Scholarly article on topic 'Analysis of the IEEE 802.15.4 MAC Parameters to Achieve Lower Packet Loss Rates'

Analysis of the IEEE 802.15.4 MAC Parameters to Achieve Lower Packet Loss Rates Academic research paper on "Computer and information sciences"

CC BY-NC-ND
0
0
Share paper
Academic journal
Procedia Computer Science
OECD Field of science
Keywords
{WSN / "packet loss" / CSMA/CA / IEEE-802.15.4}

Abstract of research paper on Computer and information sciences, author of scientific article — Imen Bouazzi, Jamila Bhar, Mohamed Atri

Abstract The IEEE-802.15.4 standard utilizes the CSMA-CA mechanism to control nodes access to the channel communication. It is becoming the popular choice for various applications of surveillance and control used in wireless sensor network (WSN). The benefit of this standard is evaluated regarding of the packet loss probability that depends on the configuration of IEEE 802.15.4 MAC parameters and the traffic load. Our exigency is to evaluate the effects of various configurable MAC parameters on the performance of beaconless IEEE 802.15.4 networks, under different traffic loads. Static values of IEEE 802.15.4 MAC parameters (macMinBE, macMaxCSMABackoffs, CW, and macMaxFrame Retries) will be evaluated. To evaluate the performance of the protocol, we use ns-2 network simulator.

Academic research paper on topic "Analysis of the IEEE 802.15.4 MAC Parameters to Achieve Lower Packet Loss Rates"

CrossMark

Available online at www.sciencedirect.com

ScienceDirect

Procedia Computer Science 73 (2015) 443 - 451

The International Conference on Advanced Wireless, Information, and Communication

Technologies (AWICT 2015)

Analysis Of The IEEE 802.15.4 MAC Parameters to Achieve Lower Packet Loss

Imen Bouazzi1, Jamila Bhar, Mohamed Atri Faculty ofScience ofMonastir, University of Monastir

Abstract

The IEEE-802.15.4 standard utilizes the CSMA-CA mechanism to control nodes access to the channel communication. It is becoming the popular choice for various applications of surveillance and control used in wireless sensor network (WSN). The benefit of this standard is evaluated regarding of the packet loss probability that depends on the configuration of IEEE 802.15.4 MAC parameters and the traffic load.

Our exigency is to evaluate the effects of various configurable MAC parameters on the performance of beaconless IEEE 802.15.4 networks, under different traffic loads. Static values of IEEE 802.15.4 MAC parameters (macMinBE, macMaxCSMABackoffs, CW, and macMaxFrame Retries) will be evaluated. To evaluate the performance of the protocol, we use ns-2 network simulator.

©2015 The Authors.PublishedbyElsevierB.V. Thisis an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-reviewunderresponsibilityoforganizingcommitteeofthelnternationalConferenceon Advanced Wireless,Information, andCommunication Technologies(AWICT2015)

Keywords: WSN, packet loss, CSMA/CA, IEEE-802.15.4

*Bouazzi Imen. Tel.: +216 21 760 519; .E-mail address: imen.bouazzi@gmail.com

1877-0509 © 2015 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license (http://creativecommons.Org/licenses/by-nc-nd/4.0/).

Peer-review under responsibility of organizing committee of the International Conference on Advanced Wireless, Information, and Communication Technologies (AWICT 2015) doi: 10.1016/j.procs.2015.12.021

1. Introduction

Advances in technology in WSN have become a major challenge for research, to attract the researchers and engineers of industry. They are used in a growing number of applications such as health [10], the environmental monitoring and surveillance of the house. Sensor networks are intended to support time-critical applications which is an important class of service dealt with the IEEE 802.15.4[2] standard. Control, actuation and the follow-up are all of the examples of applications where information must be delivered within a certain time deadline.

The IEEE 802.15.4 is a standard for short range, low rate-bit and low cost wireless personal area networks. It provides MAC and PHY layers for ZigBee. The IEEE 802.15.4 MAC standard specification describes the individual node behavior. To support time-critical applications, IEEE 802.15.4 offers a Guaranteed Time Slot GTS allocation mechanism at the network coordinator. The packets are transmitted on a super frame basis. Each super frame is divided into Contention Access Period CAP, where nodes contend among each other to send packets, and a Contention Free Period CFP, where nodes have GTSs to send packets without contention. The GTS allocation provides communication services to time critical data. It makes guarantees on packets delivery and delivery times to be transmitted to the network coordinator [3].

The rest of this paper is organized as follows: Section 2 highlights the IEEE 802.15.4 features and its slotted CSMA/CA mechanism. Section 3 presents the proposed differentiation service strategies. Section 4 presents the simulation study and performance evaluation results. Section 5 concludes the paper.

2. Related work

The improvement of CSMA/CA MAC mechanisms has drawn many research efforts. Particularly for the case of the IEEE 802.15.4 protocol, some recent research works have contributed to enhance the slotted CSMA/CA mechanism for achieving reduced (soft) delay guarantees and better reliability oftime critical events, as described next.

In [4], the authors extend a DCF with a new calculation method to increment contention window (CW) that enables each station to access the channel after a small number of attempts. Otherwise, they declare failure case early, if collision still occurs to reduce delay and packet loss ratio and increase the efficiency ofthe transmission channel.

In [5], the authors extend the previous schemes by allowing high priority frames to perform only one Clear Channel Assessment (CCA) operation instead of two, using a frame tailoring strategy, which aims to avoid collisions between data frames and acknowledgment frames when only one CCA is performed. These solutions seem to improve the responsiveness ofhigh priority frames in IEEE 802.15.4 slotted CSMA/CA, but require a non negligible change to the IEEE 802.15.4 MAC protocol to support the priority toning and frame tailoring strategies, thus turning them non-compatible with the standard.

The author in [6] analyzes via simulations the impact of different configurable MAC parameters on the performance ofbeaconless IEEE 802.15.4 networks under different traffic loads and under different levels of interference from the hidden nodes in order to have a good tradeoff between the packet loss probability and the packet latency under different conditions.

In this paper, we investigate other alternatives for improving slotted CSMA/CA without forcing fundamental changes to the MAC protocol. We particularly aim to assess different parameter settings ofthe protocol.

3. IEEE-802.15.4 MAC PROTOCOL

The features of the IEEE-802.15.4 MAC are beacon management, channel access, GTS management, frame validation, acknowledged frame delivery, association, and disassociation. In addition, the IEEE802.15.4 Mac provides hooks for implanting application appropriate security mechanisms.

3.1. superframe structure

The IEEE802.15.4 standard allows the optional use ofa super frame structure. The format ofthe super frame is defined by the coordinator. The super frame is bounded by network beacons sent by the coordinator. Thus, the beacon frame is transmitted in the first slot of each super frame. It is used to synchronize the attached devices. If a coordinator does not wish to use a super frame structure, it will turn off the beacon transmissions.

Optionally, the super frame can have an active and an inactive portion as shown in Fig 1. During the inactive portion, the coordinator may enter a low power mode. The active portion consists of CAP (Contention Access Period) and CFP (Contention Free Period). Any device wishing to communicate during the CAP shall compete with other devices using a slotted CSMA/CA mechanism. CAP portion is divided into 16 equally sized slots. On the other

Beacon

M—CFP-

Active I I I I I I I I

16 timeslots

Beacon —* i t i

Inactive

SD = abasesup erframedurati qn x 2sn symbols

BI - abasesup erframeÜuration x 211" symbols

Fig. 1. Superframe structure, hand, the CFP contains GTS (guaranteed time slots). 3.2. Slotted csma/ca algorithm

The slotted CSMA/CA maintaining three counters in each device for channel access control. Notably, NB is the number of backoff trials for packet transmission. BE is the backoff exponent for generating a random backoff duration for which a device has to wait before attempting carrier sensing, and CW is the value ofthe contention window slots for clear channel assessment (CCA) after the random backoff duration. Tab. 1 shows that range and default value of BE.

Based on the counters, the medium access process ofslotted CSMA/CA for a device is as follows [7]

• The slotted CSMA/CA shall first initialize the NB, the CW, and the BE and then locate the boundary ofthe next backoffperiod. The BE shall be initialized to the value of macMinBE. Ifthe BE subfield in the super frame sets to 1, this value shall be initialized to the lesser of 2 and the value of macMinBE.

• The slotted CSMA/CA shall delay for a random number of complete backoffperiods in the range 0 to 2BE.

• The slotted CSMA/CA request that the PHY perform a CCA then the CCA shall start on a backoffperiod

boundary.

• If the channel is assessed to be busy, the MAC sub layer shall increment both the NB and the BE by 1, ensuring that BE shall be no more than macMaxBE. And it shall also reset CW to 2. If the value of the NB is less than or equal to the macMaxCSMABackoffs, the slotted CSMA/CA algorithm shall return to step (2). If the value ofNB is greater than macMaxCSMABackoffs, the slotted CSMA/CA algorithm shall terminate with a channel access failure status.

• If the channel is assessed to be idle, the MAC sub layer in a slotted CSMA/CA system shall ensure that the contention window has expired before commencing transmission. To do this, the MAC sub layer shall first decrement the CW by one and then determine whether it is equal to O.Ifit is not equal to 0, the slotted CSMA/CA algorithm shall return to step (3). If it is equal to 0, the MAC sub layer shall begin transmission of the frame on the boundary ofthe next backoffperiod.

Table I. Range and default value ofBE.

Attribut Default value range

MacMinBE 3 0-MacMaxBE

MacMaxBE 5 3-8

MacMaxCSMABackoffs 4 0-5

4. SIMULATION SETUP

Simulation announced in this paper make the following assumptions. The IEEE 802.15.4 PHY layer operates in 2.4GHz band and no transmission is lost due to PHY level noise. The IEEE 802.15.4 MAC layer operates in beaconless mode and all packets need MAC level acknowledgement. The CCA is carried out of 18 symbols to ensure that an ACK is never implied in a collision. The simulated network topology consists of a cluster tree topology as shown in Fig.2 where we have defined a high traffic between nodes.

0 0 0 0 ooqsoo&Q

0 ffl 0 @ ® © ® O ® Fig.2. Cluster tree topology

In this section we conduct simulations to compare the performance of IEEE-802.15.4 Mac protocol on packet loss. For the performance parameters, we concentrate on average energy consumption and packet loss probability.

Our simulations are run with NS2 simulator[l]. We list in table 2 the simulation parameters that we have used it in our performance evaluation (some parameters are adopted from the work of [8]).

In table 2, CCA power is the power consumed while sensing the medium during the clear channel assessment states. The network operates under the beacon-enabled mode with the superstructure constituted by the CAP period and the CFP period. The traffic used is assumed to be both saturated (i.e., nodes have always packets to send). In the following sections we show and discuss the results ofour simulations.

5. SIMULATION RESULTS

5.1. Choice of MAC parameters

Collisions and packet loss are a vital consequence in Wireless Sensor Networks. Therefore, nodes can achieve very low duty cycles (if they alternate between sleep and active mode) by extending the length of the super frame. For instance, with the beacon parameters SO=BO=3, the duty cycle is as low as 0. 1%(210) for the coordinator (it needs to stay awake during the active period) and an order of magnitude lower for the devices that only need to wake up for the beacon reception and possibly a CCA, and frame transmission. We have evaluated the impact of different MAC parameters on the IEEE-802.15.4 performance in a cluster tree topology:

• macMaxFrameRetries: number of retransmissions (because no ACK received) before dropping a frame.

• macMaxCSMABackoffs: number of unsuccessful channel sensing before dropping a frame.

• BE: backoff exponent that determines the size ofthe contention window from which a node chooses the value ofthe random backoff interval before sensing.

• CW: Defines the number of consecutive backoffperiods at the end of which the channel must bedetected free before starting transmission.

We have used NS2 as a simulator. We assume that nodes always have a packet to send at the beginning ofeach super frame. Table II presents the simulation parameters.

TABLE II. Simulation Parameters

Average power consumed (mW) RX 40

CCA 40

Sleep 0.8

Duration 1 Timslot 0.32 ms (80 bits)

Frame length (L) 14 timeslots

ACK frame length (LACK) 2 timeslots

Simulation Time 320 s

IEEE802.15.4 settings macMaxCSMABckoffs 5

macMinBE 3

macMaxBE 5

5.2. Adapting Contention Window

In this part we will study the impact ofincreasing the value ofthe Contention Window CW on the performance ofthe protocol. It means that we will see the impact of variable CCA periods on the Efficiency of IEEE 802.15.4 MAC Protocol.

we explore in Fig.3 the impact of CW on the energy consumption. As shown in this figure, when the value of CW increases, the energy increases too. Therefore when a node goes into a sleeping state for a randomly generated time, it will consume less energy. However, if a node listen for two successive CCAs, before attempting to transmit data frame he will consume more energy while listening to the channel.

Fig.3. Energy consumption under different value ofCW

If we interfere CW=3 it means that we will add another checking ofthe channel (3 CCAs). It is equivalent that we have performed in this case a third CCA If CCA2 is sensed busy. This will avoid wasting time in a long backoff period.

If we interfere CW=1 it means that we will eliminate one checking ofthe channel (1 CCA). In this condition, we have a short clear channel assessment. No need to perform 2 CCAs.

We have obtained the following results (Fig.4) by measuring the total number of packets loss for different value of CW. Therefore, the lower value ofCW reduces significantly the packet loss probability. Since the increment of the value of CW at its maximum value decrement the probability of dropped packet.

Fig.4. Packet loss under different value ofcw

Increasing contention window size means that the devices yield an opportunity to occupy the channel prior. Therefore, the probability ofretrying transmissions increases as contention window size increases.

5.3. Impact of macMaxFrameRetries

Fig.5 shows the impact of the parameter macMaxFrameRetries on the packet loss for a cluster tree topology. We have varied the value of this parameter between 1 and 5. where the default value of macMaxCSMABackoffs= 4, macMinBE=3 and macMaxBE=5.

Traffic ioadipktys)

Fig.5. Impact of macMaxFrameRetries on Packet Loss Probability

Increasing the macMaxFrameRetries value authorizes a node to tolerate more CCA failures in a transmission attempt before a channel access failure is declared. It can also helps to reduce the packet loss probability. The drawback of grown macMaxCSMABackoffs values increase packet latency due to increasing CSMA wait times.

5.4. Impact of macMaxCSMABackoffs

As Fig.6 shown, when the numbers of hidden nodes are few, the channel access failures cause most of the packet loss and hence higher macMaxCSMABackoffs value decreases the packet loss probability. With an increase in the number of hidden nodes, collision failure becomes the reason for packet loss and hence higher macMaxCSMABackoffs value increases the packet loss probability. Normally a value of 0 for macMaxCSMABackoffs essentially means that a packet is abandoned as soon as a CCA failure is encountered.

Fig.6. Impact ofMac CSMABackoff

5.5. Impact ofmacMinBE:

Fig.7. Impact ofMacMinBE on the packet loss probability

Fig.7 shows the impact ofincreasing the BE value on packet loss probability as the number of observed nodes. In these simulations, the macMaxCSMABackoffs and macMaxFrameRetries parameters are maintained at their default values (4 and 3 respectively).

Fig.7 shows that at low number of nodes, the increase in BE can significantly reduce the packet loss probability. However, as the number of nodes increases the traffic loads increase too so the increase in the packet loss probability with fall in BE becomes more significant. It means that at very high traffic loads, the packet loss probability becomes very high irrespective ofthe BE value.

It is clearly shown that the increase in the number of nodes has a wasteful impact on the performance irrespective ofthe BE value.

6. CONCLUSION

In this paper, we have analyzed the impact of different mac parameters like macMinBE, macMaxCSMA Backoffs, CW and macMaxFrameRetries on the performance of beaconless IEEE 802.15.4 networks under different traffic load conditions and under different number of nodes. As conclusion we can say that the increase ofthe macMinBE value leads to the growth ofthe range of CSMA wait times. Therefore, competing nodes become less likely to finish their CSMA waits when another node transmitting a packet. Therefore, increasing the macMinBE values leads to reduce in both the probability of collisions as well as CCA failures.

References

1. http://nile.wpi.edu/NS/.

2. IEEE 802.15.4, "Wireless medium access control (MAC) and physical layer (PHY) specifications for low rate wireless personal area networks (LR- WPANS)," Standard, IEEE, 2003.

3. F. Charfi, O. Slama, J-M Thiriet, S. Lesecq, Improving the control performance in Wireless Network Controlled Systems, using the Beacon

mode, Journal ofTelecommunications, Vol 3, Issuel, June 2010.

4. Sedrati Maamarl, Maamri Ramdane2, Bilami Azeddinel, Benmohammed Mohamed2 , Contention Window Optimization: an enhancement to

IEEE 802.11 DCF to improve Quality of Service , International Journal ofDigital Information and Wireless Communications (IJDIWC) 1(1): 273-283 ,2011(ISSN 2225-658X)

5. T. Kim and S. Choi, "Priority-based delay mitigation for event-monitoring IEEE 802.15.4 LR-WPANs," IEEE Communications Letters, Nov.

6. D. Rohma , M. Goyala, H. Hosseinia, A. Divjakb and Y. Bashirb , A simulation based analysis of the impact of IEEE 802.15.4 MAC parameters on the performance under different traffic loads . Mobile Information Systems 5 (2009) 81-99 .

7. T. Kim, D. Lee, J. Ahn, and S. Choi, "Priority toning strategy for fast emergency notification in IEEE 802.15.4 LR-WPAN," in Proceedings of

the 15th Joint Conference on Communications & Information (JCCI), April, 2005.

8. Impact ofVariable CCA Periods on the Efficiency of IEEE 802.15.4 MAC Protocol, Mouhcine Guennoun, March 13, 2014

9. Feng Shu, Student Member, IEEE, Taka Sakurai, Member, IEEE, Moshe Zukerman, Fellow, IEEE, and Hai L. Vu, Senior Member, IEEE

Packet Loss Analysis ofthe IEEE 802.15.4 MAC without Acknowledgements .JANUARY 2007

10. Pham Thanh Hiep , Ryuji Kohno. Optimizing Packet Generation Rate for Multiple Hops WBAN with CSMA/CA Based on IEEE802.15.6. Communications and Network, 2014, 6, 112-123.