for ad

hoc network research." Wireless communications and mobile computing 2, no. 5

(2002): 483-502.

[12] Bai, Fan, and Ahmed Helmy. "A survey of mobility models." Wireless Adhoc

Networks. University of Southern California, USA 206 (2004).

[13] J. Yoon, M. Liu and B. Noble, "Random waypoint considered harmful," INFOCOM

2003. Twenty-Second Annual Joint Conference of the IEEE Computer and

Communications. IEEE Societies, San Francisco, CA, 2003, pp. 1312-1321 vol.2.

[14] Broch, Josh, David A. Maltz, David B. Johnson, Yih-Chun Hu, and Jorjeta Jetcheva.

"A performance comparison of multi-hop wireless ad hoc network routing

protocols." In Proceedings of the 4th annual ACM/IEEE international conference

on Mobile computing and networking, pp. 85-97. ACM, 1998.

[15] Resta, Giovanni, and Paolo Santi. "An analysis of the node spatial distribution of the

random waypoint mobility model for ad hoc networks." InProceedings of the

second ACM international workshop on Principles of mobile computing, pp. 44-

50. ACM, 2002.

[16] Jardosh, Amit, Elizabeth M. Belding-Royer, Kevin C. Almeroth, and Subhash Suri.

"Towards realistic mobility models for mobile ad hoc networks." InProceedings

of the 9th annual international conference on Mobile computing and networking,

pp. 217-229. ACM, 2003.

[17] G. Lin, Noubir G and Rajmohan Rajaraman, "Mobility models for ad hoc network

simulation," INFOCOM 2004. Twenty-third AnnualJoint Conference of the IEEE

Computer and Communications Societies, 2004, pp. 463.

[18] W. Navidi and T. Camp, "Stationary distributions for the random waypoint mobility

model," in IEEE Transactions on Mobile Computing, vol. 3, no. 1, pp. 99-108,

Jan-Feb 2004.

[19] Hong, Xiaoyan, Mario Gerla, Guangyu Pei, and Ching-Chuan Chiang. "A group

mobility model for ad hoc wireless networks." In Proceedings of the 2nd ACM

international workshop on Modeling, analysis and simulation of wireless and

mobile systems, pp. 53-60. ACM, 1999.

[20] Rover, Elizabeth M., and Chai-Keong Toh. "A review of current routing protocols

for ad hoc networks." IEEE Communications 6 (1999): 46-55.

[21] Liu, Changling, and Jörg Kaiser. A survey of mobile ad hoc network routing

protocols. Universität Ulm, Fakultät für Informatik., 2003.

[22] Abolhasan, Mehran, Tadeusz Wysocki, and Eryk Dutkiewicz. "A review of routing

protocols for mobile ad hoc networks." Ad hoc networks 2, no. 1 (2004): 1-22.

[23] A. Iwata, Ching-Chuan Chiang, Guangyu Pei, M. Gerla and Tsu-Wei Chen,

"Scalable routing strategies for ad hoc wireless networks," in IEEE Journal on

Selected Areas in Communications, vol. 17, no. 8, pp. 1369-1379, Aug 1999.

[24] Maurya, Ashish K., Dinesh Singh, and Ajeet Kumar. "Performance Comparison of

DSR, OLSR and FSR Routing Protocols in MANET Using Random Waypoint

Mobility Model." International Journal of Information and Electronics

Engineering 3, no. 5 (2013): 440.

[25] Clausen, Thomas, Philippe Jacquet, Cédric Adjih, Anis Laouiti, Pascale Minet, Paul

Muhlethaler, Amir Qayyum, and Laurent Viennot. "Optimized link state routing

protocol (OLSR)." (2003).

[26] A. Al-Maashri and M. Ould-Khaoua, "Performance Analysis of MANET Routing

Protocols in the Presence of Self-Similar Traffic," Local Computer Networks,

Proceedings 2006 31st IEEE Conference on, Tampa, FL, 2006, pp. 801-807.

[27] Klein, Alexander. "Performance comparison and evaluation of AODV, OLSR, and

SBR in mobile ad-hoc networks." In Wireless Pervasive Computing, 2008.

ISWPC 2008. 3rd International Symposium on, pp. 571-575. IEEE, 2008.

[28] Mbarushimana, Consolee, and Alireza Shahrabi. "Comparative study of reactive and

proactive routing protocols performance in mobile ad hoc networks." In Advanced

Information Networking and Applications Workshops, 2007, AINAW'07. 21st

International Conference on, vol. 2, pp. 679-684. IEEE, 2007.

[29] Ledy, J., Hilt, B., Boeglen, H., Poussard, A.M., Drouhin, F., Vauzelle, R. and

Lorenz, P., 2015. Impact of Realistic Simulation on the Evaluation of Mobile Ad

Hoc Routing Protocols. Emerging Topics in Computing, IEEE Transactions

on, 3(3), pp.317-334.

[30] J. J. Garcia-Luna-Aceves and M. Spohn, "Source-tree routing in wireless

networks," Network Protocols, 1999. (ICNP '99) Proceedings. Seventh

International Conference on, 1999, pp. 273-282.

[31] Hong Jiang and J. J. Garcia-Luna-Aceves, "Performance comparison of three routing

protocols for ad hoc networks," Computer Communications and Networks, 2001.

Proceedings. Tenth International Conference on, Scottsdale, AZ, 2001, pp. 547-

554.

[32] H. Ehsan and Z. A. Uzmi, "Performance comparison of ad hoc wireless network

routing protocols," Multitopic Conference, 2004. Proceedings of INMIC 2004. 8th

International, 2004, pp. 457-465.

[33] Jorg, David Oliver. "Performance comparison of MANET routing protocols in

different network sizes." Computer Networks & Distributed Systems (2003).

[34] Sesay, Samba, Zongkai Yang, Biao Qi, and Jianhua He. "Simulation comparison of

four wireless ad hoc routing protocols." Information technology journal 3, no. 3

(2004): 219-226.

[35] Broustis, Ioannis, Gentian Jakllari, Thomas Repantis, and Mart Molle. "A

Comprehensive Comparison of Routing Protocols for Large-Scale Wireless

MANETs." In Sensor and Ad Hoc Communications and Networks, 2006.

SECON'06. 2006 3rd Annual IEEE Communications Society on, vol. 3, pp. 951-

956. IEEE, 2006.

[36] David B. Johnson, David A. Maltz, and Josh Broch,“DSR: The Dynamic Source

Routing Protocol for Multi-Hop Wireless Ad Hoc Networks,” in Ad Hoc

Networking, edited by Charles E. Perkins, Chapter 5,pp. 139-172, Addison-

Wesley, 2001.

[37] T. P. Lambrou and C. G. Panayiotou, "A Survey on Routing Techniques Supporting

Mobility in Sensor Networks," Mobile Ad-hoc and Sensor Networks, 2009. MSN

'09. 5th International Conference on, Fujian, 2009, pp. 78-85.

[38] F. Maan and N. Mazhar, "MANET routing protocols vs mobility models: A

performance evaluation," Ubiquitous and Future Networks (ICUFN), 2011 Third

International Conference on, Dalian, 2011, pp. 179-184.

[39] S. Edenhofer and P. Höfner, "Towards a rigorous analysis of AODVv2

(DYMO)," Network Protocols (ICNP), 2012 20th IEEE International Conference

on, Austin, TX, 2012, pp. 1-6.

[40] C. Sommer and F. Dressler, "The DYMO Routing Protocol in VANET

Scenarios," Vehicular Technology Conference, 2007. VTC-2007 Fall. 2007 IEEE

66th, Baltimore, MD, 2007, pp. 16-20.

[41] Maurya, Ajay Kumar, Dinesh Singh, Ajit Kumar, and Ritesh Maurya. "Random

waypoint mobility model based performance estimation of On-Demand routing

protocols in MANET for CBR applications." In Computing for Sustainable

Global Development (INDIACom), 2014 International Conference on, pp. 835-

839. IEEE, 2014.

[42] Singh, Dinesh, Ashish K. Maurya, and Anil K. Sarje. "Comparative performance

analysis of LANMAR, LAR1, DYMO and ZRP routing protocols in MANET

using Random Waypoint Mobility Model." In Electronics Computer Technology

(ICECT), 2011 3rd International Conference on, vol. 6, pp. 62-66. IEEE, 2011.

[43] Agrawal, Dharma, and Qing-An Zeng. Introduction to wireless and mobile systems.

Cengage Learning, 2010.

[44] Sarkar, Subir Kumar, T. G. Basavaraju, and C. Puttamadappa. Ad hoc mobile

wireless networks: principles, protocols and applications. CRC Press, 2007.

[45] Haas, Zygmunt J., and Marc R. Pearlman. "The performance of query control

schemes for the zone routing protocol." IEEE/ACM Transactions on Networking

(TON) 9, no. 4 (2001): 427-438.

[46] Karimi, Masoumeh, and Deng Pan. "Challenges for Quality of Service (QoS) in

Mobile Ad-Hoc Networks." In IEEE Wireless and Microwave Technology

Conference (WAMICON’09). 2009.

[47] Karimi, Masoumeh. Quality of Service (QoS) Provisioning in Mobile Ad-Hoc

Networks (MANETs). INTECH Open Access Publisher, 2011.

[48] AL-Maksousy, Hassan Hadi Latheeth. "Performance Analysis of Multi-hop Wireless

Networks." Electronic Thesis or Dissertation. Kent State University, 2012.

[49] C. A. Santivanez, B. McDonald, I. Stavrakakis and R. Ramanathan, "On the

scalability of ad hoc routing protocols," INFOCOM 2002. Twenty-First Annual

Joint Conference of the IEEE Computer and Communications Societies.

Proceedings. IEEE, 2002, pp. 1688-1697 vol.3.