Empirical Study of Mobility effect on IEEE 802.11 MAC ...worldcomp-proceedings.com/proc/p2011/ICW2169.pdfsome light to answer some of the questions related to the impact of the mobility

Empirical Study of Mobility effect on IEEE 802.11 MAC protocol for Mobile Ad-

Hoc Networks

Mojtaba Razfar and Jane Dong

mrazfar, [email protected]

Department of Electrical and computer Engineering

California State University Los Angeles

ABSTRACT

To design an efficient and effective MAC layer

protocol for Mobile Ad-Hoc networks is a challenging

task. IEEE 802.11 MAC protocol, which supports ad hoc

network mode, provides a good reference for research

work in this area. In the recent years, many researchers

have investigated the performance of IEEE 802.11 on

MANET both theoretically and empirically. However,

the impact of the mobility on MAC layer design has not

been evaluated thoroughly. In our research, we used

OPNET simulator to analyze the performance of IEEE

802.11 MAC protocol under various mobility patterns

for different network topologies. The findings revealed

interesting correlation between the speed of movement/

the mobility pattern and key network performance

parameters including delay and throughput. We also

investigated the impact of mobility on fairness issues in

media access. The empirical study presented in this

paper will be useful to enhance the MAC design of

MANET with median or high mobility nodes.

Keywords: MANET (Mobile Ad-Hoc Networks), Mobility,

(Medium Access Control) MAC, IEEE 802.11, OPNET

1. INTRODUCTION

Mobile Ad Hoc Networks (MANET) are becoming more

and more popular due to its ability to offer convenient,

flexible and low cost network service for many non-

traditional applications. Unlike the widely used Wi-Fi

network which relies on the access point to attach to the

existing networking infrastructure, MANET is

infrastructure-less, where each node acts as a sender,

receiver, and router. While the freedom to deploy a mobile

ad hoc network at anytime anywhere is very attractive, to

make such network function properly presents a lot of

technical challenges. For MAC layer protocols, the well-

known challenges are imposed by hidden and exposed

terminal problems, fairness access issues, limited

bandwidth, limited power supply, as well as limited

transmission range and mobility.

IEEE802.11was primarily designed for WLAN, but it

also supports ad hoc network mode. The MAC layer

protocol in IEEE 802.11 laid the foundation for many

proposed MAC layer protocols for MANET. Therefore, it is

worthwhile to evaluate the performance of IEEE 802.11 on

MANET to see how to improve the design. Many existing

research [1-3] focused on the effectiveness on handling

hidden and exposed terminal problem, and some addressed

fairness issues. Mobility, although an important design

factor of MANET, its impact on MAC layer performance

has not been fully analyzed yet. Most of the current

researches that investigated the mobility effect are focused

on the network layer since it is a major concern in routing

[4-5]. In [6], the authors briefly compared the performance

IEEE 802.11 and other MAC protocols under network

scenarios with mobility. However, to develop a full

understanding of the mobility effect on MAC layer

performance including delay and throughput, fairness,

collision probability, a more comprehensive and in-depth

study is necessary. The objective of our research is to

conduct such study using OPNET [7] to show how mobility

impacts the key MAC layer performance parameters.

In this paper, we will present our findings of the empirical

study using OPNET simulation. Due to the nice property of

OPNET, it is possible to set up different network scenarios

with different mobility patterns, which allowed us to better

study the impact of various factors including speed,

transmission range, and moving trajectory. The network

parameters that were taken into account in our study

includes delay, throughput, collision count, overhead of

control traffic, and backoff time. The documented results

will be useful to enhance the MAC design of MANET with

different mobility.

The paper is organized as follows. Section 2 provides a

brief overview of IEEE 802.11 and highlights the important

design issues. The empirical study exploring the mobility

effect using the OPNET software is presented in section 3.

Experimental results are described in this section as well.

Finally, we will conclude our findings in Section 4.

2. OVERVIEW OF IEEE 802.11 MAC PROTOCOL

IEEE 802.11 MAC layer protocol is referred to as

Distributed Coordination function “DCF” which was based

on virtual carrier sensing and the physical carrier sensing

[8]. IEEE 802.11 DCF uses RTS/CTS/DATA/ACK when

the size of the data frame is large enough; it may just use

carrier sense or it may use both methods referred to as

CSMA/CA with RTS/CTS as a MAC protocol. The three

major issues related to MAC layer protocol over MANET

are the ability to handle hidden and exposed terminal

problems, the ability to ensure fair access of multiple

stations, and the ability to cope with mobility.

2.1. Hidden and Exposed Terminal Problems

These two problems have become a major issue in

MANET. Hidden terminal problem [3] occurs when two

stations are out of the range of each other and trying to send

to the same receiver. As a result, the effect significantly

decreases the throughput and makes the delay longer.

Exposed terminal problem, on the other hand, is when a

node is blocked from transmission to the other stations due

to the transmission of the adjacent node. This will cause

collision and bandwidth waste (less spatial reuse) and will

bring up the starvation problem of the unlucky node. IEEE

802.11 DCF proposed the RTS/CTS handshaking method in

order to alleviate the negative impact of these issues on the

whole network. In the literature, several schemes have been

proposed to solve the hidden and exposed terminal problems

using different mechanism. In [9], the authors explored the

IEEE 802.11 MAC protocol with and without using

RTS/CTS handshaking method. The total WLAN

retransmissions, data traffic sent/received, WLAN Delay

factors of the whole network was investigated using both

methods. They demonstrated that in the scenarios that the

Hidden terminal problem exists, it will be a good idea to use

this option as it decreases the delay of the network

dramatically. They also mentioned that this handshaking

method is not necessary to be used where the hidden nodes

are not present due to the overhead that it adds to the

network. The mobility factor was investigated when the

hidden nodes exist. However, the speed of the nodes and the

location of them have not been studied in this paper.

MACA [10] on the other hand, did not use the carrier

sensing option and instead, it used the RTS/CTS/DATA

handshake to reserve and use the channel. Although this

protocol was a simple design, the control channel collisions

made the scheme not effective in the MAC layer. Moreover,

in these papers, they never spoke of the effect of the

Mobility of the performance of the whole network.

2.2. Fairness issue

Another important factor that should be considered in

designing MAC protocols is to make sure that all nodes

have fair access to the channel in order to transmit their

data. So far most of the single channel MAC schemes rely

on the back-off procedure. Upon collision, the mobile nodes

will go through the back-off procedure and will try to

retransmit after a certain amount of time. Because the

backoff time is different for different nodes, some nodes

may have more chance to transmit than the others and they

are favored in data transmission. This will result in starving

problem of the unlucky nodes with long contention window

size. Therefore, designing good strategies for back-off

procedures and providing fair chances among nodes to

access the channel is one of the important aspects in

MANET. MACAW [11] for wireless LANs is another

single channel schemes which tried to improve the

performance of MACA protocol. A five handshake

RTS/CTS/DS/DATA/ACK has been used in this protocol

which leads to alleviation of the hidden and exposed

terminal problem and better fairness among nodes. By using

a different back-off approach (MILD), this protocol allowed

the nodes to access the channel in a fair manner which is

more desirable in ad hoc networks. However, the effect of

mobility on fairness issue using this protocol has not been

investigated.

2.3. Mobility issue

For the infrastructure-based networks, the access point has

the major influence on the delivery of the data to the

destination. Within a Basic Service Set (BSS), the stations

have to share information using the access point and

therefore their position towards each other is not that

important. Hence, the mobility of nodes does not have a

major effect on the MAC layer protocol [9]. For

infrastructure-less network as MANET, the mobile nodes

are in direct contact with each other. Since they can be

sender, receiver and router, mobility has a significant impact

on the performance of their data delivery. One may wonder

what influence may the mobility of the nodes cause on the

performance of the network. Will mobile nodes be treated

the same way as they move? Will the efficiency of the

transmission stays the same as the mobility varies? How

will the delay and overall throughput be affected via

different mobility pattern? Can we enhance the network

performance using the handshaking RTS/CTS method under

high mobility? Most of the questions do not have a solid

answer yet. In our research, we will explore the relationship

between mobility and all these factors using OPENT

simulation. The results presented in this paper will shed

some light to answer some of the questions related to the

impact of the mobility on MANET networks.

3. EMPIRICAL STUDY USING OPNET

3.1. OPNET Simulator

OPNET modeler is one of the powerful simulation

software allowing the users to implement different network

topologies using a friendly graphic user interface. As lots of

research papers in networking field used NS-2 simulator

[12], OPNET makes it easier to use as it provides ready-to-

use components without the need of writing codes to create

real time network simulations. It also provides the flexibility

for advanced users to create their own network node and

link by hard coding. For our research, OPNET is selected

since its Wireless Modeler includes a rich library of detailed

mobile protocols and application models that can be utilized

to create MANET with various mobility patterns.

3.1.1. MANET and Mobility in OPNET

OPNET [1] uses the IEEE 802.11 MAC protocol with

DCF for Mobile Ad-Hoc Networks. RTS/CTS handshaking

option is also included in case a user decides to implement

it. The software has different objects for MANET networks

such as the MANET station, MANET work station, and

Mobility configuration options in order to set up the

movement of the nodes. In fact, Mobility is one of the most

valuable options that are included in the simulator so that

the users can easily define the way the stations move. The

speed of the stations can also be easily defined for various

applications. This option makes it simpler in real time

simulations in comparison to the other simulators where the

mobility is a difficult task to define and implement. Figure 1

illustrates a MANET scenario with pre-defined node

mobility. The statistics that are related to this work are

explained briefly as follows:

1) MANET delay: the end to end delay of MANET

packets for the whole network (seconds).

2) Throughput: the total number MANET traffic

which is received in bits per second by all the

MANET receivers.

3) Media Access delay: The global statistic for the

total of queuing and contention delays of the data,

management, delayed block-ACK and Block-ACK

frames transmitted by all WLAN MACs in the

network (seconds).

4) Back off slots: the number of slots that a stations

needs to back off before transmission while

contenting for the medium, and the number of slots

in the contention window after the successful

transmission of the station.

5) Retransmission attempts: the total number of

retransmissions by all the WLAN MACs in the

whole network until the delivery of the packet or

being discarded as a result of reaching the short or

long retry limits. We used this factor to study the

impact of mobility on the collision counts as well

as the effectiveness of RTS/CTS handshake.

Fig.1. Mobility of the Mobile Ad-Hoc Networks

3.1.2. Simulation environment

In our study, two different network topologies were

created and analyzed to evaluate the mobility effect on the

node’s behavior and the network performance. Different

settings have been applied to the two topologies based on

the needs of the network simulations.

In the first topology, the relationship between mobility,

transmission range and the overall throughput and delay of

the network is investigated. As Figure 2 illustrates, one

subnet consists of eight nodes around each other. Another

single node is approaching this sub network with a constant

speed. To study the impact of mobility among the nodes

inside the sub network, different scenarios were created to

compare the delay and throughput where the nodes are

either static or moving randomly. We also changed the

speed of the nodes in different steps to see the influence of

this factor on the network. To evaluate the effect of the

transmission range, we also varied the transmission range in

different scenarios according to the distance and the area

that were used in the simulation. In addition, the mobility

impact on this network with different traffic loads was also

studied.

Table 1 shows the setting used for the first topology.

Fig.2. First topology where a node is approaching a static network

Attribute Value

Transmission power (W) Varies per scenario

Data Rate (bps) 11 Mbps

Physical Layer Method Direct Sequence

Buffer Size (bits) 256000

Packet Size (bits) Exponential (1024)

Traffic generated per node Varies per scenario

Node’s Speed Varies per scenario

Nodes movement method Defined/Vector trajectory

Simulation time (min) 60

Table.1. Topology 1 configurations

For the second topology, two similar subnets are created

and each consists of 7 nodes. One of the nodes is static and

it transmits to the other static node in the second subnet. The

other nodes inside the subnet are either static or moving

while trying to transmit data to the static station inside their

subnet. The two subnets are moving towards each other with

a constant speed. The internal nodes are located with

different distances from the static receiver in order to study

the effect of different movement trajectories on the fairness

among nodes. Note that the internal nodes are in the

transmission range of each other. That is, each subnet allows

their nodes to transmit inside the region of the subnet. The

transmission range for the static receiver is higher than the

others due to the fact that the receiver will need to transmit

to the other static receiver located at the second subnet. In

this topology, we not only look into the delay and

throughput factors, but also check the fairness among nodes

and the effect of RTS/CTS. Table 2 shows the setting we

used for the second topology.

Fig.3. Second topology where the mobile nodes are moving around the

static receiver inside the two subnets

Attribute Value

Transmission power of the

static receiver (W)

0.001

Mobile Nodes Transmission

power (W)

0.0003

Data Rate (bps) 11 Mbps

Physical Layer Method Direct Sequence

Buffer Size (bits) 256000

Packet Size (bits) Varies per scenario

Traffic generated per node Varies per scenario

Internal Node’s Speed (m/s) 0.2

Subnet speed (m/s) 1

Nodes movement method Defined trajectory

Simulation time (min) 30

Table.2. Topology 2 configurations

3.2. Experimental results

3.2.1. Impact of mobility on delay and throughput

A) The impact of mobility with lower transmission

range

To evaluate the impact of mobility with lower

transmission range, three scenarios were created under the

first topology (figure 2). The transmission range and traffic

load for these scenarios are the same, while the mobility

inside the subnet is different:

1) Scenario 1: inside nodes are static

2) Scenario 2: inside nodes move with speed 0.2m/s

3) Scenario 3: inside nodes move with speed 1m/s

For all these scenarios, the single node in approaching the

sub-network with a constant speed 0.2m/s. Table 3 shows

the configuration of transmission power and traffic load of

these scenarios. The Domain which covers an area of 120 ×

120 square meter allows the nodes to move inside this

region. A lower transmission range is defined so that the

nodes can sense each other at a maximum of 80 meters

distance. That is, the nodes will not be able to sense each

other at some parts of the Domain. This will allow us to see

the effect of lower transmission region on the networks

using the mobility feature.

Attribute Value

Transmission power (W) 3E-005


External Node’s Speed (m/s) 0.2

Traffic generated per node Exponential (0.1)

Table 3: Common parameters

Fig.4. Comparison of average Delay and Throughput for static and moving

inside nodes with speed 0.2m/s with lower transmission range

Figure 4 compares the results for scenario 1(static) and 2

(node moving with low speed 0.2/m). Results show that

when the nodes are not moving inside the domain, the

network has higher throughput and lower delay. This seems

be to due to the fact that when the nodes move around, the

transmission range decreases and the connection

establishment among nodes becomes weaker. Hence, the

overall throughput decreases coming up with higher delay.

Figure 5 compares the results for scenarios 2 (low

movement speed) and 3 (high movement speed. Results

demonstrate a higher delay and lower traffic being received

as a result of an increase in the speed of the mobile nodes.

Higher speed will make the node to move further from the

receiver in a shorter amount of time and therefore, less

chance to deliver their data to the destination. Higher delay

is due to the fact the nodes are having more problem in

delivering their data to the receiver and experiencing a

higher back off time and retransmissions of the data. It also

contributes to the internal collision or packet loss which

prevents the delivery of the data.

Fig.5. Comparison of average Delay and Throughput for the networks with

low movement speed (0.2m/s) and high movement speed (1 m/s) with low

transmission range.

B) The impact of mobility with higher transmission

range

Attribute Value

Transmission power (W) 0.0001


Internal Node’s Speed (m/s) 1



In this case, we increased the transmission range (0.0001

W) so that the range covers the whole area of the movement.

We also increased the Traffic generated by each node to see

how the mobile stations behave while generating more

traffic. Figure 6 shows the comparison results for the two

network scenario with static inside nodes and moving inside

notes (speed 0.2m/s). It is interesting to see that in this case,

the mobility will have a positive impact which leads to a

little higher throughput and lower delay for the entire

network.


inside nodes with speed 0.2m/s with higher transmission range

We also found out that increasing the speed of a network

with high transmission range will slightly improve the

performance of the network in case of throughput and delay.

The reason may be due to the fact that all nodes are within

the transmission range of each other no matter how they

move. Therefore, the random movement pattern of the

inside nodes may lead to a more even distribution of the

nodes that helps with channel access.

C) Impact of group mobility on delay and

throughput

Starting from this subsection, we will describe our

findings on the impact of group mobility pattern. The

simulations were created using the second topology as

discussed earlier (Figure 3). The two subnets are moving

towards each other with a speed of 1 m/s. The internal nodes

inside each subnet are moving with the speed of 0.2 m/s.

The nodes have different distances to the fixed receiver. In

the case where the nodes are moving, they move around the

receiver based on their location and distance with regards to

the receiver. Besides the delay and throughput factors, the

fairness among nodes and the effect of RTS/CST method is

investigated in the following section.

Attribute Value






In this scenario, we investigated the effect of mobility on

a network with stations generating a relatively larger traffic

inside the network. We also increased the packet size per

station. Similar to the first topology, results demonstrate a

better performance of the network with mobility in

comparison to the static one when the transmission ranges

covers the movement paths.


inside nodes with speed 0.2m/s for group mobility

3.2.2. Impact of mobility on Fairness

A) The fairness issue without RTS/CTS

To evaluate the fairness of IEEE 802.11 MAC layer

protocol for mobile network, we used the back off slot time

which is the number of slots that a stations needs to back off

before transmission while contenting for the medium, and

the contention window size after the successful transmission

of the station as the measurements. . Moreover, the

retransmission attempt is a good factor to analyze the

delivery efficiency of the packets per node when the stations

are moving around the receiver. This factor reflects the

impact of both the internal collision and the transmission

errors including the loss of acknowledgment or an error

occurred in the packet. The effect of the amount of traffic

generated by each node on the fairness issues has also been

investigated in this part. Node 3 and Node 6 are selected for

our results. As mentioned before, the reason for this

selection is the distance difference between the nodes and

the receiver on their moving paths. Therefore, we can

clearly see the effect of the different distances caused by

mobility on the fairness among these nodes.

Fig.8. Average Back off slot time and retransmission attempts for the two

selected nodes with low traffic load

Figures 8 and 9 present our simulation results for

CSMA/CA without RTS/CTS. For the low load generated

by each station, the two nodes have almost the same back

off slot time (as shown in Figure 8) demonstrating that they

have the same chance to access the channel. On the other

hand, the retransmission attempts per node are much more

less for the closer node to the receiver (Node 3). This might

be due to the internal collision or the error inside the packets

resulting in the failure of the delivery of the packets.

We repeated the same procedure but changing the amount

of traffic generated by each station to a higher level

(exponential (0.0008)). We also increase the packet size up

to eight times (exponential (8192)).

Fig.9. Average Back off slot time and retransmission attempts for the two

selected nodes with higher traffic

As shown in Figure 9, the results illustrate that the back

off slot time has increased dramatically for both nodes and

that the difference becomes obvious as the distance to the

receiver increases. This can be due to higher collision and

more competition for accessing the channel resulting in

higher back off’s and retransmission attempts for both

nodes. The closer the node is to the receiver, the higher

chance it has to access the channel while moving around the

receiver.

B) The fairness issue using RTS/CTS

Attribute Value





RTS threshold (bytes) 256

Internal Node’s Speed (m/s) 1

Subnet speed (m/s) 1


In this case, we added the RTS/CTS option to each node

to see the efficiency of this method on the network. Low

traffic has been used in this scenario. From Figure 10, we

can see that the retransmission attempts have been decreased

for Node 3, which demonstrated the effectiveness of the

handshaking method. Same results occurred for other nodes

inside the network showing the good efficiency of the

handshaking method.

Fig.10. Comparison of the average retransmission attempts for

Node 3 using CSMA/CA only (RED); and using CSMA/CA

with RTS/CTS (BLUE)

C) The network performance using RTS/CTS

We also studied the performance of whole network using

the RTS/CTS access mechanism. The same configurations

were used as the above simulation. Results depict that the

delay of the whole network decreases due to the prevention

of the collisions and allowing the nodes to have their data

delivered in a shorter amount of time.

Fig.11. Comparison of average network delay: Red plot—

without RTS/CTS; Blue plot-- with RTS/CTS

4. CONCLUSION

In this paper, the performance of the Mobile Ad-hoc

networks is investigated using the IEEE 802.11 MAC

protocol. We have shown that Mobility can affect the

network based on different factors. We studied the effect of

varying the speed of the nodes, and their location on the

network. We have also studied the fairness and the effect of

the RTS/CTS handshaking process on the performance of

the nodes inside the network. Our results show that the

performance of the network varies as the mobile nodes

move inside the network. We illustrated that the

performance of the network improves as the traffic increases

when a sufficient transmission ranges of nodes is provided.

We have also shown that Mobility will cause the nodes to

have longer back off times and retransmission attempts in

order to deliver their information to the destination. The

RTS/CTS handshaking method demonstrated its efficiency

on the mobile nodes when the number of collisions becomes

more and more.

5. REFERENCES

[1] K. Xu, M. Gerla, and S. Bae, “Effectiveness of RTS/CTS

handshake in IEEE 802.11 based ad hoc networks,” Ad Hoc

Networks, Elsevier, vol. 1, no. 1, pp. 107-123, 2003.

[2] T. S. Ho and K. C. Chen, “Performance Analysis of IEEE

802.11 CSMA/CA Medium AccessControl Protocol,” in Proc.

IEEE PIMRC ’96, pp. 407-411, 1996.

[3] Khurana, S.; Kahol, A.; Jayasumana, A.P, Effect of Hidden

Terminals on the Performance of IEEE 802.11 MAC Protocol,”

LCN ’98 proceedings, pp 12-20, 1998.

[4] Samir R. Das , Robert Castañeda, Jiangtao Yan , Rimli

Sengupta, Comparative performance evaluation of routing

protocols for mobile, ad hoc networks. In 7th Int. Conf. on

Computer Communications and Networks (IC3N), pages 153–161,

October 1998.

[5] Das, S.R, Perkins C.E, Royer, E.M., Performance Comparison

of Two On-Demand Routing Protocols for Ad Hoc Networks in

IEEE Proceedings, pp 3 – 12, 2000.

[6] Jagadeesan S, Manoj BS, Murthy CSR. Interleaved carrier

sense multiple access: an efficient MAC protocol for ad hoc

wireless networks. Proceedings of IEEE ICC’03, May 2003.

[7] Online Documentation, “OPNET Modeler,”

http://www.opnet.com/, Date visited: January 2011.

[8] IEEE 802.11 Working Group, Wireless LAN Medium Access

Control (MAC) and Physical Layer (PHY) Specification, 1997.

[9] H. Jasani, N. Alaraje, Evaluating the Performance of IEEE

802.11 Network using RTS/CTS Mechanism,” in IEEE EIT 2007

Proceedings.

[10] P. Karn, MACA—a new channel access method for packet

radio, in: Proceedings of the ARRL/CRRL Amateur Radio 9th

Computer Networking Conference September 22, 1990.

[11] V. Bhargavan, A. Demers, S. Shenker, L. Zhang, MACAW—

A Media Access protocol for wireless Lans, in: Proceedings of the

ACM SIGCOMM, 1994, pp. 212–225

[12] Online Documentation, “The Network Simulator - ns-2,”

http://www.isi.edu/nsnam/ns/, Date visited: January 2011.

Empirical Study of Mobility effect on IEEE 802.11 MAC ...worldcomp-proceedings.com/proc/p2011/ICW2169.pdfsome light to answer some of the questions related to the impact of the mobility

Documents