1
Abstract—Selecting a routing protocol for a wireless sensor
network depends on various factors like the network lifetime,
success rate and the number of nodes in the network. This
paper compares four popular routing protocols currently used
in wireless sensor networks. The experimental setup is
described, after which the various protocols are compared and
evaluated.

Index Terms—routing protocols, wireless sensor networks,
multihop routing, power consumption
I. INTRODUCTION
routing protocol is required when a source node can not
send its packets directly to its destination node but has
to rely on the assistance of intermediate nodes to forward
these packets on its behalf. Although many protocols exist
for traditional ad hoc wireless networks, none of them are
suitable to the unique requirements of wireless sensor
networks.
Routing protocols can be divided into two groups,
proactive and reactive routing protocols [1]. With proactive
routing a routing table is generated at each node, so that
routing information is kept for every node in the network.
This is used in traditional link-state and distance vector
routing protocols, with the DSDV (Destination-Sequenced
Distance Vector) protocol being proposed for wireless
sensor networks.
With reactive routing no routing tables are generated and
route discovery is done as needed. The route information is
then kept for future reference. Existing reactive protocols
include AODV (Ad hoc On-demand Distance Vector) and
DSR (Dynamic Source Routing). The problem with these
protocols is that they minimize the total energy consumed in
the network, but drain energy along paths chosen for
minimum energy consumption. Network partitioning
happens when some nodes in the network are drained from
their battery energy more quickly than others, leading to
some parts of the network being cut of from the rest of the
network.
In this paper the following routing protocols are simulated

Manuscript received March 3, 2007.
G. Niezen and G.P. Hancke are with the Department of Electrical,
Electronic and Computer Engineering, University of Pretoria, South Africa.
(e-mail: gniezen@ieee.org, g.hancke@ieee.org).
I.J. Rudas and L. Horváth are with Budapest Polytechnic, Hungary.
(e-mail: rudas@bmf.hu, laszlo.horvath@nik.bmf.hu)
The authors acknowledge the grant provided by the National Office of
Research and Technology (Hungary) and the Agency for Research Fund
Management and Research Exploitation (KPI) (Hungary) in the Hungarian-
South African Intergovernmental Science and Technology Cooperation
Programmes (project number ZA-37/2006). The authors also acknowledge
the grant provided by the OTKA Fund for Research of the Hungarian
Government (project number K 063405).

and evaluated:

Flooding protocol - no routing tables are kept and
messages are broadcast across the network,

TinyOS 1.x Multihop Routing – a shortest-path-first
algorithm with a single destination node (the sink node)
and active two-way link estimation,

Low-Energy Adaptive Clustering Hierarchy (LEACH)
– a clustering-based protocol [2], and

Ad hoc On-demand Distance Vector (AODV) – a
reactive routing protocol [3].
II. EXPERIMENTAL SETUP
A. Simulation settings
The various routing protocols were compared through the
simulation of wireless sensor networks with the OMNeT++
Discrete Event Simulation System [4]. In [5] it is stated that
traffic generation should be based on intended applications.
Since wireless sensor networks are usually designed to
handle sporadic or slowly changing events (such as
temperature and pressure variations) [6], it was assumed that
the application layer transmits information every 15 minutes.
In table 1 the settings used during the simulation are
displayed. For the radio model, a transmission distance of
1m was used, while the transmitter power was set at 2mW on
the mobility framework's physical layer implementation.

B. Calculating power consumption
To calculate power consumption, the simple radio model
introduced by [7] was used. It was initially used by the
authors to calculate the power consumption of the LEACH
protocol and was implemented in the OMNeT++ simulator
by [8].






A
TABLE I
SIMULATION SETTINGS
Setting Value
Transmission distance d=1m (for power calculations)
P max = 2mW (transmitter power)
Bit rate 10 kbps
PNRG used and seed Mersenne Twister
Seed automatically generated from
OMNeT++ simulator
Simulation package and version
number used
OMNeT++ Discrete Event
Simulation Package 3.2p1 with
Mobility Framework 2.0p2
Traffic type Data message sent once every 15
minutes
Mobility model Static, with random node placement
Comparing wireless sensor network routing
protocols
Gerrit Niezen, Gerhard P. Hancke, SMIEEE, Imre J. Rudas, FIEEE, and László Horváth, SMIEEE
1-4244-0987-X/07/$25.00 ©2007 IEEE.