Comparison of Csma Based MAC Protocols of Wireless Sensor Networks

International Journal on AdHoc Networking Systems (IJANS) Vol. 2, No. 2, April 2012
DOI : 10.5121/ijans.2012.2202 11
COMPARISON OF CSMA BASED MAC PROTOCOLS
OF WIRELESS SENSOR NETWORKS
Himanshu Singh1
and Bhaskar Biswas2
1
Department of Computer Engineering, IT-BHU, Varanasi, India.
himanshu.singh.cse07@itbhu.ac.in
2
Department of Computer Engineering, IT-BHU, Varanasi, India.
bhaskar.cse@itbhu.ac.in
ABSTRACT
Energy conservation has been an important area of interest in Wireless Sensor networks (WSNs). Medium
Access Control (MAC) protocols play an important role in energy conservation. In this paper, we describe
CSMA based MAC protocols for WSN and analyze the simulation results of these protocols. We
implemented S-MAC, T-MAC, B-MAC, B-MAC+, X-MAC, DMAC and Wise-MAC in TOSSIM, a simulator
which unlike other simulators simulates the same code running on real hardware. Previous surveys mainly
focused on the classification of MAC protocols according to the techniques being used or problem dealt
with and presented a theoretical evaluation of protocols. This paper presents the comparative study of
CSMA based protocols for WSNs, showing which MAC protocol is suitable in a particular environment and
supports the arguments with the simulation results. The comparative study can be used to find the best
suited MAC protocol for wireless sensor networks in different environments.
KEYWORDS
Wireless Sensor Networks, Medium Access Control Protocols, Energy efficiency
1. INTRODUCTION
A 'Wireless Sensor Network' (WSN) can be described as a network of sensors which
communicate with each other wirelessly. These sensors may be installed in an unattended
environment with limited computation and sensing capabilities. Hence, they need to be fault-
tolerant and reliable so that maintenance requirement is less. Since sensors are often deployed in
remote applications like forest-fire monitoring and structural health monitoring, the battery cannot
be replaced frequently due to inaccessibility of sensor nodes. To prolong network lifetime, energy
spending should be minimum. Energy conservation can be done using efficient macro-
programming of WSNs. In [4], we present a study of methods of energy efficient macro-
programming. Another approach is to design energy efficient MAC protocols. Various MAC
protocols have been designed to address this issue.
According to a survey [2] on Wireless Sensor Networks, major sources of energy waste at
medium access communication are a) collision - which requires re-transmission of collided

12
packets, b) overhearing - where a node receives a message meant for another node, c) control
packet overhead - where energy is consumed in exchange of control packets used for control data
transmission and d) idle listening - which means that node is listening to idle channel and then
over-emitting by sending packets when the destination node is not yet ready.
Among all reasons mentioned above idle listening is a major cause of energy waste. So it is
important to design a suitable MAC protocol which can reduce or prevent above energy wastes.
There are four techniques to avoid idle listening - static sleep scheduling, dynamic sleep
scheduling, preamble sampling, and off-line scheduling. Based on these techniques, many MAC
protocols based on CSMA, TDMA, hybrid and cross-layer optimizations were introduced. We
classify all the famous MAC protocols for WSNs in Table 1.
Table 1 Classification of Medium Access Control Protocols
The MAC protocols behave differently under different network scenarios with respect to energy
consumption and throughput. So there is a need of an efficient comparative study of these
protocols. In this work, we evaluate CSMA based MAC protocols. The MAC protocols were
implemented on TOSSIM [6]. Unlike other simulators (OMNet++ and NS-2), TOSSIM simulates
exactly the same code which is going to be executed on real hardware, thus narrowing the gap
between simulation and real network deployments. Each MAC protocol is then integrated with
PowerTOSSIM-Z [7], a power modelling tool, to measure the energy consumption. In Section 2 to
Section 9, we describe the implemented MAC protocols, along with their pros and cons supported
by simulation results. Section 10 describes the simulation setup and the results. Section 11
concludes the paper highlighting our future aims and scope.
2. S-MAC
Sensor-MAC (SMAC) [5] is a contention-based protocol that regulates sleep periods in a sensor
network to conserve energy and improve network lifetime. This protocol represents the baseline
of sleep-oriented, energy-efficient WSN MAC protocol designs. Out of four techniques for
avoiding idle listening: static sleep scheduling, dynamic sleep scheduling, preamble sampling, and
off-line scheduling, SMAC adopts static sleep scheduling for preserving the energy. SMAC
divides the time into frames. Every frame is divided into an active and a sleep period as shown in
Figure 1. In active period, the transmitter-receiver is switched on and it is switched off during
sleep period. The active period is further divided into Time Synchronization period and data
transfer period.

13
Figure 1: Sleep and wake-up cycles in SMAC and TMAC
Time Synchronization is required so that receiver remains awake when sender sends the message.
In Time Synchronization Period, first step in setting the sleep schedule for a node is to listen for a
SYNC packet from a neighbour. The SYNC packet contains the sleep schedule and indicates that
the sender is going to sleep after ‘t’ seconds. Once the node receives its neighbour’s sleeping
schedule, it adopts that schedule and re-transmits the schedule for other neighbouring nodes to
adopt. If a node does not receive a SYNC packet within a pre-decided timeout period, the node
will set and broadcast its own schedule. Border nodes (nodes between two active schedules) may
receive two different schedules from different nodes. They may either adopt both or one of the
schedules. We implemented SMAC where border nodes follow both the schedules. Due to this
mechanism, the network gets divided into multiple virtual clusters, each cluster surrounded by
border nodes. Each node within a cluster follows same sleep schedule, whereas border nodes
follow schedule of both its neighbouring clusters. Hence, border nodes remain awake for a larger
period, thus increasing the energy consumption. However, in every cycle border nodes and virtual
clusters keep changing. So virtual clustering does not affect the network lifetime as a whole.
In SMAC, data exchange takes place in data transfer period through RTS-CTS-DATA-ACK
handshaking for unicast communication. A node may extend its active duration if data exchange
doesn't finish in active period. However, even if data exchange finishes within active period, the
node will still remain awake until its sleep time thus wasting energy. When a node sends a RTS
(Request to Send) or CTS (Clear to Send), it puts the duration of data transmission in RTS or CTS
packets. Neighbouring nodes which overhear these RTS or CTS set an NAV (Network allocation
vector) interrupt timer based on the duration and go to sleep since they cannot communicate in
this duration due neighbour's disturbance. When NAV interrupt fires, these neighbouring nodes
again start following their normal schedule. This handshaking technique not only reduces collision
but also saves energy by overhearing avoidance. Figure 5 demonstrates this observation as the
energy consumption is low at higher traffic and increases with increase in traffic. This is because
neighbour nodes sleep for longer time on overhearing RTS or CTS, frequency of which is higher
in high traffic. However, such conservation is not achieved in Broadcast messages, since
broadcasting does not use RTS-CTS handshaking.
2. T-MAC
Timeout MAC (TMAC) [8] is also a contention-based, MAC layer protocol that is based upon the
basic features of SMAC in optimizing power efficiency by sleeping during periodic network
inactivity. However, unlike SMAC, TMAC follows dynamic sleep schedule. The TMAC protocol
introduces an active timeout mechanism that decreases the idle listening overhead by dynamically

14
adjusting the active period according to network traffic loads. TMAC allows the nodes to sleep
after sometime when all network traffic has completed, as explained in Figure 1. The end of
traffic is signalled after monitoring an idle channel for an adaptive timeout (TA) period. If no
activity occurs for this TA time duration, node switches off its radio and goes to sleep state. The
TA period should be large enough to overcome the early sleeping problem (a node goes to sleep
state when a neighbour still has packets to be sent). Such procedure makes TMAC more energy
savvy than SMAC as evident from Figure 5.
The TA duration depends on contention interval, length of RTS packet and Turnaround time. We
find the appropriate TA time for our implementation as shown in Figure 2.
Figure 2. Effect of variation of TA time on delivery ratio and energy consumption
The graph presents the trade-off analysis between energy consumed and delivery ratio, as TA is
varied. The non-zero y intercept of graph shows that an appreciable amount of energy is
consumed even when there is no data traffic. This energy is consumed during time
synchronization period. So, Synchronization is a large overhead in SMAC and TMAC.
TMAC also introduces a FRTS (Future request to send) mechanism and full buffer priority, to
avoid early sleeping problem for converging type of data communication. When a node which has
a data to send overhears a CTS packet, it broadcasts a FRTS. The duration of the data is stored in
FRTS packets. The recipient of FRTS sets its NAV and goes to sleep. After the communication,
node again wakes up to receive the data the sender of FRTS. In full buffer priority, when the
sending buffer of a node is full and it receives a RTS from some node, then instead of replying
with CTS, node transmits its own RTS thus taking the priority. When it has completed the data
sending, only then it replies with CTS to the original RTS request it received. So, this mechanism
also introduces the flow control in the data flow. It should be noted that Full buffer priority should
be employed only in converging type of mechanism and not in ad-hoc type of communication.
4. D-MAC
D-MAC [8] is a protocol which aims at real-time delivery of data, still being energy efficient. It
adopts a staggered wake-up pattern to forward the data packets to the base-station as shown in
Figure 3. Nodes are considered to be present at different levels of data-gathering tree.
Figure 3. Active and Sleep Period of DMAC

15
All the nodes at one level would wakeup simultaneously to receive the data. This receiving period,
µ, is followed by the transmitting period (µ) in which they forward data to higher level. The nodes
at next level wake-up just after the receiving period of the lower level. So, active period is a
staggered wake-up pattern, where active period of one level partially overlaps with that of lower
level as shown in Figure 3. Due to such staggered wake-up pattern, a data packet reaches from
root to leaves in one cycle only, thus minimizing latency. D-MAC adopts data prediction method
when multiple children need to send data to one parent in one cycle only. In data prediction, if a
parent receives data, it again wakes-up after 3u, hoping that there will be data from another child
as well. When same child needs to send multiple packets to same or different parent, a More-To-
Send (MTS) flag is piggybacked in data-packet, so the parent will keep waking-up every after 3µ
time, until it receives MTS flag set in last packet.
In-spite of using data-prediction and MTS flag, D-MAC is not suitable for high traffic load due to
small µ time. D-MAC does not use RTS-CTS handshaking, because at a given time only few
nodes of the network will remain active, thus reducing the chances of collision. Also, data
aggregation is possible at each node because the parent can receive packets from all children
before forwarding them. D-MAC requires local and efficient synchronization. Due to staggered
wake-up schedule, each node should know its depth-level. We implemented D-MAC where the
synchronization packet sinks down from root to leaf nodes informing each node about their depth
level. Another disadvantage is that D-MAC cannot be used to local-gossip type of
communication due to its hierarchical data forwarding design.
5. B-MAC
Berkley Media Access Control (BMAC) [9] is a MAC level protocol for Wireless Sensor
Networks which uses adaptive preamble sampling scheme. This technique consists of sampling
the medium at fixed time intervals. Figure 4 describes the working of BMAC. Sampling the
medium means to listen to the channel for some activity. In this scheme, every node samples the
medium at fixed intervals to check whether any node is willing to communicate. If any node has a
packet to send, node (sender) sense the medium if it is free, takes a small back-off and then sends
a long wake up preamble followed by data packet. Preamble is not a packet but a physical layer
RF pulse just greater than sampling period in length so that node sampling the medium notices
this activity. However, for simplicity, we have considered the preamble as a long packet in our
simulation. When receiver wakes up, it senses the medium and if it detects any noise (preamble),
it turns on its radio and waits for the preamble to end. On completion of preamble, if data packet
is destined to the node itself, it receives full data packet otherwise ignores the packet and goes to
sleep.
Figure 4. Preamble in BMAC, BMAC+ and XMAC

16
The goals of BMAC protocol are low power operation, effective collision avoidance and efficient
channel utilization at low as well as high data rate. BMAC can be scaled to a large network. It is
re-configurable by networks and its implementation is simple and requires small RAM size.
BMAC protocol uses concepts of media access functionality. It uses clear channel assessment
(CCA) and back offs for channel arbitration, acknowledgments for reliability, and low power
listening (LPL) for low power communication. B-MAC is only a link protocol, with network
services like organization, synchronization, and routing built above its implementation but BMAC
is unable to provide multi-packet mechanisms like hidden terminal support, message
fragmentation and particular low power policy.
6. B-MAC+
BMAC+ [10] is an extension of BMAC [9] protocol. BMAC+ tries to reduce waste of energy due
to long preamble of BMAC. A Preamble is the sequence of bits which does not contain any
relevant information and is used to tell receiver that some node wants to communicate. The basic
idea of BMAC+ is to replace wake up preamble with small numbers of blocks containing some
information as shown in Figure 4. This information contains address of destination node and
number of remaining blocks or countdown of data starting with highest number according to
preamble length needed. The destination address is used to avoid overhearing without receiving
remaining preamble blocks. The number of remaining blocks or countdown of sequence number
starting from zero is used to avoid idle listening by the nodes which are not recipients of data.
BMAC+ saves more energy than BMAC with same latency and throughput as evident from
simulation results in Figure 6. This is because in BMAC+, when receiver receives early preamble
it can turn its radio off for remaining preamble blocks and wait for time of data arrival to turn
radio on. It would again wake up when data arrives.
7. X-MAC
Standard MAC protocols such as BMAC use long preamble before data to wake up receiver. This
preamble scheme was enhanced in BMAC+ to reduce power consumption. In BMAC receiver
turns its radio off after getting preamble but still sender continue to send remaining part of
preamble which results in waste of energy and also introduces excess latency.
In 2006, these problems were solved when XMAC [11], a low power MAC protocol was
introduced. XMAC introduces a shortened preamble approach that retains the benefits of low
power listening such as low power communication, simplicity and decoupling of transmitter-
receiver sleep schedules. XMAC introduces a series of short preamble packets, each packet
containing the destination address and remaining number of preambles. A node sends these series
of preambles at fixed intervals as shown in Figure 4. This interval is long enough to get a response
from receiver. When receiver receives any preamble, it at once sends acknowledgement to sender
during fixed interval between preamble packets. When sender receives an acknowledgement, it
stops sending further preambles and immediately sends data packet. This reduces both energy at
both receiver and sender sides and also reduces latency per hop.
In addition, when a sender which is waiting for a clear channel to send data, detects a preamble
and then hears an acknowledgement from a node to which sender itself wants to send data, it takes
a random back-off which is long enough to complete data transmission currently going on. After
completion of this transmission, it sends data directly without any preamble. The randomized
back-off is necessary to avoid collision if more than one node is trying to send data at same time.
Since this technique requires receiver node to remain awake to receive data in next cycle as well,

17
every receiver in XMAC remains awake for a short period in case there are additional nodes
willing to send data. These two techniques reduce a lot of energy consumption and also reduce
latency.
7. WISE-MAC
WiseMAC [12] is a medium access control protocol for WSNs which is based on non-persistent
CSMA and uses preamble sampling technique to reduce power consumption. WiseMAC tries to
use minimum sized wake preamble. WiseMAC requires no set-up signalling, no network-wide
synchronization and is adaptive to traffic load. WiseMAC like BMAC is based on sampling
technique where a node listens to the channel for a short duration. All sensor nodes sense medium
at the same constant period ‘TW’ independently. Here 'independent' means that they may sense
medium at different time but they sample the medium for same period. In BMAC, if the medium
is found busy, the node continuously listens to the medium until data is received or until medium
becomes idle again. At the sender side, wake-up preamble of a size equal to the sampling period is
added in front of every data frame so that receiver will wake up at time of arrival of data packet.
This protocol provides best result when medium is idle but disadvantage of this protocol is that
long wake-up preambles cause a throughput limitation and large power consumption overhead in
transmission and reception. The main idea of WiseMAC is learning the sampling schedules of
direct neighbours of a node. These schedules are used to minimize size of preambles. To recover
packet losses, a link level acknowledgement is used in WiseMAC. The WiseMAC ACK packets
are not only used to carry acknowledgement information but also to inform other nodes (including
sender) the remaining time of next sampling. These other nodes store this time in their tables.
Using this information, a node transmits a packet with minimized size of preamble. The duration
of the wake-up preamble covers the small potential clock drift between the clock at source and
destination. WiseMAC uses following equality for calculating the minimum preamble. TP =min
(4θL, TW), where θ is the frequency tolerance of time base quartz and L the interval between
communications. L is also updated when a node gets ACK of its neighbours at Link level. The
first communication between two nodes is always done using a long wake-up preamble equal to
TW. Once some timing information is found, a wake up preamble of reduced size is used. Since
preamble is proportional to the interval L between communications, it will become small when
traffic is high. This makes WiseMAC adaptive to the traffic. The packet overhead decreases as per
increase in traffic. In the low traffic conditions, the packet overhead is high, but average power
consumption due to this overhead is low. Another important thing about WiseMAC is that it uses
a more bit present in the header of data packet like IEEE 802.11 power save protocol. When this
bit is set to 1, it indicates that more data is coming for same node. So, receiver continues checking
the medium even after sending the acknowledgement. We have simulated WiseMAC in
PowerTossim-Z and simulation results are shown in Figure 6. In figure we can see that energy
consumption in WiseMAC is less than any other MAC protocol but greater than XMAC. Reason
for this is that even though XMAC does not use minimized preamble but in XMAC all
neighbouring nodes turn off their radio after receiving any single preamble while in WiseMAC,
all neighbouring nodes also receive full preamble with actual receiver.
Thus sender’s and receiver’s energy may be saved but including all neighbouring nodes total
energy is more than that of XMAC. Still WiseMAC performs better than others and may be more
usable in low traffic conditions.
7. SIMULATION RESULTS
We evaluated all the above described MAC protocols over the parameters: delivery ratio, inter-
arrival time, energy consumption and no. of hops. After implementation, protocols were

18
integrated with PowerTOSSIM-Z [7] for measuring the energy consumption. Then each protocol
was configured to give the delivery-ratio in the range 85-100%. For such a configuration energy
consumption was evaluated. Note that, we did not integrate the MAC protocols with network
layer, which uses Collection Tree Protocol (CTP) [13] due to conflict in beacon sending period of
CTP and sleep periods of MACs. So a hard-coded hierarchical routing algorithm was manually
written at the application layer for the evaluation purposes. A graph between average energy of
each protocol and inter-arrival time is shown in Figure 5.
Figure 5. Energy consumption vs. Inter-arrival time for converge-cast communication
Figure 6. Closer view of analysis in Figure 5

19
A clean view of dense lines in figure 5 can be seen in figure 6.
In SMAC and TMAC, energy savings at high traffic is due to Overhearing avoidance. More
neighbour nodes sleep for more time on hearing RTS and CTS when traffic is high, at low traffic
they do not hear RTS/CTS thus wasting energy in idle listening during fixed duty cycle. Such
observation is not seen in BMAC, XMAC, WiseMAC and DMAC. Thus, the order of energy
consumption is:
XMAC<WiseMAC<BMAC+<BMAC<DMAC<TMAC<SMAC.
In this order if DMAC knows the levels of its node then DMAC can perform better than any MAC
protocol. For knowing the level of nodes in DMAC, we need to pass a packet say SYNC packet to
all nodes, which consume a lot of energy. In future, the work can be done on this part of DMAC.
Some of the protocols were tested for local gossip type of communication.
Figure 7. Energy Consumption for local gossip type of communication
Results shown in Figure 7 reveal that the energy consumption of TMAC was higher than that of
SMAC for inter-arrival time 1 sec. This can attributed to the fact that TMAC can dynamically
vary its active time according to data traffic. For high traffic rate, the active time (TA Time) may
even become larger than that of SMAC. So energy consumption may be larger than SMAC for
high data traffic, but decreases as data traffic decreases.
6. CONCLUSION
In this paper, we compared CSMA based MAC protocols with respect to their energy
consumption and found that in general, X-MAC performs better than all other protocols which
were considered. Protocols based on preamble sampling consume lesser energy than protocols
based on static or dynamic sleep schedule. The paper also presented the advantages and
disadvantages of these protocols when traffic is high and when it is low. Such analysis may be
used to configure the network as per user requirements. In future, we aim to present a system with
which users would be able to do such configurations using SQL like queries. We aim to integrate
the implemented protocols with TinyDB2 [14], a query driven data extraction system for WSNs.
The system would be integrated with PowerTossim-Z to enable users see the energy consumed by
their application using SQL like queries without exposing them to internal details of sensor
network platforms.

20
REFERENCES
[1] Li Deliang, Peng Fei, Energy-efficient MAC protocols for Wireless Sensor Networks, Retrieved on
16th Oct, 2011. http://www.zj.chinaunicom.com/upload/1253688110322.pdf
[2] I. Demirkol, C. E., alagoz, Mac protocols for wireless sensor networks: A survey. IEEE Commun.
Mag. vol.06 (April 2006), pp.115–121.
[3] Abdelmalik Bachir, Mischa Dohler, Mac essentials for wireless sensor networks. Communications
Surveys and Tutorials, IEEE vol.12 (2nd quarter, 2010), pp.222–248.
[4] Himanshu Singh, R.S. Singh, Configurable Task Mapping for Multiple objectives in Macro-
programming of Wireless Sensor Networks, International Journal of Advanced Smart Sensor
Networks, Volume 1, Number 1, April 2011.
[5] Wei Ye, John Heidemann, D. E. An energy-efficient mac protocol for wireless sensor networks.
INFOCOM, New York, NY, USA IEEE vol.3 (June 2002), pp.1567–1576.
[6] Philip Levis, Nelson Lee, Matt Welsh, David Culler, TOSSIM: Accurate and Scalable Simulation of
Entire TinyOS Applications. In Proceedings of the First ACM Conference on Embedded Networked
Sensor Systems (SenSys), 2003.
[7] Enrico Perla, Art Ó Catháin and Ricardo Simon Carbajo., PowerTOSSIM z: Realistic Energy
Modelling for Wireless Sensor Network Environments. 3rd ACM workshop on Performance
monitoring and measurement of heterogeneous wireless and wired networks, ACM, 2008.
[8] Gang Lu, Bhaskar Krishnamachari, C. S. R. An adaptive energy-efficient and low-latency Mac for
data gathering in wireless sensor networks, 18th International Parallel and Distributed Processing
Symposium (IPDPS04) vol.18 (April 2004), pp.224.
[9] Joseph Polastre, Jason Hill, D. C., Versatile low power media access for wireless sensor networks.
Second ACM Conference on Embedded Networked Sensor Systems (SenSys). Baltimore, MD, USA:
ACM Press vol. 2nd (November 3-5 2004), pp.95–107.
[10] Marco Avvenuti, Paolo Corsini, P. M., Vecchio, A. Increasing the efficiency of preamble sampling
protocols for wireless sensor networks. Mobile Computing and Wireless Communication
International Conference, MCWC. vol. 1st (Sept. 17-20, 2006), pp. 117–122. 11
[11] M. Buettner, E. Yee, G. V. A., and Han, X-mac: A short preamble mac protocol for duty-cycled
wireless sensor networks, International conference on Embedded networked sensor systems (SenSys)
vol. 4th (2006).
[12] El-Hoiydi, A., Decotignie, J.-D, Wisemac: An ultra-low power mac protocol for the multihop wireless
sensor networks. Lecture Notes in Computer Science(LNCS) Vol.3121 (2004), pp.18–31.
[13] Omprakash Gnawali, Rodrigo Fonseca, Kyle Jamieson, David Moss, and Philip Levis. Collection
Tree Protocol. Proceedings of the 7th ACM Conference on Embedded Networked Sensor Systems
(SenSys), 2009.
[14] Himanshu Singh, Bhaskar Biswas, TinyDB2: Porting a query driven data extraction system to
TinyOS2.x, Third International conference of Wireless and Mobile Networks, Trends in Network and
Communications in Computer and Information Science, Springer, 2011, Volume 197, Part 2, 298-
306.
Authors
Himanshu Singh is a postgraduate student at Institute of Technology, B.H.U, India. He
completed his Bachelors in Technology in Computer Engineering from Institute of
Technology, B.H.U, India in 2011. His research interests include Computer Networks and
Wireless Sensor Networks.
Bhaskar Biswas is an Assistant Professor at Department of Computer Engineering,
Institute of Technology, Banaras Hindu University. He completed his PhD from IT-
BHU, in 2011. His research interests include Database Systems, Data Mining and Web
Mining.

Comparison of Csma Based MAC Protocols of Wireless Sensor Networks

More Related Content

Similar to Comparison of Csma Based MAC Protocols of Wireless Sensor Networks (20)

More from pijans (20)

Recently uploaded (20)

Comparison of Csma Based MAC Protocols of Wireless Sensor Networks