New Mobility Metric based on MultiPoint Relay Life Duration

Ali Ouacha

, Noureddine Lakki

, Ahmed Habbani

1,2

and Jamal El Abbadi

Laboratoire d’Electronique et de Communications LEC, Ecole Mohammadia d’Ing´enieurs EMI, Universit´e Mohammed

V-Agdal UM5A, BP 765, avenue Ibn Sina Agdal 10000, Rabat, Maroc

Laboratoire SIME, Ecole Nationale Sup´erieure d’Informatique et d’Analyse des Syst`emes ENSIAS, Universit´e Mohammed

V-Souissi UM5S, BP 713, Avenue Mohammed ben Abdallah Regragui, Agdal 10000, Rabat, Maroc

Keywords:

Age of Death, MPR Selection, Link Duration, Mobility Metric, OLSR and Routing Protocols.

Abstract:

Optimized Link State Routing (OLSR) is a proactive protocol designed to operate in Mobile Ad Hoc Networks

(MANET). In this protocol, the topology is based on MultiPoint Relay (MPR) Mechanism. However, the

loss of one or many MPRs caused by their movement affects the link state of the network. Therefore, the

contribution of this paper is to keep the network links between the nodes and MPRs in stable state as long as

possible. It was done by calculating a new parameter named Average Age of Death which estimates the life

duration of MPRs. The experimental results illustrate that this parameter is affected by the environment (speed

of node, network density and others). This result provides to use this parameter as a new mobility metric that

can be used in the MPR sets Calculation.

1 INTRODUCTION

Currently, devices are becoming increasingly popu-

lar as they provide users communications by using

wireless technology. Those devices form an arbitrary

network, titled MANET(Corson and Macker, ). In

such network, the nodes communication without us-

ing any infrastructure requires a self management of

the network. Consequently, several routing protocols

are dedicated to this kind of network. Those proto-

cols can be classiﬁed on two main categories: The

ﬁrst one is a reactive protocol that builds paths at the

request and the second one is proactive protocol that

maintains an updated routing table by using periodic

exchanges of control messages. One of those proac-

tive protocols, which proved its efﬁciency in large

and dense environments, is the Optimized Link State

Routing (OLSR)(Clausen and Jacquet, 2003).

In MANETs, mobility node sthat present an ad-

vantage to the users is a major handicap in this kind

of networks. Indeed, the mobility of nodes is, in the

most case, the origin of break links caused with the

loss of one or more MPRs which are the key elements

in the deﬁnition of the network topology. Hence,

the requirement to produce techniques to quantify the

node movements degree and produce values that can

be used in the process of building MPR sets or data

paths construction. Our aim is to provide a new ap-

proach to maintain, as long as possible, links between

nodes and their MPRs. This approach is based on the

average age of death for MPRs. Our ﬁrst challenge

in this paper is to show how this new metric will be

calculated. Thereafter, we will study the impact of

the variation of environment parameter on this met-

ric. The second step is to take advantage of this new

mobility metric by its integration in the MPR set se-

lection processes. So, we aim, by selecting nodes with

large average age of death, to augment stability of the

network and prolong lifetime of links which are used

in data transfer.

The rest of this paper is organized as follows. In

the next section we provide an overview on the OLSR

protocol. The third section presents a short descrip-

tion of related works for some different metrics de-

veloped for nodes mobility calculation. The fourth

section is devoted for our contribution concerning the

average age of death for MPRs. It also illustrates

the simulation environment. Before concluding, we

present the results about our work.

2 OLSR OVERVIEW

The Optimized Link State Routing (OLSR) is a proac-

tive protocol proposed by HIPERCOM-INRIA team

and deﬁned by RFC 3626. It is recognized as one

of protocols base used in MANET networks envi-

ronment. In this protocol, every node must periodi-

305

Ouacha A., Lakki N., Habbani A. and El Abbadi J..

New Mobility Metric based on MultiPoint Relay Life Duration.

DOI: 10.5220/0004102203050309

In Proceedings of the International Conference on Signal Processing and Multimedia Applications and Wireless Information Networks and Systems

(WINSYS-2012), pages 305-309

ISBN: 978-989-8565-25-9

 2012 SCITEPRESS (Science and Technology Publications, Lda.)

cally update its routing table. This is accomplished

by broadcasting periodic control messages. However,

to avoid network congestion, OLSR uses the concept

of multipoint relays (MPR). This means, only nodes

selected as MPR are authorized to retransmit trafﬁc

control intended to be broadcasted into the whole net-

work. Therefore, the MPR mechanism reduces sig-

niﬁcantly the number of control messages forwarding

in the network. To calculate the MPR set, each node

in the network use periodic messages named HELLO-

messages received from its neighbors. To diffuse in-

formation about Topology, OLSR uses a second type

of messages named TC-messages (Topology Control

messages). Information acquired from TC-messages

allows nodes to compute and update their routing ta-

ble.

Despite the great success of the OLSR in

MANETs, several factors disrupt its performance.

But the great threat comes from physical mobility of

nodes. It has negative impact on the network topol-

ogy. To reduce this impact, the authors of several

studies (JOA-NG and LU, 1999)(Kumar and Suman,

2011)2011(Larson and Hedman, 1998)(Oudidi et al.,

2010)(Shukla, 2001) suggest many different metrics

to calculate nodes mobility and make it a quantiﬁable

value that can be used in the calculation process of

MPR-set or taken into account in the path computa-

tion. In the Following, we present brief description of

some of such metrics.

3 RELATED WORK

Several Studies propose various metrics to calculate

the mobility of nodes in ad hoc networks. They can be

classiﬁed into several categories (Kumar and Suman,

2011)(Oudidi et al., 2010).

The ﬁrst one includes direct mobility metrics.

They are ﬂexible and easy measurement methods

which are based on direct information extracted from

nodes movement. For example, the speed or average

speed can be one of this simple direct metrics. More-

over, various papers, published by many authors, fall

into this category of metrics. In (Larson and Hedman,

1998), the average speed is based on the relative ve-

locity for two nodes of the network. The average mo-

bility M (relative mobility) of node n can be also rep-

resented as the average change in the average distance

of the node n during a time interval T − ∆t being the

duration of the simulation and ∆t computation time.

Or the average distance of a node n at time t is the

average of the distances separating it from each node

i in the network.

The Paper presented by (Vazifehdan et al., 2012)

is another sample of direct mobility metrics. the main

equation for the authors is how long any two arbitrary

nodes in a wireless ad hoc network with a random

topology can communicate with each other without

interruptiondue to lack of routes between them before

their own battery runs out. To answer this question,

they utilize node-to-node communication lifetime. It

is deﬁned as the duration that two nodes can commu-

nicate with each other without any break. In the case

where communications can be established through

several alternative routes, sometimes communication

might be ended due to the failure of one of the two

nodes (source node and destination node) or failure

of the last available route between them. In other

case, batteries depletion of one or more nodes form

paths can be the reason of the communication failure.

Based on this, to determine node-to-node communi-

cation lifetime, the authors consider the energy con-

sumption rate of nodes.

The second type of mobility metrics is named de-

rived mobility metrics. Those metrics use param-

eters characterizing the state of links in the network

like the rate of changes of link states or the average

length of the path. In (Oudidi et al., 2010), authors

observe that each node in the network can be in one

of four states (the node moves and its neighbors are

ﬁxed, the node is stable and its neighbors are mov-

ing, the node and its neighbors are moving, the node

and its neighbors are immobile) which is caused by

changes in the links state. So, they deﬁne a new met-

ric of measurement of mobility that is based on the

number of nodes entering (or exiting) in the neigh-

borhood in a laps of time ∆t .

With the third category of the mobility metrics,

lifetime of links or routes is used. In the research

made in (Yawut et al., 2007), in order to ameliorate

the OLSR performance, the authors attempt to im-

prove the MPRs selection procedure. They modify

the Link tuple by adding a new ﬁeld named Start Con-

nection Time (Start

t). With this, the node calculates

in the ﬁrst step the link duration (LD) which is the

difference between current time and Start t for each

node in the neighborhood. After that, the procedure of

MPRs selection is modiﬁed in order to choose nodes

with the longest LD when D(y) of MPR candidate

nodes are the same. Where D(y), the degree of node y,

is deﬁned (Clausen and Jacquet, 2003) as the number

of symmetric neighbors of node y excluding all the

members of N and the node performing the computa-

tion and N is the subset (Clausen and Jacquet, 2003)

of neighbors of the node (node with single interface)

performing the computation. For example, with the

nodes having the same value of Willingness, reacha-

bility and Degree, we have to select the one in which

the LD is the greatest.

WINSYS2012-InternationalConferenceonWirelessInformationNetworksandSystems

306

Figure 1: Example of the AAD computation.

If the research presented above is based on the

link duration, in (Lenders et al., 2006) the authors

study the impact of nodes mobility on the link and

on the route lifetime in the network. Their ﬁrst goal

is to determine the reason of link breaks. Therefore,

they differentiate between link transmission failures

caused by movement of one of the linked nodes, and

other failures caused by interference. So, to represent

link and route lifetimes, the authors give two different

viewpoints: In the ﬁrst, the total lifetime of a link or

route describes the time interval between the moment

the link (or route) appeared until it breaks; in the sec-

ond, the residual lifetime represents the time interval

between a sample moment after the creation until the

link or path breaks. However, the residual lifetime is

also considered in the remainingpart of their research.

Since the protocol is particularly suitable for large

and dense networks as the technique of MPRs works

well in this context, we aim to increase the life dura-

tion of elements in this set of nodes. Therefore, the

stability of network becomes an important factor

4 OUR CONTRIBUTION

Before showing how we calculate Age of Death for

MPRs, let us give some deﬁnitions: n

is the node

number i executing the computation. Node is con-

sidered death when it lefts the MPR Set of the node

. ∆t

is the time between two successive MPR Sets

Computation ( Age). nb

mpr out

is the number of

nodes which left the MPR Set in ∆t

period ( Number

of Deaths). N is the number of nodes in the network.

The Average Age of Death for MPRs ( AAD) of

the node n

until the MPR Set computation number M

is deﬁned by equation (1). Figure 1 present an exam-

ple for AAD metric computation.

AAD

∑

k=M

k=0

∆t

∗ nb

mpr out

∑

k=n

k=1

∆t

(1)

The Average Age of Death for the entire network

during the simulation time is deﬁned by equation (2).

AAD =

∑

i=N

i=0

AAD

(2)

5 SIMULATION A RESULTS

5.1 Simulation Environment and

Parameters

In our study, we used a standard version of OLSR

developed by MASIMUM (MANET Simulation and

Implementation at the University of Murcia) which

we integrated in NS2 (version 2.34).

Our network is consisted of a variable numbers

of mobile nodes (20, 30, 40, 50 and 60 nodes for

each simulation) moving in an area of 1000 x 1000

m. Each node moves according to the RWP (Random

WayPoint) mobility model (Bai and Helmy, 2004)

with pause time ﬁxed to 0 s and maximum speed

varies between 5 and 30 m/s with a step of 5. For each

number of nodes and each max speed several simula-

tions are done. The scenario that deﬁnes the nodes

movement is regenerated at the beginning of all sim-

ulations. To generate trafﬁc in the network, 10 nodes

are randomly selected to be a source of CBR (Con-

stant Bit Rate) trafﬁc. And these selected nodes use

UDP (User Datagram Protocol) connections to send

Packets with 512 bytes of size in the order of one

packet every 2.5 second.

5.2 Results and Discussions

Average Age of Death: An MPR leaving the neigh-

borhood of node is considered as dead. In the Fig-

ure 2, we trace the average age of death based on the

speed for different numbers of nodes

0,01

0,02

0,03

0,04

0,05

0,06

0,07

5 10 15 20 25 30

Speed (m/s)

Average Age of Death

(AAD)

20 Nodes

30 Nodes

40 Nodes

50 Nodes

60 Nodes

Figure 2: AAD depending on speed for each number of

nodes.

From this histogram, we see that the increase

number of nodes causes increase of average age of

death for all speeds values from 5 to 30 m/s. But,

with increasing the number of nodes, the average age

of death is reduced. We also note that the average age

NewMobilityMetricbasedonMultiPointRelayLifeDuration

307

of death is greatest for maximum speed and minimum

number of nodes.

The increase in speed which cause the increase of

average age of death is interpreted by the fact that the

increase in speed increases also the rate of changes in

the network. In this respect, when the network topol-

ogy changes, the number of elements in the neighbor-

hood change as well. So the number of MPR varies

potentially with those changes. This is seen also in

highly mobile environmentseven if there is not a large

number of nodes.

Average of the Maximum Life of MPRs: The max-

imum life of MPR is the maximum life time recorded

of MPRs. The Figure 3 shows the average of the max-

imum life of MPRs based on the speed for each num-

ber of nodes.

0,5

1,5

2,5

5 10 15 20 25 30

Speed (m/s)

Average of the maximum

Life of MPRs (Second)

20 Nodes

30 Nodes

40 Nodes

50 Nodes

60 Nodes

Figure 3: Average of the Maximun life of MPRs depending

on speed for each number of nodes.

For all values of the speed, we can see that with

the increase in the number of nodes, the average of

the maximum life of MPRs decreases. But with the

increase in the average speed, the maximum life of

MPR increases.

With high speed, there is always movement in the

network. What explains the increase in the average of

the maximum life of MPRs is that there are many sets

of MPR which are calculated during these periods.

According to the results previously recorded, we

note that our new metric (AAD) shows a close re-

lationship with the network environment (density,

movement of stations). In other words, this metric

reﬂects the life of MPRs depending on the network

environment. Consequently, by using it as a new pa-

rameter in the MPRs selection presses, we give a new

OLSR protocol version that is named as OLSR-AAD.

The results depicted over here are related to the

environment approximately dense where the number

of nodes is equals to 50. In those ﬁgures shown blow

(Figure 4, 5 and 6), we respectively plot the varia-

tion of Packet Delivered Ration (PDR), Normalized

Routing Load (NRL), Average end to end Delay and

Average Throughput Trafﬁc depending on nodes max

speed for three versions of OLSR protocol. The blue

curves represent the standard OLSR (OLSR-STD).

The green curves concern the Mob OLSR (OLSR-

MOB) proposed in [10]. Our version of the OLSR

protocol (OLSR-AAD) is represented with the red

curves.

Packet Delivered Ratio (PDR): The Figure 5 shows

the behavior of the PDF (rate of successful packet de-

livery) depending on the speed of the three versions

of the OLSR protocol. We found that these three

versions retain approximately the same paces for all

values of speed. In the range of 5 to 15 m/s version

OLSR-MOB go beyond the other versions but in other

interval our version OLSR-AAD has increased com-

pared to other versions. In addition, we can see that

OLSR-AAD performs well in terms of PDF, in com-

parison with other protocols (OLSR-STD and OLSR-

MOB) for all the maximum speeds. So our protocol

is more suitable in highly mobile environments.

5 10 15 20 25 30

Speed (m/s)

Packet Delivered Ratio

(PDR)

OLSR STD

OLSR AAD

OLSR MOB

Figure 4: PDR depending on nodes speed.

Average Delay: For the Average time End-to-End

(Moy-End-to-End), we note that in the speed inter-

val from 5 to 15 m/s the values presented by the

three OLSR versions are approached. But from 15

m/s there is a change between the three versions of

OLSR. we Note also that for both values of speed

15 and 30m/s the Average time is the minimum for

the OLSR-AAD Protocol. Note that for the majority

of speed OLSR-AAD protocol optimizes the average

delay between the source and destination especially

in highly mobile environments.

0,2

0,4

0,6

0,8

1,2

1,4

1,6

5 10 15 20 25 30

Speed (m/s)

Average Delay

OLSR STD

OLSR AAD

OLSR STD

Figure 5: Average Delay Depending of nodes max speed.

Average Throughput Trafﬁc: Average throughput

allows to measure the quantity of informationper time

WINSYS2012-InternationalConferenceonWirelessInformationNetworksandSystems

308

unit transmitted in a communication channel accord-

ing to the curve we ﬁnd that for the speeds 5, 10 and

15m/s OLSRMOB protocol optimizes the other two

versions. But in the interval 15 to 30 m/s OLSR-AAD

protocol improves the other two versions of OLSR.

Therefore the modiﬁed version of OLSR protocol is

well suited especially in highly mobile areas.

2200

2400

2600

2800

3000

3200

3400

5 10 15 20 25 30

Speed (m/s)

Average Throughput Traffic

OLSR STD

OLSR AAD

OLSR MOB

Figure 6: Average Throughput Traﬁc depending of nodes

max seeds.

6 CONCLUSIONS

Our problem is the high mobility which would harm

the operating of the network. Nodes, (may be an

MPR) coming out from the neighborhood and cause

changes in the link status of the network, the result is

the rupture of one or more links, a thing which contra-

dicts the state of the routing table before the recalcula-

tion of MPR set. Our proposed solution was to calcu-

late a new metric named Age of death for MPRs that

was always updated and gives an idea of life time of

MPRs. Indeed, the results recorded during the simu-

lations show that this parameter is a close relationship

with the network environment (density, movement of

stations). In other words, this parameter reﬂects the

life of MPRs depending on the network environment.

By its integration in the MPR set calculation proce-

dure, we enhance the performance of the network.

So, compared to the other OLSR protocols versions

(OLSR-STD and OLSR-Mob), our modiﬁed version

of OLSR protocol (OLSR-AAD) is well suited espe-

cially in highly mobile areas.

Despite those better results, OLSR-AAD can be

more perfect if we add other metrics to the MPR se-

lection process or to the routing table calculation. For

example, making paths with nodes having more lev-

els of energy can reduce the link breaks probabil-

ity. Moreover, using clustering metrics can provide

a longer life for MPRs. Due to this, our next goals are

to integrate more metrics in order to increase life du-

ration of paths. Moreover, this metric that is actually

restricted to MPRs can be extended to be applied to

all types of neighbors, whether MPR or not .

REFERENCES

Bai, F. and Helmy, A. (2004). A survey of mobility models

in wireless ad hoc networks. Wireless Ad Hoc and

Sensor Networks.

Clausen, T. and Jacquet, P. (2003). Optimized link state

routing protocol (olsr). Network Working Group, RFC

3626.

Corson, S. and Macker, J. Mobile ad hoc networking

(manet): Routing protocol performance issues and

evaluation considerations. Network Working Group,

RFC 2501.

JOA-NG, M. and LU, I. T. (1999). A peer-to-peer zone-

based two-level link state routing for mobile ad roc

networks. IEEE Journal on Selected Areas in Com-

munications, 17(8):1115 1125.

Kumar, S. S. and Suman, C. S. B. (2011). Classiﬁca-

tion and evaluation of mobility metrics for mobility

model movement patterns in mobile ad-hoc networks.

International journal on applications of graph the-

ory in wireless ad hoc networks and sensor networks

(GRAPH-HOC), 3:25–38.

Larson, T. and Hedman, N. (1998). routing protocols in

wireleww ad-hoc networks- a simulation study, mas-

ter’s thesis, computer science and engineering, scool:

Luela university of technology.

Lenders, V., Wagner, J., and May, M. (2006). Analyzing

the impact of mobility in ad hoc networks realman. In

the 2nd international workshop on Multi-hop ad hoc

networks: from theory to reality : ACM New York, NY,

USA.

Oudidi, K., Hajami, A., and Elkoutbi, M. (2010). Qos rout-

ing for mobile adhoc networks andperformance analy-

sis using olsr protocol. IJCSIS (International Journal

of Computer Science and Information Security), ISSN

1947-5500, 8.

Shukla, D. (2001). Mobility models in ad hoc networks,

scool fresit-itt, bombay.

Vazifehdan, J., Prasad, R. V., and Niemegeersal., I. (2012).

On the lifetime of node-to-node communication in

wireless ad hoc networks. Computur. Networks., 56.

Yawut, C., Dhaou, R., and Paillassa, B. (2007).

The olsr performance enhancement using link

duration irit laboratory enseeiht network

& telecommunication dpt toulouse france.

http://citeseerx.ist.psu.edu/viewdoc/summary?,

doi=10.1.1.101.4124.

NewMobilityMetricbasedonMultiPointRelayLifeDuration

309