Evaluation of the Power Consumption of
Routing Protocols
for
Wireless Sensor Networks
# G. Santhosh
Kumar*
, Lino Abraham Varghese ,
Jose
Mathew
K
and
K.
Poulose
Jacob*
Department of Computer Science,
Cochin
University of Science
and
Technology
Cochin, Kerala, INDIA, *{san, kpj}[at]cusat.ac.in
Abstract
implosion problem experienced by Flooding
as
only
one copy
Sensor
networks
are one
ofthefastest growing
areas in
broad of
a
packet
is in transit
at
any
one
time.
In
GBR, each node in the network
can
look
at
its neighbors
wireless ad hoc
networking
(?Eld.
A
sensor
node,
typically'
hop
count
(depth) and
use
this
to
decide which node
to
forward
contains
signal-processing circuits, micro-controllers and
a
the
packet
on
to.
If the nodes' power level
drops
below
a
wireless transmitter/receiver
antenna.
Energy saving
is one
certain level it will increase the
depth
to
discourage
trafiE
of the critical issue for
sensor
networks since
most sensors
through
it
are
equipped with non-rechargeable batteries that have limited
thouH
it.
lifetime. Routing schemes are used to transfer data collected
LAHrdcstenme fndsta omnct
manytroutingse
directly with
the base
station
by
forming dynamic
clusters.
by
sensor
nodes
to
base
stations. In
the literature
manyThe
leaf nodes
were
connected
to
the cluster head which in
protocols for wireless sensor networks are suggested. In this
* turn to the base
station.
Cluster head nodes allocate each
work, four routing protocols for wireless
sensor
networks viz.
Flooding, Gossiping,
GBR and LEACH have been
simulated
lefndthtcncstoiatmelttoomuct.
This allows the leaf nodes
to
sleep
between its allocated
using TinyOS and their power consumption
is
studied using
communication slots and there
by saving
energy. The
dynamic
PowerTOSSIM. A
realization
of
these
protocols
has been
cluster head
mechanism
reduces the
energy
drain
on
particular
carried
out
using
Mica2 Motes.
nodes caused by static clusters and spreads
energy usage more
evenly
across
the network.
1. INTRODUCTION
All the above described protocols were chosen for imple-
A
wireless
sensor
network (WSN) [1]
can,
in
practice,
be
mentation and evaluation.
composed of
tens to
thousands of
sensor
nodes which
are
3. IMPLEMENTATION AND SIMULATION
distributed
in a
wide
area.
These nodes form
a
network by
communicating
with each other either
directly
or
through
uTinyOS
[4]
provides
a
multihop
architecture [5] to
spec-
other nodes. One
or more
nodes among them will
serve
ifytwo
major
components
viz.
a
packet
movement
logic Mul-
as
sink(s), known
as
the base
node,
that
are
capable of
tw
o
rmmnetiz,oa pakt
mv e
logicPM
communicating
with the
user
either
directly
or
through the
topnie
for
multiop Tin
a
n
tiHopE
M
.. .
.
module for
sath selection.
The
components
MHFlood-
existing
wired networks. Large
amount
of battery
power is
p
p
used up
during
internal
processing
and communication. So
ingPSM (Flooding) MHGossipingPSM (Gossiping), MHG-
other than the hardware, even the software loaded
with
the bPM(B)adMLahS
LAH
eeipe
othertanthehardwae,eve
the
software
loaded
wmented
by modifying the base
components
using nesC [6].
sensor
nodes
is
desired
to
be
power
eflkient.
This
paper These
attempts to
analyze the
power
requirements
of
some
of the
components ainairoti
popular
routing
protocols used in sensor
networks.fosectnaruefrapck.
TOSSIM
[7]
is used
to
evaluate the
implemented
protocols.
2. RoUTING PROTOCOLS
FOR
WSN
A loss
topology
is
deOiied
which allows
a
node
to
communi-
cate
with all nodes within
a
5
x
5
square
around itself.
All
According
to
nodes' participating style, routing protocols
can the protocols were simulated for a period of 500 second. The
be classiaEd into three categories, namely, direct communi-
power consumption of the routing protocols is calculated using
cation,
ceat
(Minimum Energy Transmission), and clustering
the PowerTOSSIM [8]. The simulation also tracks the number
protocols. The basic routing protocols for WSN
are
Flooding,
of messages sent, received and forwarded through each node.
Gossiping, Gradient Based Routing (GBR) and LEACH
[2],
[3].
4. RESULTS
In
the Flooding protocol packets
were
broadcasted
to
all Using PowerTOSSIM,
power usage per
node (median and
possible
routes to
its destination. Gossiping,
an
improved
standard
deviation)
and
power usage per message
for each
version of Flooding protocol, instead of
broadcasting
each protocol
were
tabulated. These data
were
used
to
compute
packet to all neighbors, the packet is sent to a single neighbor median and standard deviation and plotted
graphically
as
chosen at random from a neighbor table. Gossiping avoids the shown in Fig. 1 The median of the power usage will give the
1-4244-0731-1/06/$20.OO
©2006
IEEE.
194
Authorized licensed use limited to: COCHIN UNIVERSITY OF SCIENCE AND TECHNOLOGY. Downloaded on February 7, 2009 at

06:42 from IEEE Xplore. Restrictions apply.Page 2

8
140
10 Nodes
10
Nodes
25Nodes
25 Nodes
Flooding
Gossiping
GBR
LEACH
Flooding
Gossiping
GBR
LEACH
Protocols
Protocols
(a)
(b)
100
12010
.o 9O0
|_25 Nodes|
125 Nodes|
60
1
000
20
2000
Flooding
Gossiping
GBR
LEACH
Flooding
Gossiping
GBR
LEACH
Protocols
Protocols
©
(d)
Fig. 1: Results of simulation (a) Average hop count of the messages forwarded by the base station. (b) Power usage per

message sent. © Power usage per
node (Standard Deviation). (d) Power usage per node (Median).
average power consumption of the protocol across the network power consumption per node was less and the power usage
nodes. The standard deviation of the power usage per node will among the network nodes was even as well, making it the
give an estimate of evenness of power consumption across the ideal routing protocol for power constrained wireless sensor
network nodes. Also the power usage per message is calculated networks. The developed protocols were realized on a

wireless
to get the average power usage for each message sent across sensor network based on Mica2 Motes.
the network.
Evaluating the simulation results of the four routing pro-
RFRNE
tocols namely - Flooding, Gossiping, GBR and LEACH we
[l]
I. F. Akyildiz, W. Su,
Y.
Sankarasubramaniam, and B. Cayirci "Wireless
sensor networks: a survey", Computer Networks, Vol. 38, No. 4,
2002,
analysed that
Flooding iS
the worst in case of power
eftiEiency.
pp. 393-422.
Gossiping provided some improvement over Flooding in terms
[2]
Carlos de Morais Cordeiro, Dharma Prakash Agrawal Ad Hoc and Sen-
of power usage per message but the power usage per node was
sor Networks: Theory and Applications,
1st
edition, World
ScientiU7c
..
Publishing
Co. Pte.
Ltd., Singapore, 2006,
pp. 19-79.
stlhge.GainBaeRotnprtclw
s
giin loe
[3]
Wendi Heinzelman,
Anantha Chandrakasan, and
Hanl Balaknishnan,
power usage
Ultures
than the previous two; as the number of
"Energy-Ef(iFcient
Communication Protocols for Wireless Microsensor
node increases the power usage becomes uneven across the
Networks", In: Proc. Hawaaian Int'l
Conf:
on Systems Science, Maui,
networ node sinc manynodeswere ettin overutilied
in
Hawaii, 2000.
networ nodesinc manynodeswere gttingover tilizd in
[4]
The
ofU7cial
TinyOS website
http://tinyos.netl
routing packets. The median and standard deviation for power
L5] TinyOS tutorial found in $TOSROOT/doc/tutorial/index.html
($TOS-
usage per node is less for LEACH when compared to all the
ROOT is the TinyOS installation directory)
otherprotocolsanalysed.
L6]
David Gay, Phil Levis, Rob von Behren, Matt Welsh, Eric Brewer, and
otherprotocolsanalysed.
David
Culler, "The nesC Language: A Holistic Approach to Networked
Embedded Systems", In: Proc. Programming Language Design and
5. CONCLUSIONS
Implementation (PLDI), 2003.
[7] Luiz Felipe Perrone and David Nicol, "A Scalable Simulator for TinyOS
In this work, four routing protocols for wireless sensor
Applications", In: Proc. Winter Simulation Conference, 2002.
networks namely - Flooding, Gossiping, GBR and LEACH
[8]
Victor Shnayder, Mark Hempstead, Bor-rong Chen, and Matt Welsh,
weesimulated and their results were evaluated. The dynamic
"Simulating the Power Consumption of LargeScale Sensor Network
were
Applications",
In: Proc. 2nd
international
conference on Embedded
cluster based protocol LEACH stands out among the other
networked sensor systems , Baltimore, MD, USA, 2004, pp. 188-200