,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

Wireless Sensor Networks Analysis based

on MAC Protocols
Furat Al-Obaidy Shadi Momtahen Farah Mohammadi
Department of Electrical, Computer, and Department of Electrical, Computer, and Department of Electrical, Computer, and
Biomedical Engineering Biomedical Engineering Biomedical Engineering
Ryerson University Ryerson University Ryerson University
Toronto, Canada Toronto, Canada Toronto, Canada
furat.alobaidy@ryerson.ca smomtahen@ryerson.ca fmohamma@ee.ryerson.ca

minimize the delay. According to the results, the work improved

Abstract— Power conservation has been an important area of
interest in Wireless Sensor Network (WSN). Medium Access
Control (MAC) protocols play an important factor in power
consumption. As wireless sensors operate on batteries, periodically
replacement of which is considered a critical issue. So that a
different scenario has been designed and implemented to optimize
the power consumption with life operating increases for these
sensors through a different efficient protocol across layers.
However, there are four main classes of MAC protocols due to how
the node organizes within the shared radio channel. This work
studies and compares the typical protocols (S-MAC, T-MAC, B-
MAC, L-MAC, and Crankshaft), which cover the four classes,
analyzes their performances, and then evaluate the results. The
purposes of this experiment are to know how these protocols
interact with typical network conditions and how power
consumption can be monitored to determine the optimal protocol
for each application.
Keywords—Medium access control ,Wireless sensor network, Power
efficiency.
I. INTRODUCTION
A wireless sensor network consists of a hundred wireless
sensor, running in a harsh condition, and having monitoring, data
collecting, processing and communication in real time. These
sensors receive and transmit data to each other. The typical
architecture of each node consists of sub-units; a microcontroller
unit, a communication unit, and a battery unit. Each sensor unit is
sensing and generating measurable signal against any physical
changes in its surrounding area, for monitoring such pressure or
temperature. Besides that, routing protocols should be handled
functionally with the large number for these nodes or sensors.
Different MAC protocols are proposed to communicate between
the wireless network sensors. The main challenge with using
MAC protocols is to find an optimal protocol against power
consumption and data throughput [1-3].
Some of the recent works have studied power-efficient MAC
protocols for the WSNs; work [4] provides a proposed approach
to enhance the energy efficiency WSN. In that study, a MAC
protocol based on adaptive time slot condensing is used to
power efficiency and reduced latency for the wireless network
compared with traditional methods.
Paper [5] surveys different approaches used for MAC
protocols design for enhancing the power consumption and
increasing lifetime for sensors in healthcare applications, which
monitor vital elements for a human body. The author suggests a
hybrid protocol combined the features for both Schedule Time
Division Multiple Access (TDMA) protocol and Carrier Sense
Medium Access (CSMA) protocol. Two phases of operating are
applied through this protocol: active duty cycle (i.e. the time for
communication) and sleep duty cycle (i.e. for the rest time). In
another survey [6], the authors compared between some of MAC
protocols, Sensor-MAC (S-MAC), Timeout-MAC (T-MAC),
and Traffic- Adaptive MAC (TRAMA), and they concluded that
each of these protocols addresses different issues that arise from
power waste in sensor nodes. Furthermore, each protocol works
its best in different environments, i.e. the power consumption is
reduced significantly when the S-MAC protocol is used in light
traffics loads. But, this study showed the power consumption
with TRAMA depends on the amount of provided data
throughput.
The author in [7] Efficient Parallel MAC (EP-MAC) protocol
with a TDMA scheduling based on different traffic of data
packets in WSN. This approach removes the drawback of
scheduled based approach in MAC by utilizing contention-based
scheme. This improves the lifetime of sensor nodes. The
comparison results showed that the power consumption for each
node of EP-MAC is less than the power consumption for three
other protocols (S-MAC, T-MAC and Hybrid MAC, with
different traffic load) because the proposed EP-MAC avoids
redundant control signal. But, the main problem with the delay in
EP-MAC is higher than the delay of the other three protocols
because it has a longer latency period.
Furthermore, many different MAC protocols are being
studied for the betterment of the power efficiency property. For
instance, the authors in [8] used a MAC scheduling for their
proposed control; the algorithm is called Power-Efficient MAC
(MACPE). The prioritized frames are used for urgent
transmitting. The authors compared the MACPE performance
with three protocols (i.e., S-MAC and T-MAC). The results

k,(((
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(
concluded this approach consumes power less than S-MAC, and
T-MAC respectively. But, this algorithm has two disadvantages;
first, it is less affected when the number of send or nodes
increases; second, the power consumption is increased directly
with the number of control frames transmitted.
This paper compared the performance of a set of recent
protocols and pointed out an additional aim that needs to achieve
high power efficiency and throughput to obtain the optimal
design for MAC protocol. The rest of the paper consists of types
of MAC techniques (Section II); then, we describe the simulation
setup (section III); section IV provides simulation results and
performance evaluation. Finally, the conclusion is presented in
section V.
II. TYPES OF MAC TECHNIQUES
In WSN, many energy efficient MAC protocols are
presented. We will classify them according to the time
organization into four categories as is shown in fig. 1 [2, 9].
x Random access: this class of protocols does not restrict when
nodes access the channel. This provides a flexible handle for
different node densities and traffic data channels, so nothing
has to be decided before deployment and dynamic changes
can be installed easily. Besides that, the nodes can work in a
synchronized manner and make this protocol rather simple.
The limitation of this access is that lots of power are often
wasted due to idle listening and collisions. Under this
category, the protocol is a Berkeley-MAC (B-MAC).
x Slotted access: the principle of this protocol is periodic sleep,
listen by the node of the network. The time interval is
divided into slots, and nodes wake up at the staring of the
slot to handle pending messages waiting for transmission.
Channel access is based on contention as the random-access
protocols, but the collision rate is increased due to all
communication being grouped into the small active point of
the farm. Therefore, the effective contention resolution is
much more a priority and some slotted protocols, even
going as far as including collision avoidance signalling
despite the relatively large protocol overhead. The main
limitation is the slotted protocols mainly differ in their
policy when they are switching between active and sleep
modes. Under this category, the protocols are S-MAC and
T-MAC.
x Frame access: this type of slots is grouped into frames and
scheduling in detail according to how it can be sent in each
slot. The main features of a schedule-based approach is that
collisions do not occur, and that idle listening and
overhearing can be totally minimized. When scheduling,
communication channels, that is, specifying the sender-
receiver pair per slot, nodes only need to listen to those slots
in which they are the required receiver eliminating all
overhearing. The protocol Lightweight-MAC (L-MAC) is
an example of this category.

x Hybrid access: the protocol splitting channel into two parts,
channel control packets and data packets; in channel control
packets, information are sent in the random access and in
the data packets, information is transmitted in the scheduled
channel. The example of this access is in the Crankshaft
protocol.
This work will study and compare power - efficiency of the
protocols (B-MAC, S-MAC and T-MAC, L-MAC,
Crankshaft), which are included in the above mentioned
categories.
WSN MAC Protocols
Hybrid Frame Slots Random
Crankshaft PEDAMACS S-MAC RATE EST
L-MAC STEM
P-MAC D-MAC
AI-LMAC
Z-MAC TRAMA
Preamble sampling
T-MAC
Wise MAC
FLAMA SCP-MAC
CSMA-MPS
LPL
B-MAC
X-MAC
Fig.1. Types of mac protocols according to time
organization and historic development.
III. SIMULATION SETUP
Our protocol designed and evaluated based on OMNeT++
simulator. This simulator is a widely-used network simulation
package because it is a fully programmable, modular, and open-
source simulation tool. For our experiments, we use the layout of
the equivalent real-world potato-field experiment in [10]. This
setup includes 96 nodes on a field of approximately 90×50
meters. The simulated nodes have a radio range of 25 meters,
which is like the radio range in the real-world experiment. The
base station is situated near a corner. The topology is simple, but
it is enough to show the basic characteristics of the MAC
protocols [11].
In our simulations, we have included five protocols: S-MAC,
T-MAC, B-MAC, L-MAC, and Crankshaft. All nodes periodically
send information to a head node, which then processes the data
and stores them for further processing. Broadcast floods are
typically found in routing protocols, and in distributing queries
over the network. The power consumption is evaluated for each
node based on the various MAC protocols and under different
loads of traffic. The simulation time consumed 500 seconds. For
each, with different average sending data rates (50,
75,100,150,175,200,250) bits/sec. The optimal model reflects an
optimization in terms of latency, power consumption, and delivery
ratio. Table I lists the simulation parameters. The routing is done
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

using a static routing table. Therefore, there is no routing traffic
exchanged during the simulation. In this simulation, the power
consumption in each node is calculated by multiplying the time
with the required power to operate the radio in that mode.
IV. SIMULATION PARAMERESULTS AND PERFORMANCE
EVALUATION
Fig.2 shows the results of the power consumption for each
selected protocol with the data flow rate changing. Our results
prove that Crankshaft protocol consumes less power against the
other protocols. So that Crankshaft protocol is very power
efficient, and the power consumption does not exceed (1
milliwatt) at 250 bits/sec. There are two factors which contribute
to the power efficiency: firstly, because only a subset of the
nodes is awake in each slot overhearing is reduced. Secondly, the
contention is reduced. While the power consumption in L-MAC
protocol remains nearly constant around 1.5 milliwatts because
the L-MAC protocol already saturates at low message rates. The
small increase in power consumption is due to more messages
In Fig.4, all the protocols are compared and proved that B-
MAC has the lowest latency (0.1 sec). This is caused by a lack of
information and retries. Messages are not kept in queues waiting
for resending of other packets, which means latency is kept
down. Crankshaft’s latency rises quickly (from 4.1 sec to 10.2
sec) as the delivery ratio drops. This is partly caused by messages
having to wait in queues due to congestion and partly caused by
messages having to wait for a full frame when contention is lost.
L- MAC suffers from the highest latency of up to 11.6 sec, due to
the multiple slots within a time frame and which obliges each
node to exchange information within their space all the time. The
latency in the protocols S-MAC and T-MAC seems low for high
message rates (about 3 sec), but the data that can be delivered by
this type, does not exceed 30% of the data sent.
Table. I. SIMULATION PARAMETERS.
Frequency 2.4 GHz Battery Voltage 3 Volt
Battery capacity 1000 mAh Transmission current 17 mA
Sleep Current 0.02 mA Receiving current 16.4 mA

that are sent one hop, only to be discarded because of full Frame length 610 ms Poll length 0.3 ms
message queues at the receiver. A large number of slots also Contention window 9.15 ms Max. data length
induce a high message latency for the L-MAC protocol. At the T- 64 bytes

T-MAC
MAC protocol, the power consumption increases with a flow

Crankshaft
Packet header 8 bytes Sift node parameters 512
rising of data in the network. The results cleared a rise in power Max. data length 250 Packet header:
consumption by 0.75 milliwatts when the flow of data changed
bytes 11 bytes
from 50 to 250 bits / sec. Since the T-MAC protocol adjusts
Activity timeout 15 ms Unicast slots 8
Adaptive Timeout (TA) to fit the data flow, with the increase TA,
the power consumption is also increasing. In the Protocol B- Max. data length 64 bytes Broadcast slots 8
MAC, when the flow of data was low, power consumption Slots per frame 20 Check Interval 0.3 Sec
increased. It increased consumption from 4.1 milliwatts at 50

B-MAC
Packet header 24 bytes Slot Duration 1 Sec
L-MAC

bits/sec to 7.5 milliwatts at 100 bits/sec. Then, power

consumption, backed and settled at a moderate increase equal to
8.4 at 250 bits/sec. Higher consumption of power for S-MAC Slot Number 8 Active period 20 ms

S-MAC
protocol against the rest of the protocols, which increasingly is Queue Length 50 Sleep period 180 ms
proportional to the increase in the rate of data flow. The
simulation showed increased power consumption up to four times
within our studied limits.
Fig.3. shows the results of the delivery ratio. This result
proves that the B-MAC protocol has the highest rate of delivery
of data (i.e. 97%) against other protocols. Especially when the
data flow is low and decreases the delivery ratio to 11%, while
increasing the flow of data. The same case with the crankshaft
protocol, but it is reduced by 89% at data flow equal 50 bits/sec
to be up to 32% at data flow equal 250 bits /sec. For L-MAC
protocol, note that the delivery ratio is 53% of data flow equal 50
bits/sec. The reason for this low rate is because the protocol L-
MAC suffers from the need to assign a contention-free slot to all
sensors within the network. To allow such assignment, existing
20 slots are required. Even at low message rates, the L-MAC
protocol does not achieve a perfect delivery ratio. Finally, in
protocols S-MAC and T-MAC, the exchange of messages or data
are only in the active period and are entered in a forced idle state,
which prevents the protocols to respond to increased data flow
Fig.2. The comparative results based on
rate and produces a low delivery ratio. average power consumption.
 ,(((&DQDGLDQ&RQIHUHQFHRI(OHFWULFDODQG&RPSXWHU(QJLQHHULQJ&&(&(

REFERENCES
[1] H. Singh and B. Biswas, "Comparison of CSMA based MAC protocols of
wireless sensor networks," Int. J. AdHoc Netw. Sys., vol. 2, no. 2, 2012.
[2] Khatarkar, S. and Kamble, R.,” Wireless Sensor Network MAC Protocol:
SMAC & TMAC,” Indian J. of Comput. ,Sci. and Eng., vol.4, no.4,
pp.304-310, 2013.
[3] Furat, AL-Obaidy, Hossein Zereshkian, and Farah A. Mohammadi. "An
energy-efficient routing algorithm in ZigBee-based cluster tree Wireless
Sensor Networks," in Proc. IEEE 30th Canadian Conf. of Elect. Comput.
Eng. (CCECE), pp. 1-5. 2017.
[4] Liu, Y., Ota, K., Zhang, K., Ma, M., Xiong, N., Liu, A. and Long,
J.,“QTSAC: An energy-efficient mac protocol for delay minimization in
wireless sensor networks,” J. Trans. IEEE Access, vol. 6, pp.8273-8291,
2018.
[5] Sruthi, R., “ Medium access control protocols for wireless body area
networks: A survey,” J. Procedia Technol., vol. 25, pp. 621-628, 2016.
[6] M. GUNN and S. G. M. KOO, "A Comparative Study of Medium Access
Control Protocols for Wireless Sensor Networks," Int. J. Commun,
Netw.and Syst. Sci., vol. 02, no. 08, pp. 695-703, 2009.
[7] M. Arifuzzaman and M. Matsumoto, "An Efficient Medium Access
Fig.3. The comparative results based on delivery ratio. Control Protocol with Parallel Transmission for Wireless Sensor
Networks," J. Sensor and Actuator Netw., vol. 1, no. 3, pp. 111-122, 2012.
[8] S. Agarwal, V. Jain and K. Goswami, "Energy Efficient MAC Protocols
for Wireless Sensor Network," Int. J. Comput. Sci. Applicat., vol. 4, no. 1,
pp. 153-160, 2014.
[9] M. Khan, A. Alajlan and M. Almasri, "Power Efficient Scheduled-Based
Medium Access Control Protocol over Wireless Sensor Networks," J.
Wireless Sensor Netw., vol. 08, no. 02, pp. 13-23, 2016.
[10] R. Ramya, G. Saravanakumar and S. Ravi, "MAC Protocols for Wireless
Sensor Networks," Indian J. Sci. Techol., vol. 8, no. 34, 2015.
[11] J. Pan, "A survey of network simulation tools: Current status and future
developments," Tech. Rep., Washington University, St. Louis,USA. vol.2,
no.4, 2008.

Fig.4. The comparative results based on latency.

V. CONCLUSIONS
We compared five MAC protocol types for WSN with respect
to their power consumption, delivery ratio, and latency. We have
shown through simulation results that crankshaft protocol
consumes lesser power than all other protocols which were
considered. Further simulations show that the B-MAC protocol
has the highest rate of delivery of data (i.e. 97%) with the lowest
latency (0.1 sec) comparing with the other protocols. In the
future, simulations can be done with a greater number of nodes
and other existing protocols with different parameters.