Available via license: CC BY 4.0
Content may be subject to copyright.
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
DOI : 10.5121/ijasuc.2012.3408 73
PERFORMANCE COMPARISON OF
VARIOUS ROUTING PROTOCOLS IN
DIFFERENT MOBILITY MODELS
Neha Rani
1
,
Preeti Sharma
2
and Pankaj Sharma
3
1
Department of Computer Science,
A.B.E.S. Engineering college ,Ghaziabad, U.P,
India
(nehachandel83@gmail.com)
2
Department of Computer Science,
A.B.E.S. Engineering college ,Ghaziabad, U.P,
India
(Preeti.sharma860@gmail.com)
3
Department of Computer Science, Asst. Professor,
A.B.E.S. Engineering college,
Ghaziabad, U.P, India
(sharma1.pk@gmail.com)
A
BSTRACT
Mobile Ad hoc Network (MANET) is a infrastructure less network in which two or more devices have
wireless communication which can communicate with each other and exchange information without need
of any centralized administrator. Each node in the ad hoc network acts as a router, forwarding data
packets for other nodes. The main issue is to compare the existing routing protocol and finding the best
one. The scope of this study is to test routing performance of three different routing protocols (AODV,
OLSR and DSDV) with respect to various mobility models using NS2 simulator. In this paper the
parameters used for comparison are packet delivery fraction (PDF), average end to end delay (AEED),
normalized routing load (NRL) and throughput.
K
EYWORDS
NS-2 simulator, Performance parameters, AODV, OLSR, DSDV, RPGM, RWPM.
1.
I
NTRODUCTION
A Mobile Ad-Hoc Network is a collection of mobile nodes with no pre-established
infrastructure, self organizing wireless network which forms a temporary network [1].
Figure 1: Infrastructured and ad-hoc networks.
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
74
Each of the nodes has a wireless interface and communicates each other over either radio or
infrared signals.
In ad hoc networks [2] all nodes are mobile and can be connected dynamically
in an arbitrary manner. One area of research, which has been a focal point of research in Ad hoc
networks, is Routing. Generally, Adhoc routing protocols can be classified broadly into two
categories, these are Proactive, Reactive. A brief classification of Ad-hoc routing protocols is
given in Figure 2.
Table-Driven routing protocols (Proactive)
In table driven/ proactive routing protocols
[4], nodes periodically exchange routing information and attempt to keep up-to-date routing
information [5]. Proactive protocols are called so because they have to maintain information
about routing prior to its use. Information about routes is maintained in tables called routing
tables and when topology changes these tables are updated.
On Demand routing protocols (Reactive) I
n on demand/ reactive routing protocols [6],
nodes only try to find a route to a destination when it is actually needed for communication.
Proactive protocols are named so, because routing table [7] is not maintained in it. When
communication between nodes is required then route is discovered in on-demand manner.
Many routing protocols have been proposed, but just few comparison studies have been
performed. Almost all available comparative studies have performed simulations for proactive
and reactive protocols by varying number of nodes, network topology, and network density.
This paper focuses on varying mobility speed from low to high by keeping other parameters like
pause time, number of nodes, mobile connections, simulation duration constant. This study is
performed by varying mobility speed and measuring different quantitative metrics in different
mobility models i.e. RPGM and RWPM. Reference Point Group Mobility (RPGM) model
shows the random motion of mobile nodes associated with a group [3]. Random Waypoint
Model (RWPM) assumes that each host is initially placed at random position within the
simulation area. As the simulation progresses, each host possess at its current location for a
determinable period called the pause time [3].
The remainder of the paper is organized as follows. After presenting the related work in section
2, section 3 presents routing in MANET. Section 4 presents mobility models. Section 5
describes simulation environment. The results of our simulation are analyzed in section 6.
Finally, section 7 concludes the paper.
Figure 2: Classification of Routing Protocols in MANET
AODV
DSR
LMR
DYMO
Table
-
driven
OLSR
DSD
V
WRP
TORA
On
-
Demand
Adhoc routing
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
75
2.
R
ELATED
W
ORK
Several researchers have done the quantitative and qualitative analysis of Ad hoc Routing
Protocols by means of different performance parameters. Also they have used different
simulators for this purpose.
1. Nilesh P.Bopade, Niket N.Mhala performed simulations for comparison of Proactive and
Reactive protocols. They have used NS2 Simulator. They have used varying number of mobile
nodes but other factors like pause time, speed were not taken into consideration.
2. J Broch et al., performed experimental performance comparison of both Proactive and
Reactive routing protocols. In their NS-2 simulation, a network density of 50 nodes with
varying pause times and various movement patterns were chosen.
3. Jorg D.O. [3] studied the behaviour of different routing protocols for the changes of network
topology which resulting from link breaks, node movement, etc. This paper has focussed on
performance evaluation by changing number of nodes. But he did not investigate the
performance of protocols under high mobility, large number of traffic sources and larger
number of nodes in the network which may lead to congestion situations.
4. Arunkumar B R et al. Authors perform simulations by using NS-2 simulator. Their studies
have shown that reactive protocols perform better than table driven (proactive) protocols.
5. N Vetrivelan & Dr. A V Reddy analyzed the performance differentials using varying network
density and simulation times. They performed two simulation experiments for 10 & 25 nodes
with simulation time up to 100 sec.
3. ROUTING IN MANET
This section explains the rouing methodology of various routing protocols taken in comparative
study
.
3.1. Destination sequenced distance vector (DSDV) protocol
One of the examples of proactive protocol is DSDV. This protocol adds a new attribute,
sequence number, to each route table entry at each node. Each node maintains a routing table at
its own and which helps in packet transmission.
3.1.1. Protocol overview and working
For the transmission of packets each node maintains routing table. The routing table maintained
by each node also contains the information for the connectivity to different stations in the
network. These stations shows all the available destinations and the number of stations (hops)
required to reach each destination in the routing table. The routing entry is tagged with a
sequence number which is originated by the destination station. Each station transmits and
updates its routing table periodically. The packets being broadcasted between stations indicate a
list of accessible stations and number of nodes required to reach that particular station. Routing
information is advertised periodically by broadcasting or multicasting the packets. In DSDV
protocol each mobile station in the network must constantly advertise its routing table to each of
its neighbouring stations. As the information in the table may vary frequently, thus the
advertisement should be done on the continuous basis so that every node can locate its
neighbours in the network. It ensures the shortest number of stations (hops) required from
source station to a destination station.
The data broadcast by each node will contain its new sequence number and the following
information for each new route:
– The destination address
– The number of hops required to reach the destination and
– The new sequence number, originally stamped by the destination
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
76
After receiving the route information, receiving node increments the metric and broadcasts it.
Once the mobile nodes receive the route information they broadcast it on an immediate basis.
As the mobile nodes change their position within the network results in breaking their links.
These broken links may be detected by the layer2 protocol. When there is a broken link in a
network, then immediately that metric is assigned an infinity metric determining that there is no
hop and the sequence number is updated. Sequence numbers for infinity metrics are odd
numbers and the sequence numbers originating from the mobile hosts are defined to be
even number. The broadcasting is done in two ways: full dump and incremental dump. Full
dump broadcasting will carry all the routing information and requires multiple network protocol
data unit (NPDU) while the incremental dump will carry only information that has changed
since last full dump and requires only one NPDU to fit in all the information .When an
information packet is received from another node, In the first step, it compares the sequence
number of the Node with the available sequence number for that entry. If the number of the
sequence is larger than the previous one then it will update the routing information with the new
sequence number else if the information arrives at node with the same sequence number it looks
for the another metric entry and if the number of hops is less than the previous entry the new
information is updated at the node (if information is same or metric is more then it will discard
the information). While the nodes information is being updated the metric is increased by 1 and
the sequence number is also increased by 2. Similarly, if a new node enters the network area, it
will announce itself in the network and the nodes in the network update their routing table.
During the process of broadcasting, the mobile hosts transmits their routing tables periodically
but due to the continuous movements by the hosts in the networks, this will lead to continuous
burst of new routes transmissions upon every new sequence number from that destination. The
probable solution for this problem is to delay the advertisement of such routes until it shows up
a better metric. The Address stored in the routing table corresponds to the layer at which the
DSDV protocol is operated [8].
3.2. Opitmized link state routing (OLSR) protocol
OLSR is an optimization version of a pure link state protocol. Whenever there is any change in
the topology then information is flooded to all nodes. This causes overheads and such overheads
are reduced by Multipoint relays (MPR). Two types of control messages are used in OLSR they
are topology control and hello messages. There is also Multiple Interface Declaration (MID)
messages which are used for informing other host that the announcing host can have multiple
OLSR interface addresses [9]. The MID message is broadcasted throughout the entire network
only by MPRs. There is also a “Host and Network Association” (HNA) message which provides
the external
routing information by giving the possibility for routing to the external addresses.
Routing in OLSR is described as follows.
3.2.1. Neighbour Sensing
The link in the ad hoc network can be either unidirectional or bidirectional so the host must
know this information about the neighbors. The Hello messages are broadcasted periodically for
the neighbor sensing. Only nearby neighbour receives hello messages.
3.2.2. Multipoint Relays
Overheads are reduced with the help of MPRs .Instead of pure flooding the OLSR uses MPR to
reduce the number of the host which broadcasts the information throughout the network [10].
3.2.3. Multipoint Relays Selection
Proposed algorithm for selecting Multipoint Relay set:
1. All neighbour which want to become MPR are taken.
2. For every neighbour a host degree is calculated. Host degree is actually the number of
neighbours whose distance from source is two hops.
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
77
3. Then neighbour is added to MPR set .If it is the only neighbour from which is possible to get
to the specific two hop neighbour, and then remove the chosen host neighbours from the two
hop neighbour set.
4. If there are still some hosts in the two hop neighbour set, then calculate the reach ability of
the each one hop neighbour, meaning the number of the two hop neighbours, that are yet
uncovered by MPR set.
3.2.4. Routing Table Calculations
The host maintains the routing table, the routing table entries have following information:
destination address, next address, number of hops to the destination and local interface address.
3.3. Ad Hoc On-Demand Distance Vector Routing (AODV)
The Ad Hoc On-Demand Distance Vector routing protocol (AODV) is designed to improve the
performance of the Destination-Sequenced Distance Vector routing protocol (DSDV). The main
goal of AODV is to broadcast discovery packet when necessary and to distinguish between
local connectivity and topology maintenance. In AODV [11] overhead is reduced as number of
broadcast is minimized.
3.3.1. Path Discovery Process
The process of path discovery starts when a node needs communication with other node by
sending route request packet (RREQ) packet [11] which contain broadcast id, source address,
destination address, source sequence number, destination sequence number, hop count. When an
intermediate node receives RREQ it checks that it had received over bi-directional link. If this
has already processed then RREQ packet is discarded. Otherwise, it checks for route entry for
destination. The reply is send to source only if the destination sequence number in RREQ is
greater than destination sequence number in its route table. A route reply packet (RREP) [11]is
send by intermediate node as a response to RREQ packet. As RREP travels back to source all
information are updated. Finally, RREP reaches source and route entry is modified.
3.3.2. Maintaining Routes
In AODV [12] each node maintains a routing table with its entries. An active route entry is one
in which is in use by active neighbours. Path which is followed by packets from source to
destination with active route entries is called an active path. To transmit data from source to
destination each time route entry is used.
Figure 3: Source node S initiates the
path discovery process
Figure 4: A RREP packet is sent back
to the source
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
78
4. MOBILITY MODELS
A model that depicts the movement of mobile nodes, and changes in their velocity and
acceleration over time is called Mobility model. Basic parameters related to node movement are
number of nodes, mobility speed, pause time, sending rate, number of connections, simulation
duration. Mobility models can be classified in to two types group and entity models. In entity
models, the motion of mobile nodes are independent from each other, while in group models the
movements of nodes are dependent on each other or on some predefined leader node [13].
4.1. Reference Point Group Mobility (RPGM)
It is a group mobility model that shows the random motion of mobile nodes. In this model nodes
are dependent on some predefined leader node that determines the group motion behaviour.
4.2. Random Waypoint Mobility (RWPM)
It is an entity model, in which a node can choose any random velocity and any random
destination. The node starts moving towards the selected destination. After reaching the
destination, the node stops for a small duration defined by the “Pause Time” parameter and it
repeats the complete process again until the simulation process ends.
5. SIMULATION ENVIRONMENT
5.1. Simulation Model
Here we perform the experiments for the evaluation of the performance of Ad Hoc routing
protocol AODV, DSDV, OLSR with varying the mobility speed. We have 30 simulation run in
total out of which 15 trace files has been generated for RPGM and RWPM each. We tested all
performance metrics in our experiment under varying mobility speed of node (10 to 50m/sec)
and while other parameters (nodes = 20, data sending rate = 5 pkts/sec and no. of connections =
10, simulation duration=100sec, pause time=null) are constant.
5.2. NS-2 simulator
The network simulations have been done using network simulator NS-2 [14]. The network
simulator NS-2 is discrete event simulation software for network simulations. It simulates
events such as receiving, sending, dropping and forwarding packets. The ns-allinone-2.34
supports simulation for routing protocols for ad hoc wireless networks such as AODV, DSDV
and DSR. NS-2 is written in C++ programming language with Object Tool Common Language.
Although NS-2. 34 can be built on different platforms, for this paper, we chose a Linux platform
i.e. FEDORA 13, as Linux offers a number of programming development tools that can be used
with the simulation process. To run a simulation with NS-2.34, the user must write the OTCL
simulation script. The performance parameters are graphically visualized in GRAPH. Moreover,
NS-2 also offers a visual representation of the simulated network by tracing nodes events and
movements and writing them in a file called as Network animator or NAM file.
5.3. Simulation Parameters
This paper considers a network of nodes placing within a 2000m X 2000m area. The
performance of AODV, OLSR and DSDV is evaluated by varying the node speed and keeping
the other parameters such as number of nodes, transmission rate, pause time, simulation
duration constant. Table 1 depicts the simulation parameters used in this evaluation. All
performance metric are checked under the varying nodes speed from 10 to 50 m/sec in RPGM
and RWPM mobility model.
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
79
SIMULATION
PARAMETERS
Simulators NS-2.34
Protocols
AODV,OLSR ,DSDV
Simulation duration
100 sec
Simulation area
2000m X 2000m
Pause time Null
Movement model RPGM,RWPM
MAC layer protocol IEEE 802.11
Traffic type CBR
Data payload
51
2 bytes
5.4. Performance Metrics
While analyzed the AODV, DSDV and OLSR, protocols, we focused on four performance
metrics for evaluation which are Packet Delivery Fraction (PDF), Average End-to-End Delay,
Normalized, Routing Load (NRL) and Throughput.
5.4.1. Packet delivery fraction
Packet delivery fraction (PDF) is the ratio of number of received data packets successfully at the
destinations over the number of data packets sent by the CBR sources.
5.4.2. Average End to end delay
It is the average time from the transmission of a data packet at a source node until packet
delivery to a destination which includes all possible delays caused by buffering during route
discovery process, retransmission delays, queuing at the interface queue, propagation and
transfer times of data packets.
5.4.3. Normalized Routing Load
The normalized routing load (NRL) it is the ratio of all routing control packets send by all nodes
to number of received data packets at the destination nodes.
5.4.4. Throughput
It is the average number of messages successfully delivered per unit time number of bits
delivered per second.
Table 1: Simulation table
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
80
6. SIMULATION RESULTS AND ANALYSIS
The results after simulation are viewed in the form of bar graphs. The performance of AODV,
OLSR and DSDV based on the varying the mobility speed is done on parameters like packet
delivery fraction, average end-to-end delay, normalized routing load and throughput.
6.1. Packet Delivery Fraction
Figure 5 and Figure 6 shows that the group model RPGM is superior compared to entity model
i.e. Random Waypoint. This happens because it is a group mobility model and the whole
communication process occurs between a few groups. When node speed is 10m/sec, Figure 5
depicts that AODV performed better compared to OLSR and DSDV. But as mobility speed goes
up the delivery ratio of routing protocols also went down.
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
10 20 30 40 50
P
D
F
(
%
)
MOBILE NODE SPEED (m/sec)
RPGM
AODV
OLSR
DSDV
Figure 5: Mobility speed Vs Packet delivery fraction in RPGM Model
Figure 6: Mobility speed Vs Packet delivery fraction in RWPM Model
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
81
6.2. Average E2E Delay
This experiment shows that RPGM Model (
Figure 8) demonstrates little high average delay
for AODV than OLSR. In RWPM, as the mobility speed increases to 50m/sec, the delay of
AODV drops down. The performance of OLSR and DSDV is almost equal (an average delay of
5.5ms) .
0.00345
0.0035
0.00355
0.0036
0.00365
0.0037
0.00375
0.0038
0.00385
10 20 3 0 40 50
AVERAGE DELAY(sec)
MOBILE NODE SPEED (m/sec)
RPGM
AODV
OLSR
DSDV
Figure 7: Mobility speed Vs Average delay in RPGM Model
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
10 20 3 0 40 50
AVERAGE DELAY(sec)
MOBILE NODE SPEED ()m/sec)
RWPM
AODV
Figure 8: Mobility speed Vs Average delay in RWPM Model
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
82
0
0.001
0.002
0.003
0.004
0.005
0.006
0.007
0.008
0.009
10 20 3 0 40 50
AVERAGE DELAY(sec)
MOBILE NODE SPEED (m/se c)
OLSR
DSDV
Figure 9: Mobility speed Vs Average delay in RWPM Model
6.3. Routing Load
This experiment investigated the routing load of 20 nodes as speed varies from 10 to 20m/sec.
Figure 10 and Figure 11 shows routing load of routing protocols in RPGM and RWPM mobility
models respectively. Compared with the AODV, OLSR and DSDV, AODV demonstrates the
lowest and OLSR shows highest average routing load for all mobility models. In RWPM, the
routing load of OLSR is comparatively more. AODV performs better than DSDV and OLSR at
the lowest speed level because it is on-demand protocol. This protocol performs best with the
RW model.
Figure 10: Mobility speed Vs Routing load in RPGM Model
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
83
0
1
2
3
4
5
6
7
8
9
10 20 30 40 50
ROUTING LOAD
MOBILE NODE SPEED (m/sec)
RWPM
AODV
OLSR
DSDV
Figure 11: Mobility speed Vs Routing load in RWPM Model
6.4. MAC Load
In this experiment, Figure 12 depicts that AODV, OLSR and DSDV all shows nearly
same performance in RPGM. However Figure 13 demonstrates that the MAC load of
AODV is approx. 8.573(low) and for DSDV is 19.614(high) in RWPM.
0
1
2
3
4
5
6
10 20 3 0 40 50
N
M
L
MOBILE NODE SPEED (m/sec)
RPGM
AODV
OLSR
DSDV
Figure 12: Mobility speed Vs Normalized MAC load in RPGM Model
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
84
Figure 13: Mobility speed Vs Normalized MAC load in RWPM Model.
7.
C
ONCLUSION
This paper studied performance of the three widely used MANET routing protocols (AODV,
OLSR and DSDV) with respect to group (RPGM) and entity (RWPM) mobility models. We
have developed a set of simulation scripts for the NS2 simulation environment merged with the
BonnMotion scenario generation tools. Simulation results have indicated that the relative
ranking of routing protocols may vary depending on mobility model. The relative ranking also
depends on the node speed.
Analysis on RPGM model shows that message delivery rate (packet delivery ratio) of AODV,
OLSR, and DSDV was almost same but as the mobility increases the message Delivery Rate
goes on decreasing. Hence it is concluded that the PDF of AODV is comparatively higher. In
case of Average End to End delay (AEED), the OLSR shows least delay as compared to AODV
and DSDV. The study shows that the AODV demonstrates lowest routing load and OLSR
shows its highest values. For performance metric, MAC load OLSR shows maximum values.
Hence it can be concluded that In RPGM model, AODV shows good performance. In RWPM
the AODV showed maximum packet delivery ratio. In case of average delay as mobility
increases the OLSR shows low delays. Routing load and MAC load of AODV is least as
compared to OLSR and DSDV routing protocols.
RPGM model is suitable for military battlefield, disaster management and other rescue
operations areas, where high message delivery rate with low routing load, low MAC load and
delay is required, Hence considering simulation results AODV is best suited protocol for this
model. Future work should be focused to extending set of the experiments by taking into
consideration the simulation parameters, different propagation models and MAC protocols.
R
EFERENCES
[1]. Mehdi Alilou, Mehdi Dehghan.t, (2005) “Upgrading Performance of DSR Routing Protocol in
Mobile Ad Hoc Networks”, World Academy of Science, Engineering and Technology.
[2]. Magnus Frodigh, Per Johansson and Peter Larsson. Wireless ad hoc Networking-The art of
networking without a network.
International Journal of Ad hoc, Sensor & Ubiquitous Computing (IJASUC) Vol.3, No.4, August 2012
85
[3]. A. Laouiti, A. Qayyum and L. Viennot, (2002) “Multipoint Relaying for Flooding Broadcast
Messages in Mobile Wireless Networks”, 35th Annual Hawaii International Conference on System
Sciences (HICSS).
[4]. Koojana Kuladinithi, Asanga Udugama, Nikolaus A. Fikouras, and Carmelita Gorg, “Experimental
performance evaluation of AODV implementations in static environments”,
http://www.comnets.unibremen. de/~koo/AODV-Perf-ComNets.pdf.
[5]. Jeremie Allard, Paul Gonin, Minoo Singh, and Golden G. Richard III, ( 2002) “A user level
framework for ad hoc routing”, In LCN ’02: Proceedings of the 27th Annual IEEE Conference on Local
Computer Networks, page 13, Washington, DC, USA,. IEEE.
[6]. Elizabeth Belding-Royer, Ian Chakeres, David Johnson, and Charlie Perkin, (2004). “DYMO –
dynamic MANET on-demand routing protocol”, In Rebecca Bunch, editor, Proceedings of the Sixty-First
Internet Engineering Task Force, Washington, DC, USA, IETF.
[7]. Charles E. Perkins and Pravin Bhagwat, “ DSDV routing over a multi hop wireless network of mobile
computers”, Technical report, IBM Research and University of Maryland, USA.
[8]. Charles E. Perkins and Pravin Bhagwat, “Highly dynamic destination-sequenced distance-vector
routing(DSDV) for mobile computers”, Technical report, IBM Research and University of Maryland,
USA.
[9]. Aleksandr Huhtonen, “Comparing AODV and OLSR routing protocols”, Helsinki University of
Technology, Telecommunication software and multimedia laboratory.
[10]. T. Clausen and P. Jacquet, (2003) “Optimized link State Routing protocol (OLSR)”, RFC-3626,
IETF Networking Group.
[11]. C. E. Perkins and E. M. Royer,( 1999) “Ad-hoc on demand distance vector routing,” in Proceedings
of the 2nd IEEE Workshop on Mobile Computing Systems and Applications (WMCSA’99), vol. 3, New
Orleans, LA, USA, pp. 90–100.
[12]. C. Perkins, E. Belding-Royer, and S. Das, (2003) “RFC 3561: Ad hoc on-demand distance vector
(AODV) routing”, category: experimental. [Online]. Available: ftp://ftp.isi.edu/in-notes/rfc3561.txt.
[13]. Valentina Timcenko, Mirjana Stojanovic, Slavica Bostjancic Rakas (2009), “MANET Routing
Protocols vs. Mobility Models: Performance Analysis and Comparison”, Proceedings of the 9th WSEAS
International Conference on APPLIED INFORMATICS AND COMMUNICATIONS (AIC '09).
[14]. NS -2, The ns Manual, Available at http: //www. isi.edu/nsnam/ns/doc.
