Multicast Routing Performance Analysis for Different Mobility Models

On the IEEE 1609.4 Standard using Random Dijkstra Algorithm

Doan Perdana

Department of Electrical Engineering

Faculty of Engineering, University of Indonesia

Kampus Baru UI, Depok 16424, Indonesia

doan.perdana@ui.ac.id

Riri Fitri Sari

Department of Electrical Engineering

Faculty of Engineering, University of Indonesia

Kampus Baru UI, Depok 16424, Indonesia

riri@ui.ac.id

Abstract- Providing an efficient and scalable VANET

multicast routing has multiple issues due to its highly

dynamic topology, the high mobility, and change trajectory On the other hand, mobility models represent real world

scenarios and according that issues above, we need to evaluate

shortest path performance in the mobility models using

random Dijkstra algorithm for IEEE 1609.4 standard.

In this paper, we evaluate the performance mobility

model for multicast routing protocol compared with

broadcast routing in IEEE 1609.4 standard, in terms of

throughput, queuing delay, and number of delivered packets.

The mobility models observed in this work are Manhattan

Mobility Model, STRAW Mobility Model, Traffic Sign Model

and Intelligent Driver Management Model (IDM_IM). The

mobility models is also evaluated performance using random

Dijkstra algorithm. We also evaluates performance safety and

non-safety application using multicast routing for IEEE

1609.4 standard. We use VanetMobiSim and ns 2.34

simulator for the evaluation.

From the simulation, it was found that IDM_IM mobility

model has the worst delay for broadcast routing compared

with multicast routing. We can analyze IDM_IM using spatial

map randomly to generate traffic and multicast routing is

more efficient and scalable routing protocol in VANET

compared with broadcast routing. On the other hand,

STRAW mobility model for multicast routing has the highest

throughput among others. We also evaluate performance

safety and non-safety application impact of different mobility

models for IEEE 1609.4 standard. For non-safety application

using Dijkstra’s algorithm for multicast routing protocol,

IDM_IM mobility model has the worst delay. On the other

hand, for safety application, STRAW mobility model highest

throughput and number of delivered packets among others.

Keywords : IEEE 1069.4/802.11p, Multicast Routing Protocol,

Random Dijkstra Algorithm, Manhattan Mobility, STRAW

Mobility, Traffic Sign, Intelligent Driver Management

I. INTRODUCTION

The Intelligent Transportation Systme (ITS) is moving

significantly toward advanced active safety in the recent

years, and various kinds of research and efforts has carried

out in academia and in industry [1]. Vehicular Ad hoc

Networks (VANETs) serve as the basis for Intelligent

Transportation Systems (ITS) that aims to apply modern

Information and Communication Technologies (ICTs) to

improve the quality, effectiveness, and safety of

transportation systems of future wireless communication

networks.

One of the standards of the IEEE 1609 protocols

family, which supports multichannel operation and

enhances IEEE 802.11p MAC layer, is IEEE 1609.4 [1].

This standard describes seven different channels with

different features and usage [1], there are six service

channels (SCH) and one control channel (CCH) with

different characteristics [1]. Multicast is defined by delivering multicast packets

from a single source vehicle to all multicast members by multi-hop communication [5]. Alam, M. et al. [6] evaluate the performance of mobility models IEEE 1609.4 standard, i.e Manhattan Mobility Model, Random Freeway Mobility Model, Traffic Sign Model, and Intelligent Driver Management Model.

The problem that is being addressed in this paper is to

evaluate the performance multicast routing protocol

impact of different mobility model using random Dijkstra

algorithm for IEEE 1609.4 standard, in terms of

throughput, queuing delay, and number of delivered

packet. The mobility models proposed for IEEE 1609.4

standard is designed and simulated using VanetMobiSim

[7], which represent real world scenarios and realistic

mobility models.

The contribution of this paper is to conduct the best

mobility models for IEEE 1609.4 standard. In addition,

other novell contribution is the performance evaluation of

multicast routing protocol impact of different mobility

models compared with broadcast routing protocol for

IEEE 1609.4 standard. This paper evaluates the random

Dijkstra algorithm implementation in different mobility

models. Finally, this paper also evaluates the performance

safety and non-safety application impact of different

mobility models for IEEE 1609.4 standard.

This paper is organized as follows. In section II, we

provide are related work and motivation. In Section III,

we provide a scenario and simulation for the performance

of different mobility models for the IEEE 1609.4 standard.

In section IV, we evaluate the performance multicast

routing impact of different mobility and evaluates the

performance safety and non-safety application impact of

different mobility models for IEEE 1609.4 standard.

Finally, we conclude the paper and suggest the future work

in Section V.

II. RELATED WOK

Perdana, D. et al. [3] perform study the performance

of the IEEE 1069.4 based on multi-hop dissemination in

vehicular network to get SCH and CCH channel

performance using Enhanced Distributed Channel Access

Function (EDCA). We use [3] to simulate and evaluate the

performance of Multi-channel operations of the IEEE

1609.4 based on channel perfomance using EDCA

implementation.

Alam, M. et al. [6] introduced a new mobility model,

called Integrated Mobility Model (IMM) for Vehicular Ad

hoc Networks. They evaluated the routing protocols

AODV, DSR, and OLSR using Integrated Mobility Model,

Manhattan, and Freeway mobility models are also

evaluated. We use [6] to simulate and evaluate Manhattan

and Traffic sign mobility models.

Choffnes R. D. et al. [8] introduced STRAW, a new

mobility model for VANETs in which nodes move

according to a realistic vehicular traffic model on roads

defined by real street map data. We refer [8] to simulate

and evaluate STRAW mobility model performance for

IEEE 1609.4 standard.

Jagdeep K. et al. [13] analyzes the QoS performance

metrics (routing overhead, packet delivery ratio and

average delay) for unicast and multicast applications in

highway and urban environment with packet size 512

bytes. We refer [13] to simulate and evaluate the multicast

routing performance for IEEE 1609.4 standard.

A. Multicast Routing

According to Yun, W.L. et al. [5], Multicast is defined

by delivering multicast packets from a single source

vehicle to all multicast members by multi-hop

communication. Geocast routing is to deliver a geocast

packet to a specific geographic region. Vehicles located in

this specific geographic region should receive and forward

the geocast packet. On the other hand, Jagdeep K. et al.

[13] defined a multicast group is composed of senders and

receivers. For connecting senders and receivers, each

protocol constructs either a tree or a mesh as the routing

structure. There are some nodes called forwarding nodes

in the routing structure that are not interested in multicast

packets but act as routers to forward them to receivers.

Group members (senders and receivers) and forwarding

nodes are also called tree or mesh nodes depending on the

routing structure.

Figure 1. Multicast/Geocast Routing [5]

B. Broadcast Routing

According to Yun, W.L et al. [5], Broadcast protocol

is utilized for a source vehicle sends broadcast message to

all other vehicles in the network. Broadcast is the last

important operation for a vehicle to disseminate a

broadcast message to all the others in a VANET Tonguz

et al. [5] demonstrate that the broadcast storm problem

causes serious packet collision and packet loss since too

many vehicles simultaneously broadcast messages in a

VANET.

Figure 2. Broadcast Routing [5]

C. STRAW (STreet RAndom Waypoint)

The idea of STRAW mobility model [8] is similar to

that of Random Waypoint, but unlike traditional random

waypoint models, this model determines a vehicle’s

trajectory at each intersection. A vehicle will make a turn

at an intersection with a specified probability that can be

independently assigned to each vehicle [8]. STRAW is a

new mobility model for VANETs in which nodes

move according to a realistic vehicular traffic model on

roads defined by real street map data [8].

D. Manhattan Mobility Model

The Manhattan mobility model was first introduced by

F. Bai et al. [10]. The Manhattan mobility model [10] is

expected to have high spatial dependence and high

temporal dependence. It too imposes geographic

restrictions on node mobility [6,10]. However, it differs

from the Freeway model in giving a node some freedom

to change its direction [6,10]. Node moves on these

horizontal and vertical streets. In real life a node that is

reaching an intersection can turn to left, right or move in a

straight direction [6,10]. In Manhattan model, when the

nodes reaches the intersections they move with 0.5

probability on the same street, 0.25 turning to left and 0.25

turning to a right [6,10]. In Manhattan model the velocity

of the nodes is restricted to the lanes and also there is a

velocity dependency between the two nodes. So

Manhattan mobility model has high spatial and temporal

dependencies and also impose geographic restrictions on

nodes mobility but give some freedom to nodes as to

change its street [6,10]. The Manhattan mobility model is

useful in modeling movement in urban areas [6,10].

Figure 3. Manhattan mobility model using VanetMobiSim [10]

E. Dijkstra Algorithm

According to [9], the original Dijkstra's algorithm,

given a start and an end point, always selects the same

path, even in the presence of multiple available paths with

same weights. For traffic balancing, cars should be able to

select different shortest paths. Random Dijkstra's

algorithm based on traffic congestion changes randomly

generated by traffic mobility model in vehicular network.

III. SIMULATION AND SCENARIO

By using VanetMobiSim [7] and ns 2.34 simulator

[11], we evaluate the performance multicast routing

compared with broadcast routing impact of different

mobility model using random Dijkstra algorithm for IEEE

1609.4 standard. We also evaluate the performance of the

safety and non-safety application of different mobility

models for IEEE 1609.4 standard. We simulate the

scenario with the payload size of 168 bytes (Voice), 200

bytes (video), 256 byte (background and best effort traffic).

The number of cars range from 30 to 70 nodes. The

channel configuration using default values for control and

service channel intervals are 50ms, and the guard intervals

value is 4 ms. Table I. presents all parameters used in our

simulation.While some parameters stay fixed, others are

varied in order for us to observe the changing behavior of

the network.

TABLE I.SIMULATION PARAMETERS

Parameters Values

Number of vehicles 30 - 70 Payload size 168, 200, 256 bytes

Services VoIP, Video,

Background, Best

Effort

CCH slot duration 50ms

SCH slot duration 50ms

Simulation time 20 seconds

Mobility Models Manhattan,

STRAW, Traffic

sign, IDM_IM

MAC Protocol IEEE 1609.4

Transmission range 250 m

n1

n4n2

n3

n6

n5 nX

nY

Fig.4. Path of forwading data from source to destination

on Single-hop Dissemination [3]

Fig. 4 shows the single-hop dissemination. The path of

forwarding data from the source to destination has a direct

route to destination based on Multi-channel operations

IEEE 1609.4 [3].

In our work, we use VanetMobiSim [7] to evaluate the

performance multicast routing of the mobility models in

IEEE 1609.4 standard. Based on [9], VanetMobiSim can

produce detailed vehicular movement traces employing

different motion constraints or traffic generator models.

Taking into account the interaction of the two, we simulate

different traffic conditions through fully customizable

scenarios. We simulate the vehicular movement traces for

IEEE 1609.4 standard using ns 2.34 simulator [11], we

evaluate the performance impact of different mobility

model using random Dijkstra algorithm [15].

IV. RESULT AND EVALUATION

Base on the scenario, in Figure 5, 6, 7, 8, and 9 the

simulation was performed to obtain data according to three

aspects to be measured, i.e the average delay, number of

delivered packet, and throughput.

1) Performance of Multicast Routing for Different

Mobility on the IEEE 1609.4 standard

a) Average Delay

Fig. 5 shows the average of delay multicast

routing compared with broadcast routing for different

mobility models on the IEEE 1609.4 standard using

Dijkstra’s algorithm, by varying the number of nodes.

We focus on the average access delay which

calculate the MAC layer. Delay and access will be used

interchangeably on this work by varying the number of

nodes. This can be seen in Fig. 5.

The following is the equation for the average

delay derived as [4] :

Fig. 5. Average delay of node density with different mobility

using Dijkstra’s algorithm for IEEE 1609.4 standard

From Fig. 5, we found that the average delay of the

IDM_IM mobility models using broadcast routing has the

highest average delay compared with multicast routing.

We can analyze the IDM_IM mobility models generating

traffic from a spatial map randomly. We also evaluate the

average delay using Dijkstra’s algorithm and found that

IDM_IM mobility has the highest average delay compared

with other mobility models. We also evaluate the multicast

routing more efficient and scalable compared with

broadcast routing for the IEEE 1609.4 standard.

b) Throughput

Fig. 6 shows the throughput of multicast routing

compared with broadcast routing for different mobility

impact on the IEEE 1609.4 standard, by varying the

number of nodes using Dijkstra’s algorithm. Throughput Ti

(t) is the rate of successful packet delivery through a

network connection per unit time. We focus on the

throughput calculated at the MAC layer, then Ti (t) derived

as [3,14] :

Throughput Ti (t) = x* (1-p)*d*data rate Where d = DATA/(DIFS+PACKET+SIFS+ACK) x is the number of nodes Ti (t) is the throughput

a is the distance of nodes p is the collision probability for a transmission

Fig. 6 System throughput of node density with different

mobility models using Dijkstra’s algorithm

for IEEE 1609.4 standard

From Fig. 6 shows that performance of STRAW

mobility model has the highest throughput compared with

broadcast routing. We evaluate the throughput

performance using Dijkstra’s algorithm. From [8],

STRAW (STreet RAndom Waypoint), that constrains node

movement to streets defined by map data for real US cities

and limits their mobility according to vehicular congestion

and simplified traffic control mechanisms. We can analyze

STRAW mobility model using realistic vehicular traffic

model on roads defined by real street map data [8], so we

have the mobility impact in IEEE 1609.4 standard

compared with other mobility models.

c) Numbers of Packet Delivered

Fig. 7 shows the numbers of delivered packets

of multicast routing for different mobility impact on the

IEEE 1609.4 standard using Dijkstra’s algorithm, by

varying the number of nodes. The numbers of delivered

packets is defined by varying number of nodes as

parameters change. We focus on the numbers of delivered

packets which calculate on MAC layer.

Let pidle be the probability that a channel is idle

in a given slot, and pbusy its converse [4]. Similarly, let

psuccess be the probability that a slot is occupied by a

successful transmission, and pcoll is the probability that a

collision occurs during a slot.

If we assume a scenario with M nodes with the

above mentioned assumptions, it is easy to verify that pidle,

pbusy, psuccess,and pcoll are as follow [4]

P idle = (1− τ) M

P busy = 1–P idle

P success = M·τ·(1−τ)M−1

Fig.7 Number of delivered packet of node density with different

mobility models using Dijkstra’s algorithm

for IEEE 1609.4 standard

From Fig.7 shows that the performance of

STRAW mobility model has the highest number of

delivered packet compared with boradcast routing. We also

evaluate the number of delivered packet using Dijkstra’s

algorithm. We found that the STRAW mobility model has

the highest number of delivered packet performance

compared with other mobility model. From [8], STRAW

(STreet RAndom Waypoint), constrains node movement in

streets defined by the map data for real US cities and limits

their mobility according to vehicular congestion and

simplified traffic control mechanisms. We can analyze

STRAW mobility model using realistic vehicular traffic

model on roads defined by real street map data [8].

2) Performance of Safety and Non-Safety Application

for Different Mobility Models on the IEEE 1609.4

standard using Dijkstra’s Algorithm

a) Average Delay

Fig.8 shows the average delay for different

mobility models on the IEEE 1609.4 standard using

Dijkstra’s algorithm to evaluate safety and non-safety

application for multicast routing protocol.

Fig.8 Average delay of node density safety and non-safety application

for IEEE 1609.4 standard

From Fig. 8 shows that the average delay of the

IDM_IM mobility models compared with among others

using multicast routing protocol.

We evaluate IDM_IM mobility models generating

traffic from a spatial map randomly. We also analyze that

the performance of safety application more better

performance compared with non-safety application

because from [12], we can get non-safety application (e.g.

multimedia application) need more higher bandwith and

capacity compared with safety application, so average of

delay more higher for non-safety application.

b) Numbers of Packet Delivered

Fig.9 shows the numbers of delivered packets of

different mobility models using Dijkstra’s algorithm by

varying nodes to evaluate safety and non-safety application

for multicast routing protocol on the IEEE 1609.4 standard.

We focus on the numbers of delivered packets which

calculate the MAC layer.

Fig.9 The number of delivered packet of node density with service differentiation capability for IEEE 1609.4 standard

From Fig. 9 shows that the STRAW mobility

model has the highest number of delivered packets for

multicast routing. Based on [12], we evaluate the non-

safety application has more channel (six SCH channel)

compared with safety application (one CCH channel), so

the number delivered packets more higher for safety

application. We can also analyze STRAW mobility model

using realistic vehicular traffic model on roads defined by

real street map data [8].

c) Throughput

Fig.10 shows system throughput of multicast

routing for different mobility models on the IEEE 1609.4

standard using Dijkstra’s algorithm by varying nodes. We

evaluate the safety and non-safety performance based on

different mobility models. We focus on the numbers of

delivered packets the MAC layer.

.

Fig.10 System throughput of node density with service

differentiation capability for IEEE 1609.4 standard

From Fig. 10 shows that the system throughput of

STRAW mobility model has the best performance

compared with among others using Dijkstra’s algorithm.

Based on [8], we evaluate STRAW (STreet RAndom

Waypoint), constrains node movement in streets defined by

the map data for real US cities and limits their mobility

according to vehicular congestion and simplified traffic

control mechanisms. We can also analyze, based on [12],

the non-safety application has more channel (six SCH

channel) compared with safety application (one CCH

channel), so the throughput more higher for safety

application.

V. CONCLUSION

We have conducted that best performance of multicast

routing for different mobility on the IEEE 1609.4

standard. Using Dijkstra’s algorithm, we conclude that the

performance of IDM_IM mobility model has the worst

delay performance, we can analyze IDM_IM using spatial

map randomly to generate traffic. We also evaluate the

multicast routing more efficient and scalable compared

with broadcast routing for the IEEE 1609.4 standard. On

the other hand, STRAW mobility model has the highest

throughput, since this model determines a vehicle’s

trajectory at each intersection; namely, a vehicle will

make a turn at an intersection with a specified probability

that can be independently assigned to each vehicle.

Finally, we also evaluate the performance of safety

and non-safety application for different mobility on the

IEEE 1609.4 standard. We evaluate STRAW mobility

model has the best throughput and number of delivered

packets for non-safety application performance. On the

other hand, IDM_IM mobility model has the worst

average delay for non-safety application.

REFERENCE

[1] Jafari, A., Performance Evaluation of IEEE 802.11p for Vehicular

Communication Networks. PhD Dissertation, Faculty of Arts, Computing, Engineering, and Sciences, Postgraduate Program, Sheffield Hallam University, South Yorkshire, England, UK, 2011.

[2] IEEE Std 1609.4-2010, IEEE Standard for Wireless Access in Vehicular Environments (WAVE) Multichannel Operation, IEEE, 2011.

[3] Perdana, D. Sari, RF (2013), “Performance Evaluation of Multi-Channel Operation IEEE 1609.4 Based On Multi-hop Dissemination”, International Journal of Computer and Network Security (IJCSNS), VOL.13 No.3, March 2013, pp. 32-41.

[4] Ali J. Ghandour, Marco Di Felice, Hassan Artail and Luciano Bononi, “Modeling and Simulation of WAVE 1609.4-based Multi-channel Vehicular Ad Hoc Networks”, 5th ACM International Conference on Simulations Tools and Techniques (SIMUTools 2012), March 19-23, 2012, Sirmione-Desenzano, Italy.

[5] Yun-Wei Lin, Yuh-Shyan Chen, Sing-Ling Lee, Routing Protocols in Vehicular Ad Hoc Networks : A Survey and Future Perspectives, Journal of Information Science and Engineering 26, 913-932, 2010.

[6] Muhammad Alam, Muhammad Sher, Syed Afaq Husain, Integrated Mobility Model (IMM) for VANETs Simulation and Its Impact, International Conference on Emerging Technologies, 2009.

[7] VanetMobiSim project home page. [Online]. Available: http://vanet.eurecom.fr/

[8] Choffnes, R. D., Bustamante, E. Fabian, STRAW – An Integrated Mobility and Traffic Model for VANETs, Department of Computer Science, Northwestern University.

[9] Fiore, M., J. Härri, F. Filali and C. Bonnet. 2007. Vehicular Mobility simulation for VANETs. In Proceedings of the 40th Annual Simulation Symposium (ANSS ’07).

[10] F. Bai, N. Sadagopan, and A. Helmy, “The IMPORTANT framework for analyzing the Impact of Mobility on Performance of Routing Protocols for Adhoc Networks”, Elsevier Ad Hoc Networks vol. 1 (2003), pp. 383-403, 2003.

[11] The networksimulator-2,http://www.isi.edu/nsnam/ns/ [12] Wireless LAN Medium Access Control (MAC) and Physical

Layer (PHY) specifications Amandement 7 : Wireless Access in

Vehicular Enviroments, IEEE Unapproved Draft Std.

P802.11p/D8.0, Jul. 2009.

[13] Jagdeep Kaur and Er. Parminder Singh, Performance

Comparison Between Unicast and Multicast Protocols in

VANETs, International Journal of Advanced Technlogy and

Engineering Research (IJATER) Volume3, Issues 1, Jan 2013,

ISSN : 2250-3536.

[14] Jonathan C.L. Liu, Ph.D. Paper#4 : Overload Issues in 802.11b

Ad-Hoc Network. Lecture Slides. Department of Computer,

Information Science and Engineering (CISE), University of

Florida.

[15] Doan Perdana and Riri Fitri Sari, “Mobility Models Performance

Analysis using Random Dijkstra Algorithm and Doppler Effect

for IEEE 1609.4 Standard”, International Journal of Simulation,

Systems, Science, and Technology, United Kingdom Simulation

Society, Current status : Accepted.