Chapter 1: Introduction ----------------------------------------------------------------------------------------------------------------------------------------------------------

Wireless networking in recent years is becoming an integral part of residential, commercial, and military computing applications. The elimination of unsightly and cumbersome wiring and the increase of mobility made possible by wireless networks are only some of the advantages that have led to the widespread acceptance and popularity of such networks.

Recent developments in ad-hoc wireless networking have eliminated the requirement of fixed infrastructure (central base station required in cellular networking) for communication between users in a network and expanded the horizon of wireless networking. These networks termed as mobile ad hoc networks (MANET) are a collection of autonomous terminals that communicate with each other by forming a multihop radio network and maintaining the connectivity in a decentralized manner.MANET and, in particular, wireless sensor networks (WSNs) are finding increasing applications in communication between soldiers in a battlefield, emergency-relief-personnel coordinating efforts, earthquake aftermath, natural disaster relief etc.

Minimal configuration and quick deployment make ad hoc networks suitable for emergency situations like natural disasters or military conflicts. The presence of a dynamic and adaptive routing protocol will enable ad hoc networks to be formed quickly.

1.1History

The concept of ad hoc network was first introduced by Packet Radio Network (PRNET) research for military purpose in 1970, which evolved into Survivable Adaptive Radio Networks (SURAN) program in the early 1980.

In the early 1990s, ad hoc networks entered a new phase of development due to the popularity of notebook computers with communication equipments based on Radio Frequency and infrared. The idea of an infrastructure less collection of mobile hosts was proposed, and the IEEE 802.11 subcommittee adopted the term “ad hoc networks”. Non-military applications were also suggested.

Since mod 1990s, a lot of work has been done on the ad hoc standards and the Mobile Ad Hoc Networking (MANET) working group was born, and made effort to standardize routing protocols for ad hoc networks.

1.2 Basic Concepts of Ad Hoc Network

Ad hoc networks are defined by the manner in which different nodes organize themselves to provide the path for the data to be routed from source nodes to destination nodes. The term “Ad hoc” means either “using what is on hand” or “for one specific purpose”. Ad hoc networks follow both the meaning i.e. they are formed as they are needed and are configured to handle what is needed by each node.All available nodes are aware of all other nodes within range. The collection of nodes is interconnected in many ways. The network is ad hoc because each node is willing to forward data for other nodes, and so the determination of which nodes forward data is made dynamically based on the network connectivity. This is in contrast to wired networks in which

1

routers perform the task of routing. It is also in contrast to managed (infrastructure) wireless networks, in which a special node known as an access point manages communication among other.The connection is established for the duration of one session and requires no base station. Instead, devices discover others within range to form a network for those computers. Devices may search for target nodes that are out of range by flooding the network with broadcasts that are forwarded by each node. Connections are possible over multiple nodes (multihop ad hoc network). Routing protocols then provide stable connections even if nodes are moving around.

Fig 1 Ad Hoc Network

In the above diagram nodes have been highlighted, showing paths through several nodes. If one of the intermediate nodes fails (i.e. that node leaves that area), the network will automatically reorganize itself, by identifying an alternative path from the source to destination. Typically all available nodes are also network users, each sharing the total data transfer capacity of the particular hardware and operating protocol being used. Since there is no need for central administration of the network configuration, it is most efficient to design the system for autonomous operation of each node.

1.3 MANET Challenges

Being one of the most popular fields of study during the last few years, almost every aspect ofad hoc networks have been explored in some level of detail. Yet, no ultimate resolution to any of the problems is found or, at least, agreed on. On the contrary, more questions have arisen than been answered. The major open problems are:

A. Autonomous- No centralized administration entity is available to manage the operation of the different mobile nodes.

B. Mobility- Nodes are mobile and can be connected dynamically in an arbitrary manner. Links of the network vary timely and are based on the proximity of one node to another node.

C. Device discovery- Identifying relevant newly moved in nodes and informing about their existence need dynamic update to facilitate automatic optimal route selection.

2

Node

Node

Node

Node

Node

Node

D. Bandwidth optimization- Wireless links have significantly lower capacity than the wired links.

E. Limited resources -Mobile nodes rely on battery power, which is a scarce resource. Also storage capacity and power are severely limited.

F. Scalability- Scalability can be broadly defined as whether the network is able to provide an acceptable level of service even in the presence of a large number of nodes.

G. Limited physical security- Mobility implies higher security risks such as peer-to- peer network architecture or a shared wireless medium accessible to both legitimate network users and malicious attackers. Eavesdropping, spoofing and denial-of-service attacks should be considered.

H. Infrastructure-less and self operated- Self healing feature demands MANET should realign itself to blanket any node moving out of its range.

I. Poor Transmission Quality- This is an inherent problem of wireless communication caused by several error sources that result in degradation of the received signal.

J. Ad hoc addressing- Challenges in standard addressing scheme to be implemented.

K. Network configuration- The whole MANET infrastructure is dynamic and is the reason for dynamic connection and disconnection of the variable links.

L. Topology maintenance- Updating information of dynamic links among nodes in MANETs is a major challenge.

1.4 Effect of mobility on routing in Ad Hoc network

An ad hoc network consists of nodes that communicate with each other without the help of pre-existing infrastructure. The links between the nodes may change and the network adapts rapidly to the new situation. The freedom of movement makes wireless communication very attractive. But at the same time mobility brings challenges owing to bandwidth and power constraints, limited or no infrastructure and mobility of users.

When a link between two nodes that is in use disconnects, the routing protocol needs to adapt to the new situation. This creates a cost both in the amount of control traffic and in the message delay. When signaling overhead increases the energy consumed by the network will in turn increase which leads to a reduced network lifetime. Also because of the quick topology changes due to mobility of the nodes, ordinary routing protocol fails to give good performance. Since nodes are mobile, most of the time a lot of undesirable effects such as disconnection, bit errors, reduction in throughput, etc. take place In order to evaluate the impact of mobility while simulating a MANET routing protocol, it is crucial that the underlying mobility model accurately emulates real-world node mobility or at least the essential characteristics.

The mobility of nodes in a network may be studied via mobility models. Mobility models imitate node mobility on real networks and come in two varieties: network traces and synthetic models. Network traces are those mobility patterns that are observed in real life systems. Traces provide accurate information, especially when they involve a large number of participants and an appropriately long observation period. However, new network environments (e.g. ad hoc networks) are not easily modeled if traces have not yet been created. In these situations it is necessary to use Synthetic models.

3

Synthetic mobility models mimic node mobility through stochastic processes. Simulation studies use synthetic mobility models to generate mobility data algorithmically without concern for storage or collection. A common synthetic mobility model is random waypoint mobility model. Random waypoint forces nodes to select a random waypoint within a predefined set of coordinates, and then travel in a straight line to the waypoint and pause between direction changes. Random waypoint accepts a node speed and pause time. Simulation studies commonly use random waypoint models to simulate MANET behavior [T camp].

The synthetic models used for movement pattern generation should reflect the movement of the real mobile devices, which are usually carried by humans, so the movement of such devices is necessarily based on human decisions. ‘Regularity’ is an important characteristic of human movement patterns. All simulated movement models are suspect because there is no means of accessing to what extent they map reality. However it is not difficult to see that random mobility models such as Random Walk, Random Waypoint (default model used in almost all network simulations), etc., generate movements that are not suitable in all situations. Hence we need to focus on more realistic situation oriented mobility models such as Gauss Markov, Manhattan Grid and Reference Point Group Mobility Model (RPGM).

4

Chapter 2: Background Work----------------------------------------------------------------------------------------------------------------------------------------------------------

Most of the researchers use Random Way Point mobility model to determine the performance of different ad hoc routing protocols. This is a widely used mobility model. In this model, each node selects a random point in the simulation area as its destination, and a speed v from an input range [vmin, vmax]. The node then moves to its destination at its chosen speed. When the node reaches its destination, it rests for some pause time. At the end of this pause time, it selects a new destination and speed and resumes movement.

In [1] the researchers have used Random Waypoint Mobility model. They have compared the performance of four ad hoc routing protocols namely AODV, DSDV, TORA and DSR. Their protocol evaluation is based on 50 wireless nodes, moving about over a rectangular (1500mX300m) flat space for 900 seconds of simulation time. They have generated 7 different pause times: 0, 30, 60, 120, 300, 600 and 900 seconds. With CBR traffic sources, they have experimented with 1, 4 and 8 packets sending rate per second. Performance metrics used are Packet delivery ratio, Routing overhead and Path optimality. They concluded that DSDV performs quite predictably, delivering virtually all data packets when node mobility rate is low but and failing to converge as node mobility increases. The performance of DSR was very good at all mobility rates and movement speeds, although its use of source routing increases the number of routing overhead bytes required by tie protocol.Finally, AODV performs almost as well as DSR at all mobility rates and movement speeds and accomphshes its god of eliminating source routing overhead but it still requires the transmission of many routing overhead packets and at high rates of node mobility is actually more expensive than DSR.

In [2] the researchers have used Random Way Point Model to compare the performance of AODV and DSR protocol. Two field configurations are used-1500 mX300 m field with 50 nodes and 2200m X 600m field with 100 nodes. Node speed is randomly distributed between 0-20 m/sec. They have used three performance metrics namely Packet delivery fraction, Avg. End-to-end delay and Normalized routing load.Their conclusion is that DSR outperforms AODV in less “stressful” situations, i.e., smaller number of nodes and lower load and/or mobility. AODV, however, outperforms DSR in more stressful situations, with widening performance gaps with increasing stress (e.g., more load, higher mobility). DSR, however, consistently generates less routing load than AODV.

In performance comparison of ad hoc routing protocols using Random way point model [3] the researchers used field configuration of 500 m X 500 m field with 50 nodes and node speed is randomly distributed between 0-20 m/sec. The performance metrics used are Packet delivery fraction, Avg. end to end delay, normalized routing load and Normalized MAC load.

The general observation from the simulation is that for application-oriented metrics such as packet delivery fraction and delay. AODV, outperforms DSR in more “stressful” situations (i.e., smaller number of nodes and lower load and/or mobility), with widening performance gaps with increasing stress (e.g., more load, higher mobility).

Papers in [4], [5] and [6] also use Random way point model as movement model in their experiments.

5

Except Random way point model there exists a variety of mobility models that have been used in the simulation based studies of mobile systems. A survey and simulation based comparison of variety of mobility models can be found in [7]. They have presented the survey on Entity mobility models (like Random walk, Random direction, Gauss Markov, Manhattan Grid etc.) and group mobility models (Reference point group mobility model, Column mobility model etc.). A brief description of some of the mobility models from this paper has been given below.

a) Random walk mobility model: The Random Walk Mobility Model was first described mathematically by Einstein in 1926 [1]. Since many entities in nature move in extremely unpredictable ways, the Random Walk Mobility Model was developed to mimic this erratic movement. In this mobility model, an MN moves from its current location to a new location by randomly choosing a direction and speed in which to travel. The new speed and direction are both chosen from pre-defined ranges, [speedmin; speedmax] and [0;2π] respectively. Each movement in the Random Walk Mobility Model occurs in either a constant time interval t or a constant distance traveled d, at the end of which a new direction and speed are calculated. If an MN which moves according to this model reaches a simulation boundary, it “bounces” off the simulation border with an angle determined by the incoming direction. The MN then continues along this new path.

b) Random direction mobility model: In this model, mobile nodes (MNs) choose a random direction in which to travel similar to the Random Walk Mobility Model. An MN then travels to the border of the simulation area in that direction. Once the simulation boundary is reached, the MN pauses for a specified time, chooses another angular direction (between 0 and 180 degrees) and continues the process.

c) Gauss Markov mobility model: Gauss-Markov mobility model creates random movement changes that are dependent on node's current speed and direction. At fixed intervals the simulator generates a new speed and direction based on their current values and standard deviations. In addition the model keeps nodes away from the edges by changing their direction away from them should they get too close.

d) Manhattan Grid mobility model: This model emulates the movement pattern of mobile nodes on streets defined by maps. It is useful in modeling movement in an urban area. Maps are used in this model which is composed of a number of horizontal and vertical streets. Each street has two lanes for each direction (north and south direction for vertical streets, east and west for horizontal streets). The mobile node is allowed to move along the grid of horizontal and vertical streets on the map. At an intersection of horizontal and Vertical Street, the mobile node can turn left, right or go straight. This choice is probabilistic: the probability of moving only on the same street is 0.5, the probability of turning left or right is 0.25. The velocity of a mobile node at a time slot is dependent on it’s velocity at the previous time slot [18-26.pdf].

e) Reference point group mobility model: The Reference Point Group Mobility (RPGM) model represents the random motion of a group of MNs as well as the random motion of each individual MN within the group [13]. Group movements are based upon the path traveled by a logical center for the group. The logical center for the group is used to calculate group motion via a group motion vector, G~M. The motion of the group center completely characterizes the movement of its corresponding group of MNs, including their direction and speed. Individual MNs

6

randomly move about their own pre-defined reference points, whose movements depend on the group movement. As the individual reference points move from time t to t+1, their locations are updated according to the group’s logical center. Once the updated reference points, RP (t+1), are calculated, they are combined with a random motion vector, R~M, to represent the random motion of each MN about its individual reference point.

f) Column mobility model: The Column Mobility Model proves useful for scanning or searching purposes. This model represents a set of MNs that move around a given line (or column), which is moving in a forward direction (e.g., a row of soldiers marching together towards their enemy). A slight modification of the Column mobility Model allows the individual MNs to follow one another (e.g., a group of young children walking in a single-file line to their classroom). For the implementation of this model, an initial reference grid (forming a column of MNs) is defined. Each MN is then placed in relation to its reference point in the reference grid; the MN is then allowed to move randomly around its reference point via an entity mobility model. (The authors propose using the Random Walk Mobility Model)

The paper in [8] the authors have examined two on demand routing protocols for Mobile ad hoc networks based on Manhattan Grid mobility model. In their study CBR (Constant bit rate) traffic sources are used with packet size of 512 bytes. Map size is 1000 m X 1000 m and means speed varies from 5-25 m/sec. Connection rate is 10 packets/sec and pause time is 10 sec. The simulation is carried out for 200 sec. Performance metrics used are packet delivery fraction, Avg. end to end delay, Normalized routing load and normalized MAC load.

They have concluded that relative performance of protocols may vary with the mobility model used. For varying speed of nodes AODV protocol has the worst performance for packet delivery ratio despite being having the best results in Random way point model. In this model which has high relative speed AODV seems to achieve as good a throughput as DSR.

The paper in [9] analyses the performance of AODV and DSR under two mobility models namely boundless simulation area and probabilistic random walk mobility model. They have observed that AODV yields good performance for high/low mobility; but the performance of DSR is good for low traffic and low mobility.

In [10] the authors have evaluated the performance of ad hoc routing protocols under Random way point model and Reference point group mobility model. Different parameters for the mobility models are as below.Simulation area is 1000m x 1000m; minimum node speed for RPGM is 1 m/sec and for RWP is uniform; number of nodes is 20; traffic type is CBR; packet size is 512 bytes; connection rate is 10 packets/sec; pause time is 25 sec.; number of connection is 5; maximum node speed varies from 10 m/sec to 25 m/sec in an interval of 5 m/sec.; simulation time is 200 sec. and number of nodes in each group is 2. Performance metrics used are packet delivery fraction, Normalized routing load and Avg. end to end delay.

Their observation is that the mobility pattern influences the performance of MANET routing protocols. DSR and AODV achieve the highest throughput and least overhead with RPGM when compared to RWP mobility models. From their results, it is analyzed that AODV has better throughput and less delay in RPGM model when compared to RWP model.

In [11] the authors have studied the effect of various random mobility models namely random way point, random walk with reflections and random walk with wrapping on the performance of AODV.

7

The simulation period for each scenario is 900 seconds and the simulated mobility network area is 800 m x 500 m rectangle. Simulation runs are made with the number of nodes varying from 5 to 25. In each simulation scenario, the nodes are initially located at the center of the simulation region. The nodes start moving after the first 10 seconds of simulated time. The MAC layer protocol IEEE 802.11 is used in all simulations. The transmission range is 250m. The application used to generate is CBR traffic and IP is used as Network layer protocol. They have selected the Packet delivery ratio, Average end to-end delay and Control overhead as metrics during the simulation in order to evaluate the performance of the different protocols.

They have observed that the Random way point produces the highest throughput. But the throughput of the Random walk with reflection and Random walk with wrapping drastically falls over a period of time and also Random walk with reflection shows moderate packet delivery ratio. They have observed that Random walk with wrapping has the highest delay with node density as well as mobility. Also the effect of the routing overhead is very less with Random walk model with wrapping. But the other two models suffer a lot from routing overhead packets.

In [12] influence of three mobility model namely Column, Pursue and RPGM mobility model, has been explored on AODV, DSDV and DSR. The simulation parameters are as follows: transmitter range 250 m, bandwidth 2 Mbps, simulation time 900 sec, number of nodes 50, environment size 1500 m x 300 m, traffic type CBR (Constant bit rate), packet rate 4 pkts/sec, number of connections 20, the speed of the nodes used here are 1 m/sec, 5 m/sec, 10 m/sec, 15 m/sec and 20 m/sec.Performance metrics used are Packet delivery ratio, Routing overhead, Throughput and Path optimality.

They have concluded that DSR outperforms the other protocols, DSDV performs better than AODV, AODV has high routing overhead, AODV performs better with Column and RPGM model, DSR fares better with Pursue and Random way point model.

In [13] the authors have evaluated the routing strategies of three routing protocols namely AODV, OLSR and SRMP, and compare their performances on two mobility models: Random way point and RPGM model. The network parameters used here are: network size 500 m x 500 m, traffic model is CBR, node transmission distance is 250 m, data packet size is 512 bytes, RPGM group size is 4, and sending buffer is 64 packets and simulation time is 100 seconds, node speed is fixed at 20 m/sec and pause time varies from 0 to 100 sec.Performance metrics used are network throughput and end to end delay.

They have conclusion that proactive protocols are better suited to CBR traffic, as the results in this paper showed that the proactive protocol SRMP gets high throughput and low delay in spite of the mobility model and mobility scenarios, Source routing strategy combines multicasting can get good performance. Proactive and reactive strategy can not determine the protocol performance alone, OLSR is better suited to group communication model RPGM.

Most of MANET simulations based on random mobility models, e.g. random waypoint model, models insufficient to reflect the environmental constraints. In [14] the authors have used a combine mobility model to analyze the effect of diverse mobility pattern (Random Waypoint Mobility Model and Manhattan Mobility Model) in indoor and outdoor environment to get a realistic simulation. In these models the movements of mobile nodes are

8

either influenced by building or restricted to street. They have demonstrated the utility of the mobility model by evaluating various MANET routing protocols, including DSR, AODV and DSDV. Simulation environment consists of 10 wireless nodes forming an ad hoc network, moving bout over a 1000 X 1000 flat space for 900 seconds of simulated time. Each run of the simulator accepts as input a scenario file that describes the exact motion of each node and the exact sequence of packets originated by each node, together with the exact time at which each change in motion or packet origination is to occur. They have generated different scenario files with varying movement patterns and traffic loads (CBR), and then ran all three routing protocols against each of these scenario files. After that they have compared the performance results of the three protocols. When Nodes are in street they are move according to Manhattan grid model; otherwise they move as Random Waypoint model. They have run their simulations with movement patterns generated for 6 different maximum speeds, 10, 20, 30, and 40, 50, 60 with constant pause time. In comparing the protocols, they have chosen to evaluate them according to the following metrics: packet sent throughput, packet delivery ratio, packet overhead, packet dropped.

They have concluded that AODV and DSDV give better performance than DSR when throughput is considered as metric. Since DSR pre-computes the routes before sending the packets its packet delivery ratio is better than other protocols. Routing overhead of DSDV and DSR protocols is significantly low. DSR drops few packets but its throughput is very low. DSDV is better than AODV protocol in dropping packets.

The majority of existing mobility models for ad hoc networks do not provide realistic movement scenarios; they are limited to random walk models without any obstacles. In [15] the authors have proposed a more realistic movement model through the incorporation of obstacles. These obstacles are utilized to both restrict node movement as well as wireless transmissions. Performance metrics used are data packet reception, control packet overhead and end to end delay.The simulation area is 1000m × 1000m, and the maximum node transmission range is 250m. However, in the presence of obstructions, the actual transmission range of each individual node is likely to be limited. The propagation model is the two-ray path loss model. At the MAC layer, the IEEE 802.11 DCF protocol is used, and the bandwidth is 2Mbps. Because here a campus environment has been modeled, the mobility of the nodes, unless otherwise stated, is randomly selected between 0 and 5 m/s to represent walking speeds. The pause time in the simulations is also randomly selected between 10 and 300 seconds. Hence, when a node reaches its intended destination, it pauses for a certain period of time and then selects a new destination.

The conclusion is that the effect of mobility significantly impacts the performance of an ad hoc network routing protocol. Through the use of the AODV protocol, the authors have shown that the mobility model affects a variety of characteristics, including the connectivity of the nodes and network density, as well as the packet delivery and overhead of the routing protocol.

9

Chapter 3: Design of the problem statement----------------------------------------------------------------------------------------------------------------------------------------------------

3.1 Introduction

Mobility is human’s nature. Mobility is an unmistakable truth in human lives and time is always a constraint, while communication is a necessity. Communicating while moving to save time has become a challenge. Mobility management has widely been recognized as one of the most important and challenging problems for a seamless access to wireless networks and mobile services. It is the fundamental technology used to automatically support mobile terminals enjoying their services while simultaneously roaming freely without the disruption of communications. Two main aspects need to be considered in mobility management, i.e. location management (ex. addressing, location registration and update, tracking and paging, etc.) and handoff management (ex. handoff trigger and initiation, connection routing, smoothing, etc.). Also in Mobile Ad Hoc Networking (MANET) nodes are mobile. So, mobility is to be managed here also and is a challenging feature of MANET.

3.2 Mobility for Mobile communication

Mobility affects mobile communications on all the components, including devices, networks, and services. To a mobile device, besides the physical requirements like weight, size, power, display, and shape, there still exist other functional requirements ex. different user interfaces suitable to mobility scenario and the computing and communication capabilities distribution. To a service for mobile case, the most important effect is the requirement on adaptation in which a mobile service should be adaptive to different transmission links, different user mobile devices, and different using contexts. In particular here, the focus is on the impacts of mobility on the protocols of networks.

3.2.1 Effect of mobility on network architecture

For network architectures different mobility modes can be distinguished resulting in different types of network architectures and communication usages. The mobility modes can be divided into three main classes according to the different spatial-temporal relations, including:

1) Nomadic or portable communications: Here no network connection is needed during the movement and a new connection will be re-established only after the mobile node has arrived at its new location. Mobility can be either transparent or aware to other nodes. In this scenario only location management is significant and no handoff management is needed. Moreover, portable communications are not necessarily based on wireless networks. This is a quasi-mobile communication mode.

2) Cellular communications: Here the wireless network is organized as a cellular structure in order to enable frequency reuse. This is also traditionally known as mobile communications. Each cell encompasses a certain distance and coverage. Continuous connectivity should be

10

provided when an on-served mobile node is moving from one cell into another (maybe either neighboring or overlapping cell). Both location management and handoff management are necessary in the scenario, while handoff management is only invoked when on-serving movement out of a certain area.

3) Pervasive communications: Here, the communications between mobile nodes are ubiquitous and even invisible. The scenario is mostly based on a dynamic on-the-fly set-up without using any pre-existing network infrastructure, known as mobile ad hoc networking. It is an autonomous system in which mobile hosts connected by wireless links are free to move randomly and often act as routers at the same time. Location management scheme in ad hoc networks is involved into the routing strategies and mainly used in the cluster-based hierarchical multi-hop ad hoc networks, while handoff management is only treated with fast location management.

3.2.2 Effects of Mobility on Protocol stack

The feature of mobility also affects the whole protocol stack, from the physical, data link, and network layers up to the transport and application layers.

1) At the physical layer, mobility influences are remarkable since most mobile communications are based on wireless media like radio. A wireless channel varies with most mobility factors e.g. velocity, direction, place (outdoor or indoor), etc. Resource reuse and avoiding interference are two important problems at the physical layer.

2) At the data link layer, mobility based on wireless networks brings problems of bandwidth, reliability, and security, for which compression, encryption, and error correction techniques are needed. Other problems include fixed or dynamic channel allocation algorithms, collision detection and avoidance measures, QoS resource management, etc.

3) At the network layer, mobility of mobile nodes means that new routing algorithms are needed in order to change the routing of packets destined for a moving node to its new point of attachment in networks. How to track a node’s movement and how to keep the moving node’s connectivity are two basic issues at the network layer. This in turn forms the two main operations of mobility management.

4) At the transport layer, an end-to-end connection may mix wired and wireless links. This makes congestion control a complex task due to the different characteristics of wired and wireless networks, since packet loss is caused mainly by high error rates and handoff in wireless networks instead of because of congestion—the situation on wired links. Retransmission mechanism based on increasing interval may lead to an unnecessary drop in the date rate. Function distribution between the transport and the data link layer is a new problem caused by mobility.

5) At the middleware and application layer, mobility brings new requirements on middleware supports. Examples include service discovery schemes, QoS management, and environment auto configuration. Device-aware applications are important to adapt to different types of user devices, while connection-aware applications are needed to adapt to the changing conditions of network connectivity. Besides these challenges, mobility also brings new opportunities to applications. Context-aware applications are possibly based on the measures for sensing miscellaneous context information of mobile end users.

11

3.3. Mobility management

Mobility management is the essential technology that supports roaming users with mobile terminals to enjoy their services through wireless networks when they are moving into a new service area. The serving networks can be of any type, e.g. the Internet or intranet, mobile ad hoc networks, personal communications systems (PCS), or the mix of these networks. The mobile node can freely change its point of attachment to the networks. The main function of mobility management is then to efficiently support the seamless roaming of the mobile users and/or devices within the whole serving networks. From the viewpoint of functionality, mobility management mainly enables communication networks to:

Locate roaming terminals in order to deliver data packets, i.e. function for static scenario.

Maintain connections with terminals moving into new areas, i.e. function for dynamic scenario.

According to the concept above, mobility management contains two distinct but related components: location management and handoff management. The former concerns how to locate a mobile node, track its movement, and update the location information, while the latter focuses mostly on the control of the change of a mobile node’s access point during active data transmission. One usage scenario may invoke either one or both of the two managements.

There are still many other aspects concerning the network management of mobility, ex. mobile QoS and resource management, mobile security and privacy, billing, power management, etc.

3.3.1 Mobility management at Network layer

Network layer offers routing for packets from one network to another through independent links according to the destination address. The physical location of a mobile unit can no longer decide its address in a network. Since mobility, modeled as changing node’s point of attachment to the network is essentially an address translation problem, it is naturally best solved at the network layer by changing the routing of datagram destined to the mobile node to arrive at the new point of attachment.

The major change which layer may undergo in case of mobility is the change of IP address when ever mobile device enters new network. The address change leads to the challenge of location management which apparently is another responsibility of network layer. Change of network requires the device to be configured to the new network setting, device should be able to get a new IP address and updating any naming service so that it can be reached by the corresponding hosts (location update).

12

3.4 Evaluation of mobility management schemes

3.4.1 Simulation based evaluation method

The performance evaluation of the mobility management schemes of the future mobile communication systems is becoming more and more difficult and complex a task. Future mobile communication systems evolve with the trend of global connectivity through the internetworking and interoperability of heterogeneous wireless networks. Roaming in such a network architectures is very complex a situation and causes many new problems. The future mobile systems should support a huge number of subscriber population with diverse movement modes. The complexity in future mobile networks brings the performance evaluation many new challenges under study.

Currently simulation based study has gained more attention. In simulation based method, simulation model of the target mobile communication system is used for the evaluation.The main challenge of simulation-based method is that to what details the simulation should be made. The complexity in detailed modeling both target mobile system and workload characteristics can make experiments unfeasible due to the experiment time burst and massive computing power needed. Another problem may be that it’s difficult to generalize the results from a series of experiments in one specific mobile system in order to predict the performance parameters of other related mobile systems.

3.4.2 Simulation modeling issues

The performance of mobility management schemes strongly depends on workload characteristics. Consequently, accurate workload models are needed to specify the different behaviors of subscribers in terms of both user mobility patterns and communication traffic patterns, as illustrated in Fig. 3. Two user patterns need to be simulated for evaluation experiments, including mobility and traffic patterns. Mobility models characterize user movement patterns. Traffic models describe the condition of mobile services. The combinations between user units (individual or group) with user behavior patterns (mobility and traffic) lead to different models that are finally used for various evaluation purposes.

For mobility modeling, both analytical and simulation models can be employed to describe the activity of user’s movement. Analytical mobility models base on simplifying assumptions regarding the movement behaviors of subscribers and can provide performance parameters for simple cases through mathematical calculations. Simulation models consider more detailed and realistic mobility scenario and then can derive valuable solutions for more complex cases. Typical mobility models include Brownian model, random walk model, random waypoint model, Gauss-Markov model, Markovian model, Manhattan grid model, mobility vector model, reference point group model, pursue model, nomadic community model, column model, fluid flow model, exponential correlated random model, etc.

13

3.5 Analysis and modeling of mobility models

To thoroughly and systematically study a new Mobile Ad hoc Network protocol, it is important to simulate this protocol and evaluate its protocol performance. Protocol simulation has several key parameters, including mobility model and communicating traffic pattern, among others.

The mobility model is designed to describe the movement pattern of mobile users, and how their location, velocity and acceleration change over time. Since mobility patterns may play a significant role in determining the protocol performance, it is desirable for mobility models to emulate the movement pattern of targeted real life applications in a reasonable way. Otherwise, the observations made and the conclusions drawn from the simulation studies may be misleading. Thus, when evaluating MANET protocols, it is necessary to choose the proper underlying mobility model. For example, the nodes in Random Waypoint model behave quite differently as compared to nodes moving in groups. It is not appropriate to evaluate the applications where nodes tend to move together using Random Waypoint model. Therefore, there is a real need for developing a deeper understanding of mobility models and their impact on protocol performance.

One intuitive method to create realistic mobility patterns would be to construct trace-based mobility models, in which accurate information about the mobility traces of users could be provided. However, since MANETs have not been implemented and deployed on a wide scale, obtaining real mobility traces becomes a major challenge. Therefore, various researchers proposed different kinds of mobility models, attempting to capture various characteristics of mobility and represent mobility in a somewhat 'realistic' fashion. Much of the current research has focused on the so-called synthetic mobility models [T. Camp] that are not trace-driven.

In the previous studies on mobility patterns in wireless cellular networks, researchers mainly focus on the movement of users relative to a particular area (i.e., a cell) at a macroscopic level, such as cell change rate, handover traffic and blocking probability. However, to model and analyze the mobility models in MANET, the focus is the movement of individual nodes at the microscopic-level, including node location and velocity relative to other nodes, because these factors directly determine when the links are formed and broken since communication is peer-to-peer.

3.5.1 Requirement of analyzing routing protocols under different mobility models In the performance evaluation of a protocol for an ad hoc network, the protocol should be tested under realistic conditions including, but not limited to, a sensible transmission range, limited buffer space for the storage of messages, representative data traffic models, and realistic movements of the mobile users (i.e., a mobility model). There are two types of mobility models: the mobility model in which the movements of the mobility models are independent of each other is called entity mobility model and the mobility model in which node movements are dependent of each other is called group mobility models.

In chapter 2 some entity mobility models (random way point, random direction, gauss Markov, and Manhattan grid) and some group mobility models (column model, reference point group mobility model) have been discussed.

One frequently used mobility model in MANET simulations is the Random Waypoint model, in which nodes move independently to a randomly chosen destination with a randomly selected velocity. The simplicity of Random Waypoint model may have been one reason for

14

its widespread use in simulations. However, MANETs may be used in different applications where complex mobility patterns exist. Hence, recent research has started to focus on the alternative mobility models with different mobility characteristics. In these models, the movement of a node is more or less restricted by its history, or other nodes in the neighborhood or the environment.

3.5.2 Limitations of random Way point model

For some mobility models, the movement of a mobile node is likely to be affected by its movement history. This type of mobility models are referred as mobility model with temporal dependency. In some mobility scenarios, the mobile nodes tend to travel in a correlated manner. These models are known as mobility models with spatial dependency. Another class is the mobility model with geographic restriction, where the movement of nodes is bounded by streets, freeways or obstacles.

The Random Waypoint model and its variants are designed to mimic the movement of mobile nodes in a simplified way. Because of its simplicity of implementation and analysis, they are widely accepted. However, they may not adequately capture certain mobility characteristics of some realistic scenarios, including temporal dependency, spatial dependency and geographic restriction:

Temporal Dependency of Velocity: In Random Waypoint and other random models, the velocity of mobile node is a memory less random process, i.e., the velocity at current epoch is independent of the previous epoch. Thus, some extreme mobility behavior, such as sudden stop, sudden acceleration and sharp turn, may frequently occur in the trace generated by the Random Waypoint model. However, in many real life scenarios, the speed of vehicles and pedestrians will accelerate incrementally. In addition, the direction change is also smooth.

Spatial Dependency of Velocity: In Random Waypoint and other random models, the mobile node is considered as an entity that moves independently of other nodes. This kind of mobility model is classified as entity mobility model. However, in some scenarios including battlefield communication and museum touring, the movement pattern of a mobile node may be influenced by certain specific 'leader' node in its neighborhood. Hence, the mobility of various nodes is indeed correlated.

Geographic Restrictions of Movement: In Random Waypoint and other random models, the mobile nodes can move freely within simulation field without any restrictions. However, in many realistic cases, especially for the applications used in urban areas, the movement of a mobile node may be bounded by obstacles, buildings, streets or freeways.

Random Waypoint model and its variants fail to represent some mobility characteristics likely to exist in Mobile Ad Hoc networks. Thus, several other mobility models were proposed.

Gauss Markov mobility model is a mobility model considering temporal dependency. In this model, the velocity of mobile node is assumed to be correlated over time and modeled as a Gauss-Markov stochastic process. When the node is going to travel beyond the boundaries of

15

the simulation field, the direction of movement is forced to flip 180 degree. This way, the nodes remain away from the boundary of simulation field. In the Gauss-Markov model, the temporal dependency plays a key role in determining the mobility behavior. By observing the mobility behavior of users in real life, it is also observed that the temporal dependency is an important mobility characteristic that should be captured.

The Random way point model does not capture many realistic scenarios of mobility. For example, on a freeway to avoid collision, the speed of a vehicle cannot exceed the speed of the vehicle ahead of it. Moreover, in some targeted MANET applications including disaster relief and battlefield, team collaboration among users exists and the users are likely to follow the team leader. Therefore, the mobility of mobile node could be influenced by other neighboring nodes. Since the velocities of different nodes are 'correlated' in space, thus we call this characteristic as the Spatial Dependency of velocity. Reference point group mobility (RPGM) model is a mobility model considering spatial dependency in velocity. In the RPGM model, each group has a center, which is either a logical center or a group leader node. For the sake of simplicity, it is assumed that the center is the group leader. Thus, each group is composed of one leader and a number of members. The movement of the group leader determines the mobility behavior of the entire group.

In most real life applications, it is observed that a node’s movement is subject to the environment. In particular, the motions of vehicles are bounded to the freeways or local streets in the urban area, and on campus the pedestrians may be blocked by the buildings and other obstacles. Therefore, the nodes may move in a pseudo-random way on predefined pathways in the simulation field. This kind of mobility model is called a mobility model with geographic restriction. In Manhattan grid mobility model, the movement of mobile node is also restricted to the pathway in the simulation field.

3.6 Problem statement

The main aim of this study is

Acquiring of detailed understanding of ad hoc routing protocols

Implementing the mobility models such as random way point, Gauss Markov, Group mobility model

Analyzing the difference in performance of routing protocols under mobility.

16

Chapter 4: Experimental results----------------------------------------------------------------------------------------------------------------------------------------------------------4.1 Experimental setup

4.1.1 Network simulator 2

The simulations have been carried out using NS 2.33. NS is a discrete event simulator, where the advance of time depends on the timing of events which are maintained by a scheduler. An event is an object in C++ hierarchy with a unique ID, a scheduled time and the pointer to an object that handles the event. NS is based on two languages: an object oriented simulator, written in C++ and an OTcl (an object oriented extension of Tcl) interpreter, used to execute user’s command scripts. NS has a reach library of network and protocol objects. There are two class hierarchies: the compiled C++ hierarchy and the interpreted OTcl one, with one to one correspondence between them.

In the Tcl script provided by the user, one can define a particular network topology, the specific protocols and applications that one wish to simulate (whose behavior is already defined in the compiled hierarchy) and the form of output that is to be obtained from the simulator.4.1.2 Mobility generator

4.1.2.1 The setdest mobility generator

The Random Way Point model is most commonly used mobility model in research of MANETs. This model is provided by the setdest tool in the standard NS 2 distribution. Usage: The syntax to run “setdest” with arguments is as shown below:Syntax:/setdest [-n num_of_nodes] [-p pause time] [-s max speed] [-t simtime] [-xMax x] [-y max y] > [outdir/movement-file]The table below shows the parameters used to generate the Random Way Point model.

PARAMETER VALUE

Model Random Way Point

Mobility Generator Setdest in NS 2

Number of nodes 40

X dimension 500

Y dimension 500

Simulation time 600

Pause time 0.1

Max. speed Varied as speed varies from

5 m/sec to 25 m/sec

17

Table 1: Parameters used to Generate Random Way Point

Movement Pattern

4.1.2.2 The BonnMotion mobility generator

BonnMotion is Java-based software which creates and analyses mobility scenarios. It is developed within the Communication Systems group at the Institute of Computer Science IV of the University of Bonn, Germany. It serves as a tool for the investigation of mobile ad hoc network characteristics. The scenarios generated in this mobility generator can be exported for NS 2. The mobility models that are supported are RW model, GM model, MG model and RPGM model.Usage: All applications described above are started via the "bm" wrapper script. Syntax:/bm <parameters> <application> <application parameters>.Using this mobility generator, three mobility scenarios have been generated, namely Gauss Markov, Manhattan Grid and Reference point group mobility. For Gauss Markov and Manhattan Grid, movement files are generated for five different speeds: 5 m/sec, 10 m/sec, 15 m/sec, 20 m/sec and 25 m/sec. The parameters used for generating these two mobility models are given in Table 2 and 3.Table below shows the parameters used to generate Gauss Markov mobility model. The angle standard deviation varies between 0 and 1.

PARAMETER VALUE

Model Gauss Markov

Mobility Generator Bonn Motion

Number of nodes 40

X dimension 500

Y dimension 500

Simulation time 600

Random Seed 1

Angle Standard Deviation 0.39269908169872414

Max. speed Varied as speed varies from

5 m/sec to 25 m/sec

Table 2: Parameters used to generate Gauss Markov

Movement Pattern

Table 3 below shows the different parameters used to generate Manhattan Grid

mobility file.

18

PARAMETER VALUES

Model Manhattan Grid

Mobility Generator BonnMotion

Number of Nodes 40

Simulation Time 600

X Dimension 500

Y Dimension 500

Number of X blocks 7

Number of Y blocks 7

Maximum Pause time 0.1

Pause Probability 0.1

Mean Speed Varies from 5 m/sec to 25 m/sec

Table 3 Parameters used to generate Manhattan Grid movement

pattern

For RPGM model, five different group scenarios have been generated to study the effect of inter-group dependency of the mobile nodes on the performance of different routing protocols. The number of groups in RPGM is 1, 2, 5, 10 and 20 keeping the number of nodes in the scenarios constant at 40. The different scenario movement files generated and parameters used are shown in Table 4 and 5.

Group scenario Number of Groups Number of nodes per

Group

RPGM1 1 40

RPGM3 5 8

RPGM4 10 4

Table 4 Different RPGM scenarios

PARAMETER VALUES

Model Reference Point Group Mobility

Mobility Generator BonnMotion

Number of Nodes per Group Varied from: 40, 20, 8, 4, 2

Simulation time 600

X dimension 500

Y dimension 500

Maximum Pause 0.1

Minimum Speed Varied from: .5 m/sec, 1 m/sec, 1.5 m/sec, 2

m/sec

19

Maximum distance between two nodes in a

group

50

Maximum Speed Varied from 5 m/sec, 10 m/sec, 15 m/sec,

20 m/sec and 25 m/sec

Table 5 parameters used to generate Reference point group mobility model

4.1.3 Traffic model

Here Constant Bit Rate (CBR) traffic connections with send rate of 4 and packet size of 512 bytes for UDP sources has been generated. Random traffic connections of CBR can be setup between mobile nodes using a traffic-scenario generator. This script is available in NS 2.33. It can be used to create CBR and TCP traffics connections between wireless mobile nodes.

Syntax: ns cbrgen.tcl [-type cbr] [-nn nodes] [-seed seed] [-mc connections] [-raterate]

4.1.4 Performance metrics used

Three performance metrics have been used. They are Packet delivery fraction (PDF), Normalized routing load (NRL) and Average end to end delay.

Packet delivery fraction (PDF): this is the ratio of total number of packets successfully received by the destination nodes to the number of packets sent by the CBR source nodes throughout the simulation: PDF = (number of received packets)/ (number of sent packets)

Normalized routing load (NRL): this is calculated as the ratio between the numbers of routing packets transmitted to the number of packets actually received

NRL = (number of routing packets sent)/ (number of data packets received)

Average end-to-end delay (AED): this is defined as the average delay in transmission of a packet between two nodes and is calculated as follows:

AED= (time packet received − time packet sent)/ (total number of packets received).

20

21