Network Management in a Heterogeneous and Sparsely Populated Mobile Ad Hoc Network

Network Management in a Heterogeneous and Sparsely Populated Mobile Ad Hoc Network

By

Shanmugasundaram Natarajan

Submitted to

Dr. Dan Marinescu

School of Electrical Engineering and Computer Science

University of CentralFlorida

OrlandoFL

1. INTRODUCTION

Mobile ad hoc network consists of wireless mobile hosts that communicate with each other, in the absence of a fixed infrastructure. These networks have no fixed routers. All the nodes are capable of movement and can be connected dynamically in an arbitrary manner. Nodes of these networks act as routers, which discover and maintain routes to other nodes in the network. Due to the movement of the hosts, the network’s wireless topology may change rapidly and unpredictably. Due to transmission distance limitations, the nodes may not be able to communicate with each other directly. Hence a multihop scenario occurs, and several hosts may need to relay a packet before it reaches the destination.

Mobile ad hoc networks commonly referred to as MANETs have applications in situations like battlefields, rescue operations, festival grounds, assemblies or major disaster areas where users need to deploy networks immediately without the benefit of base stations or fixed network infrastructures.

Here we are concerned with a special class of ad hoc networks, those where communicating entities range from primitive devices with a very limited communication range, power, and processing capabilities to considerably more sophisticated devices with a far greater range and virtually unlimited power as well as computing capabilities.

1.1Characteristics of MANETs

• Dynamic topologies: Nodes are free to move arbitrarily. Thus the network topology, which is typically multihop, may change rapidly at unpredictable times, and may consist of both bi-directional and unidirectional links.

• Bandwidth Constrained: Wireless links have significantly lower capacity than the hardwired links. Often the realized throughput is much less than the medium’s maximum transmission rate. Usually in this scenario congestion is a norm than exception.

• Energy Constrained Operation: Some or all of the nodes in a MANET may rely on batteries or other exhaustible means for their energy. For these systems the most important design criteria may be energy conservation.

• Limited Physical Security:Mobile wireless networks are generally more prone to physical security threats than the fixed cable networks. The increased possibility of eavesdropping, spoofing and denial of service attacks should be considered.

1.2 Classification of MANETs

MANETs can be typically classified into

• Sparsely populated and
• Densely populated

Each of these types has specific characteristics. Sparsely populated MANETs, due to lesser number of nodes have less competition to the available limited resources. But the lesser number of nodes may result in the partition of the network and some of the nodes may become unreachable due to host mobility. In a densely populated network, on the other hand there is a stiff competition to the limited resources. Even though host reachability is not a problem, the routing overhead is usually high.

1.3 Problem Definition

This paper attempts to find an efficient routing protocol for a sparsely populated heterogeneous MANET taking their specific characteristics into consideration. The paper analyzes proactive and reactive approaches to routing in a sparsely populated MANET.

This paper is organized as follows. Section 2 defines a sparsely populated and heterogeneous MANET and discusses its characteristics. Section 3 discusses the routing protocol for a sparsely populated MANET. Section 4 discusses the performance measures of such a system. Finally Section 5 presents the conclusion.

2Heterogeneous and Sparsely Populated MANETs

2.1Model definition

A heterogeneous network consists of communicating entities that range from primitive devices with limited communication range, power, processing capability to sophisticated devices with a far greater communication range, power etc. A heterogeneous system is shown in Fig.1

These heterogeneous entities combine together to form a sparsely populated mobile ad hoc network. Such a sparsely populated MANET can be formally defined as follows. Each of the nodes in the MANET can be thought of as a set of Mobile Entities MEs.

Let,

MEi and MEj be two mobile stations,

di,j = distance(MEi, MEj)

N be the expected number of mobile stations

R be the radius of the area covered by the whole network

ri be the radius of coverage area of each station

The density,ρ, of MEs in a S-MANET is defined as:

ρ =

where 1< ρ < k with k small.

The proximity set of MEi, pi,j ={ MEj | di,j < ri , } is the number of nodes that can be reached by a transmission from MEi is also relatively small:

E{ pi,j } < k.

Each of the mobile entities provides information sufficient to establish its mobility equation:

MEi (info)= ( ri, location vector( xi , yi,, t ), mobility vector (dxi , dyi))

The mobility vector dxi and dyi will give the speed and the direction of movement of the nodes.

Each of the mobile entities are characterized by the following resource vector,

RVi =(Class,Proci, Poweri, CRi, Storagei,)

Where,

Class gives the general classification of the device. It can be L1, L2 or L3 denoting high resource, medium resource and low resource device.

Proci gives the processing power of the mobile entity

Poweri gives the power available in the entity, which may be represented as the remaining battery life

CRi gives the coverage range of the mobile entity

Storagei gives the secondary storage capacity of the entity

A heterogeneous sparsely populated MANET is hereafter referred to as HSP-MANET.

Such a heterogeneous sparsely populated MANET has asymmetric links. Since some of the entities have less communication capabilities and resources, they can only receive messages and may not be able to send messages.

In such a heterogeneous system some of the less powerful entities may be clustered around a more powerful entity. Such a system should be capable of multi-mode routing. The powerful entities should help the less powerful entities around them in routing. As each of the system is equipped with a GPS receiver, location information should be exploited.

2.2 Characteristics of a HSP- MANET

Due to the lesser number of nodes an HSP-MANET has typically less competition to the limited wireless resources. There is usually enough bandwidth to support the traffic generated by such a system. Due to lesser number of nodes it is possible to maintain the global network topology.

3. Routing in HSP-MANETs

Given the characteristics of a sparsely populated MANET it is tempting to consider a proactive approach to routing. Maintaining the global network topology is always feasible. But the focus should be in reducing the routing overhead to the maximum extent possible to preserve power and perform other optimizations. Taking this into consideration a reactive approach can be followed with efficient caching strategies and broadcasting techniques.

3.1 Related Work

Although numerous protocols have been proposed for routing in mobile ad hoc networks almost all of them consider a densely populated system. Many of the protocols follow a proactive or reactive approach. Many of the protocols follow different approaches such as making use of location information, using hierarchical routing etc. But to the best of our knowledge, none of these protocols give special consideration to a sparsely populated heterogeneous system.

This paper attempts to find an efficient approach for routing in a sparsely populated system taking the heterogeneity into consideration

3.2 Routing in MANETs

One of the main design issues in an ad hoc mobile network is routing. Numerous protocols have been developed for such networks. These protocols have to deal with the typical limitations of these networks such as low bandwidth, higher power consumption and higher error rates. The routing protocols may be generally categorized as

• Table driven or Proactive protocols
• Demand driven or Reactive protocols

Table driven routing protocols attempt to maintain consistent up-to-date routing information from each node to every other node in the network. These protocols require each node to maintain one or more tables to store routing information, and they respond to changes in network topology by propagating updates throughout the network in order to maintain a consistent network topology. Some of the table driven protocols include Destination-Sequenced-Distance-VectorRouting, Cluster Head Gateway Routing and the Source-Tree Routing.

While table driven routing protocols decrease the route acquisition latency in a MANET, it typically increases the routing overhead associated with maintaining the global network topology at any given time.

Source-Initiated or On-Demand routing creates routes only when desired by the source node. These protocols are usually less costly than a proactive protocol when the host mobility is high. When a node requires a route to the destination, it initiates a route discovery process within the network. This process is completed once when a route is found or all possible route permutations have been examined. Once a route has been established it is maintained by a route maintenance procedure until the destination becomes inaccessible along every path from the source or until the route is no longer needed. Some of the on-demand protocols include Ad Hoc On Demand Distance Vector Routing (AODV), Dynamic Source Routing (DSR), Temporally Ordered Routing Algorithm (TORA), Location Aided Routing etc.

On-demand routing protocols have typically less routing overhead when compared to table-driven protocols. But the route acquisition time may be more. The performance of on-demand protocols may be improved in many ways including the use of efficient caching strategies, broadcasting techniques and making use of the location information of the nodes.

3.3 Routing in a heterogeneous ad hoc network

Almost all the protocols for ad hoc networks assume that all nodes have symmetric links. This means that two neighboring nodes are in the transmission range of each other. But in a network of heterogeneous devices, the processing and communication capabilities of devices differ to a large extent. Even between devices of same capabilities, power consumption may vary due to individual computational and communication loads, resulting in different transmission ranges. The links are asymmetric in such a network. Many existing protocols do not support the asymmetry of the links. Some approaches like GAHA (GPS based Hop by Hop Acknowledgement) and GAPA (GPS based Passive Acknowledgement) enhance routing between nodes having asymmetric links. Both GAHA and GAPA support asymmetry at link level and hence they can be applied to other routing protocols. But these approaches do not assume heterogeneous devices.

A protocol for routing between heterogeneous entities having asymmetric links between them is presented here. Section 1.1 describes the model of the system and discusses the assumptions about the system. Section 1.2 discusses the types of messages passed between the entities. Section 1.3 describes the protocol.

3.3.1 The Model and Assumptions

The studied model can be pictorially represented as follows

The devices in the model can be grouped as level 1, level 2, and level 3 devices. Level 3 devices have more computational and communication power than level 1 devices. In general, higher the level, more are the resources. Typically lower level devices are grouped around a higher-level device. Thus the lower level devices form a logical cluster around the higher-level device. The higher-level device can be called the clusterhead. The lower level devices are called the members of the cluster. Note that the nodes do not maintain any separate information pertaining to the cluster they belong to. It is just a logical classification.

3.3.1.1 Assumptions

We have the following set of assumptions about the system.

1. Each of the devices is assumed to have a GPS receiver to provide accurate location information.
2. Each of the devices is assumed to be capable of operating in different power modes; i.e. they are capable of increasing their transmission range by going to a higher power mode.
3. It is assumed that there is always at least one higher-level device within the transmission range of a low level device.
4. The low level device has enough power to actually transmit data, once the route is known.

3.3.2 Types of Messages passed between nodes

The nodes in the network pass different types of messages to facilitate routing between them. The following are the different types of messages or signals.

Beacon Signals: These signals are sent to advertise the position of the node. They typically include the location and resource information.

Beacon Reply Signals: These signals are replies to the beacon signal. These also include the location and resource information of the replying node.

Route Request Message: When a node wants to communicate with any other node and if it does not already know a route to the destination, it sends the route request signal to a higher level device.

Route Reply Message: Any node hearing a route request can reply to that message if it happens to know the route to the destination.

Actual Data: Besides the control messages, the nodes transmit actual data between them.

3.3.3 Protocol Description

3.3.3.1 Data Structures needed to be maintained.

Each node has to maintain a neighborhood table to maintain information about the nodes, for which a link exists in at least one direction. The table has two columns, the Location and the Resources. Typically a node stores the location and resource information of other nodes. A neighborhood table can look like the following:

Whenever a node A hears a beacon signal from node B, it adds the details of node B in its neighborhood table. If node A replies to the beacon signal or if it communicates with node B for some other reason, node B will add details about node A in its neighborhood table column. Initially there will be lesser number of entries in the neighborhood tables, but as communication between nodes grow, more entries will be added in the table.

Each node depending upon the resources can cache the previously used routes for future use. A separate data structure has to be maintained for the caches. Typically level 3 nodes have enough resources and they maintain a cache of previously used routes.

3.3.3.2 Working of the protocol

Whenever a node hears a beacon signal, it adds the sending node’s information in its neighborhood table and it may either choose to reply to the signal or not. In case it chooses to reply to the beacon signal, it helps the sending node to maintain the topology of the cluster.

Whenever a lower level node A wants to communicate with some other node, it contacts one of the higher level nodes B from its neighborhood table by sending a burst signal, either directly or by going into a higher power mode. The node includes its location information, the destination’s location information and the amount of data to transmit. If B has a route to the destination, it sends the route to A (Fig. 2a). If B does not have a route, it initiates a route discovery, obtains a route and sends that route to A (Fig. 2b).

Whenever the clusterhead or any other level 2 nodes give a route to the lower level node, it arbitrates with the intermediate nodes so as to provide an optimum route in terms of power rather than the number of hops. In Fig. 3, node S wants to communicate with node D. The clusterhead may arbitrate with intermediate nodes, A, B, I and select the route S-A-B-D rather than S-I-D if the former route has less overall power consumption.

If any of the higher level nodes happen to be in the path, and the destination node is within its transmission range, then it can transmit the data packets in a single hop to the destination. Thus the heterogeneity of nodes allows the possibility of single hopandmultiple hopsin the same route.

Whenever two clusterheads communicate with each other, they can use a different frequency to avoid interference. The clusterhead nodes can frequently exchange the information about the nodes they know between them, so that they know a route to a node outside their transmission range. The information in the beacon signals has to be encrypted to avoid security breaches.

Whenever a broadcast is sent for route discovery, the approach mentioned in Location Aided Routing protocol (Y. B. Ko and N. H. Vaidya) can be used. This approach avoids broadcast storm by limiting the broadcast to a forwarding zone based on the location information of the source and the destination. Any node receiving the broadcast message re-broadcasts it, only if it is in the forwarding zone. Otherwise it ignores the message.

3.3.3.3 The protocol as a finite state machine

4PERFORMANCE ISSUES

To judge the merit of a routing protocol one needs both qualitative and quantitative metrics with which we can measure the suitability and performance of the protocol.

A list of qualitative properties of a routing protocol should be

• Distributed operation: The protocol should be able to perform a distributed operation

• Loop Freedom: This is a generally desirable to avoid packets spinning around the network. Ad hoc solutions such as TTL values can bound the problem, but a more structured and well-formed approach is generally desirable.

• Demand Based Operation: Instead of assuming a uniform traffic distribution within the network, the algorithm should adapt to traffic pattern on a demand or need basis. If this is done properly, network energy and bandwidth resources can be utilized efficiently at the cost of increased route discovery delay.

• Proactive Operation: If additional latency of demand-based operation is not acceptable and if bandwidth and energy resources permit, a proactive operation is desirable.

• Security: Sufficient security protection to avoid snooping of network traffic, manipulation of packet headers, and redirection of routing messages is needed.

• Sleep Period Operation: As a result of energy conservation, nodes of a MANET may stop transmitting and/or receiving. The protocol should be able to accommodate such sleep periods without adverse consequences.

The following list of quantitative metrics can be used to access the performance of a routing protocol.