06295776
-
Upload
dhirajkapila -
Category
Documents
-
view
219 -
download
0
Transcript of 06295776
-
7/29/2019 06295776
1/5
Abstract-Wireless sensor networks have gainedconsiderable attention in the past few years. They have
found application domains in battlefield communication,
homeland security, pollution sensing and traffic
monitoring. As such, there has been an increasing need
to define and develop simulation frameworks for carrying
out high-fidelity WSN simulation. For our simulation
based study we have considered three different scenarios
where a Adhoc routing protocol can be applied, namely a)
Phenom nodes with Mobility, Sensor nodes with mobilitywhile the SINK node is stationary b) Phenom nodes with
Mobility, Sensor nodes are stationary and Sink node is
stationary c) Phenom nodes are stationary, Sensor nodes
are stationary and SINK node is stationary. We have
considered AODV and DSR as the Adhoc routing
protocols for comparison in these scenarios.Simulation
has been carried out by using qualnet 5.0.2 simulator.
The matrices used for this simulation are packet Deliver
Ratio to SINK, Average End to End Delay to SINK,
Overhead, Throughput. Finally, through our analysis we
show the various intricacies that are associated with
Adhoc routing protocols when used in a sensor
environment. We also list the various problems we faced
while simulating the routing protocols in each of the
scenarios.
Keywords- Sensor networks, Adhoc routing protocols,
scenario study, simulation, AODV, DSR.
I. INTRODUCTION
A Wireless Sensor Network (WSN) contains
hundreds or thousands of these sensor nodes. These
sensors have the ability to communicate either among
each other or directly to an external base-station (BS).
A greater number of sensors allows for sensing over
larger geographical regions with greater accuracy.
Even though sensor networks are a superset of ad
hoc routing protocols, the routing protocols proposed
for ad hoc routing protocols cannot be used as it is for
sensor networks because of various reasons as given
in [1, 2]. But surprisingly we found out that there is
lack of simulation based study or research work [5] as
to show why ad hoc routing protocols cannot be used
in a sensor network environment. The main
contribution of this paper is that we have carried out a
simulation based study of ad hoc routing protocols to
understand their behavior when used in a sensor
network environment.
This paper is organised as follows. In the second
section 2, we give brief introduction of AODV and
DSR [6] routing protocol and in section 3 deals the
simulation setup and results obtained on the execution
of simulation. Finally we draw the conclusion in
section 4.
II. DESCRIPTION OF ROUTING
PROTOCOL
A. Ad-Hoc on Demand Distance Vector (AODV)
We give brief description of routing protocols andthe simulator, which we have considered for studying
the behavior of ad hoc routing protocols in WSN [7].
Adhoc On Demand Distance Vector (AODV) Routing
Protocol: AODV routing protocol uses on demand
approach for finding routes. In AODV the source
node and the intermediate nodes store the next hop
information corresponding to each flow for data
packet transmission. To find a route to the destination,
the source node floods the network with
RouteRequest packets. The RouteRequest packets
create temporary route entries for the reverse path
through every node it passes in the network. When it
reaches the destination a RouteReply is sent back
through the same path the RouteRequest was
transmitted. For route maintenance, every routingtable entry maintains a route expiry time which
indicates the time until which the route is valid. Each
time that route is used to forward a data packet; its
expiry time is updated to be the current time plus
Active Route Timeout. A routing table entry is
invalidated if it is not used within such expiry time.
AODV [5] uses an active neighbor node list for each
routing entry to keep track of the neighbors that are
using the entry to route data packets. These nodes are
notified with routeerror packets when the link to the
next hop node is broken. Each such neighbor node, in
turn, forwards the routeerror to its own list of active
neighbors, thus invalidating all the routes using the
broken link. The main advantage of this protocol is
that routes are established on demand and destinationsequence numbers are used to find the latest route to
the destination. The disadvantage of this protocol is
that the intermediate nodes can lead to inconsistent
routes [3].
III. DYNAMIC SOURCE ROUTING (DSR)
Dynamic source routing protocol (DSR) [4]: DSR
is an on-demand routing protocol. The major
difference between DSR and the other on demand
routing protocols is that, it is beacon less and hence
Scenario Based Analysis of AODV and DSR ProtocolsUnder Mobility in Wireless Sensor Networks
Varsha Sahni Pankaj Sharma Jaspreet Kaur Sohajdeep singh
CSE Dept. CSE Dept. CSE Dept. CSE Dept.CTIEMT CTIEMT CTIEMT Jalandhar, India
Jalandhar, India Jalandhar, India Jalandhar, India [email protected]
[email protected] [email protected] [email protected]
2012 International Conference on Advances in Mobile Network, Communication and Its Applications
978-0-7695-4720-6/12 $26.00 2012 IEEE
DOI 10.1109/MNCApps.2012.42
160
-
7/29/2019 06295776
2/5
does not require periodic hello packets. Consider a
source node that does not have a route to the
destination. When it has a data packet to be sent tothat destination, then it initiates a RouteRequest
packet. This RouteRequest is flooded throughout the
network. Each node upon receiving a RouteRequest
broadcasts the packet to its neighbors if it has notforwarded already or if the node is not the destination
node. Each RouteRequest carries a sequence number
generated by the source node and the path it as
traversed. A node, upon receiving a RouteRequestpacket, checks the sequence number on the packet
before forwarding it. The packet is forwarded only if
it is not a duplicate Route Request packet. Thus, all
the nodes except the destination node, forwards aRouteRequest packet during the route construction
phase. A destination node upon receiving the
RouteRequest packet, replies to the source node
through the reverse path the RouteRequest packet hadtraversed.
IV. SIMULATION SETUP
We have used Network Simulator Qualnet 5.0.2in our evaluation. The simulated sensor area is
501*501m rectangle. The size of each of the message
sent is 100 bytes. We have simulated the sensor
nodes from 50 nodes to 100 sensor nodes. We havekept hte simulation time constant at 100 sec. The
transmission and recieving power of 175mW and
initial energy of about 5 joules is made available. The
sensing energy for the phenom nodes is kept at1.75mW. The parameters used for carrying out
simulation are summarized in the table 1.
Table 1: Simulation Parameters
Parameters Value
Routing
ProtocolsAODV, DSR
MAC Layer 802.11
Packet Size 100 bytes
Terrain Size 501m * 501m
Nodes 50, 100
Data Traffic
TypeCBR
No. of Source50,60,70,80,9
0,100
Simulation
Time
100
sec.
TransmissionPower
175mW
Energy 5 Joules
Transmission
Range50 m
V.PERFORMANCE METRICS:
We have used the following metrics forevaluating the performance of two on-demand
reactive routing protocols (AODV & DSR):
A. PACKET DELIVERY RATIO TO SINK
It is the ratio of data packets delivered to thedestination to those generated by the sources. It is
calculated by dividing the number of packet received
by destination through the number packet originated
from source. PDF = (Pr/Ps)*100
Where Pr is total Packet received & Ps is the total
Packet sent. The greater packet delivery ratio is, themore realiable the network is.
B. AVERAGE END-TO-END DELAY
(SECOND) TO SINK:This includes all possible delay caused by
buffering during route discovery latency, queuing at
the interface queue, retransmission delay at the MAC,
propagation and transfer time.
(Time packet received Time packet sent)Total No. of Connection Pairs
Where Tpr is receive Time and Tps is sent Time
C. OVERHEAD
It is the sum of all the control packets sent by allthe sensor nodes in the network to discover and
maintain routes to the SINK node.
D. THROUGHPUTThroughput can be defined as the :
Node Throughput Data Transmission
Total Number of Nodes
A high network throughput indicates a small error
rate for packet transmission and a low level for
contention in the network.
Result Analysis
We have considered three scenarios to study the
behavior of ad hoc routing protocols in wirelesssensor networks.
A. Scenario One
In scenario one, we consider Phenom Nodes withMobility, Sensor Nodes with Mobility and SINK
without Mobility.
161
-
7/29/2019 06295776
3/5
Figure 1. Number of Nodes vs. PDR to SINK
In fig 1 we have mapped the PDR to SINK v/s
number of nodes. As can be seen from the graph DSR
is having the highest packet delivery ratio to the
SINK. There is an increase and decrease pattern
among the packets delivered to the sink. The decreasein the packet delivery may be mainly due to the
collisions caused by the flocking of phenom nodes
and sensor nodes together. If the sensor nodes arelittle far away from each other, while getting the
phenom packets in large number then there will beincrease in the packets delivered to the SINK.
Figure 2. Number of Nodes vs. EED to SINK
In fig 2 we have mapped the delay taken by thesource node packet to reach the SINK node in the
network. With both phenom nodes and sensor nodes
having mobility then the chances of link failure are
very much high. Link failures triggers new routediscoveries in AODV since it has at most one route
per destination in its routing table. The delay in
AODV is less than DSR since AODV finds the routes
only when it is needed. Also, as the number of nodes
increases, the time taken by the DSR to find theroutes to the SINK node in the cache increases. DSR
has longer delay because their route discovery
mechanism takes more time as every intermediatenode tries to extract information before forwarding
for reply.
Figure 3. Number of Nodes vs. Overhead
In fig 3 we have mapped the overhead occurring
in the network v/s number of nodes. The initial
overhead in DSR is high, but it decreases when the
nodes >80. This happens due to the cachingmechanism in DSR. Since DSR finds the routes
required to reach the SINK from its cache, the
overhead decreases. But as the number of nodes inthe network is large both AODV and DSR will have
more or less the same amount of overhead. The majorcontribution to the AODVs routing overhead is from
route requests, while route replies constitute a largefraction of DSRs routing overhead. Since DSR
replies to every request in source routing it causeshuge overhead than AODV.
Figure 4. Number of Nodes vs. Throughput
In fig 4 we have mapped the throughput to SINKv/s Number of packets. The initial throughput of DSRis more than AODV. But when the number of nodes
> 70, then the throughput of AODV and DSR slowly
starts to converge with each other. AODV has
difficulty in delivering the packets when the nodesare moving fast resulting in fewer throughputs than
DSR. In DSR the source routing reveals more
information in one route discovery than AODV.
Therefore, within the same time more routes arediscovered and so more packets are delivered. AODV
catches up when the mobility of the nodes gets lower.
This is because routes become more stable and so
eventually everybody can find all the routes it ever
needs.
B. Scenario Two
In scenario two, we consider Phenom Nodes withMobility, Sensor Nodes without Mobility and SINK
without Mobility.
Figure 5. Number of Nodes vs. PDR to SINK
In fig 5 we have mapped the number of packetsdelivered to the sink node. The packet delivery to
SINK of both AODV and DSR decreases as thenumber of nodes increases in the network. But the
number of packets delivered is low in DSR and startsto decrease once the number of nodes >70. But with
162
-
7/29/2019 06295776
4/5
high mobility, the chances of the caches being stale
are quite high in DSR. Eventually when a route
discovery is initiated, the large numbers of replies
received in response are associated with high
overhead and cause increased interference to data
traffic. Hence, the cache staleness and high overhead
together result in significant degradation inperformance for DSR in high mobility scenarios.
Figure 6. Number of Nodes vs. EED to SINK
In fig 6 we have mapped the delay taken by the
source node packet to reach the SINK node in the
network. The end to end delay of the DSR is more
than AODV when the number of nodes >70. DSR has
high end to end delay as its route discovery
mechanism takes more time since every intermediate
node tries to extract information before forwarding
the reply. Also, the routes become invalid due to the
phenom nodes movement. To adapt to this change,
each sensor node constantly monitors the links it uses
to forward the packets resulting in more end to end
delay.
Figure 7. Number of Nodes vs. Overhead
In fig 7 we have mapped the overhead occurring
in the network v/s number of nodes. The overhead ofDSR is more than AODV when the number of nodes
>80. In DSR, a source node reaches the destination
by flooding the network with a route request (RREQ)
message. When a RREQ message reaches the
destination a route reply message is sent back to the
source, including the newly found route. This route
request and route reply messages increases as the
number of nodes increases leading to increasing in
the overhead of the network.
Figure 8. Number of Nodes vs. Throughput
In fig 8 we have mapped the throughput to SINK
v/s Number of nodes. Throughput of both AODV and
DSR decreases as the number of nodes increases in
the network. But in DSR the decrease is very rapid
with the throughput of DSR decreasing to nearly 2%
when the number of nodes in the network is more
than 100.
C. Scenario Three
In scenario three, none of the nodes that is, the
phenom nodes, the sensor nodes or the sink nodes
have mobility.
Figure 9. Number of Nodes vs. PDR to SINK
In fig 9 we have mapped the number of packets
delivered to the sink node. The packet delivery of
both AODV and DSR decreases when the number of
nodes increases in the network. But the decrease is
severe in the case of DSR. In DSR, routing is done
through source routing i.e. the sender of a packet
determines the complete route from itself to the
destination, and includes the route in the packet. And
since the whole route is carried in each data packet,
intermediate nodes have the opportunity to inspect
the whole route and extract information from it
resulting in more delay to reach the SINK node. Alsothis results in traffic congestion leading to loss of
data packets. When the nodes are stationary the
caching mechanism does not have much effect in
improving the performance of DSR.
163
-
7/29/2019 06295776
5/5
Figure 10. Number of Nodes vs. PDR to SINK
But when the route is discovered DSR spends
more time in replying to other intermediate nodes,
since it will be having more information to the
destination node. The delay in DSR is more when the
number of nodes increases >70 in fig 10. But AODV
will be having one single route and packets will be
delivered through that route resulting in less delay.
Figure 11. Number of Nodes vs. Overhead
In fig 11 we have mapped the overhead occurringin the network v/s number of nodes. As can be seen
from the graph the overhead for DSR increases
astronomically when the number of nodes in the
network is >90. But AODV is having less overhead.
As we have discussed in previous section DSR makes
use of source routing. If there is a link failure, then
many nodes will be affected. Furthermore, there is a
possibility that a node extracts route informationfrom a stale source route. Using the invalid route
information causes unnecessary overhead. DSR
always generates more RREP( route reply) and
RREQ (route requests) than AODV resulting in more
overhead. It has been reported in various ad hoc
research papers that the overhead decreases as themobility decreases. But we found a contrasting
situation where even after all the nodes are stationary,
there was no decrease in the overhead of DSR but it
increased proportionally.
Figure 12. Number of Nodes vs. Throughput
In fig 12 we have mapped the throughput to SINK
v/s Number of packets. Throughput of both AODV
and DSR decreases as the number of nodes increases.But in DSR, the fall in throughput is severe and it is
less than 5%, when the number of nodes is >90. Even
though DSR by virtue of its source routing can
maintain caching mechanism it does not matter when
the nodes are stationary.
VI. CONCLUSION
In this paper we have studied the behavior of thead hoc routing protocols in wireless sensor networks
under various scenarios (ref Table 1). Through
simulation and our analysis, we show that DSR is the
routing protocol to be used in an environment like
scenario one and AODV is the routing protocol to be
used in environments, where scenario two and
scenario three are considered. Also we list the various
problems we had to face while simulating the routing
protocols in a sensor network paradigm. Our future
work includes designing an energy efficient routingprotocol for Wireless Sensor Networks. With all
these research challenges we firmly believe that wehave a very exciting time ahead of us in the area of
Wireless Sensor Networks.
REFERENCES[1] I. Akylidiz, W. Su, Sankarasubramaniam, and E.Cayrici, A
survey on sensor networks, IEEE Communications
Magazine, Volume: 40 Issue: 8, August 2002, pp.102-114.
[2] K. Akkaya and M. Younis, A survey of Routing Protocols
in Wireless Sensor Networks, Elsevier Ad Hoc NetworkJournal, 2005, pp 325-349.
[3] C.Perkins, E.B.Royer and S.Das,AdHoc On-Demand
Distance Vector (AODV) Routing, RFC 3561, IETF
Network Working Group, July 2003.
[4] E.M.Royer and C.K.Toh, A Revisew of Current Routing
Protocols for Ad-Hoc Mobile Wireless Networks, IEEEPersonal Communications Magazine, April 1999, pp. 46-55.
[5] Bertocchi, F et.al, Performance Comparison of Routing
Protocols for Ad hoc networks, In proceedings of Global
Telecommunications Conference, 2003, pp 1033-1037.[6] H. Jiang and J. Garcia-Luna-Aceves,Performance
comparison of three routing protocols for ad hoc networks,
In Proceedings of IEEE ICCCN 2001.
[7] Gowrishankar. S, T.G. Basavaraju, SubirKumarSarkar,
Issues in Wireless Sensor Networks, In proceedings of the2008 International Conference of Computer Science and
Engineering, (ICCSE 2008), London, U.K., 2-4 July, 2008.
164