Energy Aware Routing in IEEE 802.11s Wireless Mesh Networks

Maria Zogkou

, Aggeliki Sgora

and Dimitrios D. Vergados

Department of Informatics, University of Piraeus,80 Karaoli & Dimitriou St., GR 18534, Piraeus, Greece

VTT Technical Research Centre of Finland, FI 90571, Oulu, Finland

Keywords:

IEEE 802.11s WMNs, HWMP, Airtime Link Metric, Energy Aware Routing, QoS.

Abstract:

Wireless Mesh Networking is a continuous growing technology that can be used used for several application

scenarios, such as military tactical operations, etc. in next generation wireless networks. The IEEE 802.11s

Standard deﬁnes the procedures that wireless nodes follow in order to interconnect and create a Wireless Local

Area Network (WLAN) mesh network. It, also, deﬁnes the routing protocol and the metric that are used by

a IEEE 802.11s mesh network to route data. However, although the energy consumption of mesh nodes is a

crucial parameter for the network’s lifetime in speciﬁc purpose operations (e.g. military and health) the default

metric proposed by the standard doesn ’t take into account the energy of the nodes. In this paper, a new energy

- aware routing metric for the IEEE 802.11s mesh networks has been implemented. Simulation results showed

that the proposed metric prolongs the lifetime of a WMN in comparison with the the default metric used by

IEEE 802.11s Standard while causing a little higher total delay in the network.

1 INTRODUCTION

Wireless Mesh networking is a promising solution for

next generation wireless networks that envisages sup-

plementing wired infrastructure with a wireless back-

bone for providing Internet connectivity to mobile

nodes (MNs) or users in residential areas and ofﬁces,

and could be called the Web-in-the-sky (Nandiraju

et al., 2007). Wireless Mesh Networks (WMNs) at-

tract more and more the attention of the research com-

munity due to their low cost implementation and ro-

bustness. Mesh networking is considered the most ap-

propriate technology for Military/Government wire-

less networks (Shyy, 2006). However, in these net-

works several factors should be taken into acount,

such as mobility, QoS support, power management

etc.

IEEE developed an amendment of the IEEE

802.11 Standard, namely the IEEE 802.11s Standard

(802.11s 2011/D12.0, 2011), for mesh networking

in Wireless Local Area Networks (WLANs). The

aforementioned Standard deﬁnes the Hybrid Wireless

Mesh Protocol (HWMP) as the path selection proto-

col used by IEEE 802.11s WMNs. The default metric

of HWMP is the Airtime Link Metric (ALM) which

does not take into account the energy of the nodes in

the network and thus, the routing mechanism is not

energy - efﬁcient.

In this paper, we deal with problems caused by

Figure 1: An IEEE 802.11s Architecture.

odes that suffer from energy exhaustion and we pro-

pose an energy aware routing metric that can ad-

dress connectivity problems in IEEE 802.11s net-

works. This energy aware metric is used by HWMP

routing protocol instead of the airtime link metric. In

particular, the proposed metric takes into account the

residual energy of nodes and selects the route that is

made up by nodes with the maximum residual energy

in the mesh network. Thus, nodes with low energy

levels do not participate in the routing procedure and

remain operational for longer periods of time. This

results in the prolonging of the network lifetime.

The rest of the paper is organized as follows: The

basic components of an IEEE 802.11s WMN archi-

215

Zogkou M., Sgora A. and Vergados D..

Energy Aware Routing in IEEE 802.11s Wireless Mesh Networks.

DOI: 10.5220/0004634302150220

In Proceedings of the 10th International Conference on Signal Processing and Multimedia Applications and 10th International Conference on Wireless

Information Networks and Systems (WINSYS-2013), pages 215-220

ISBN: 978-989-8565-74-7

 2013 SCITEPRESS (Science and Technology Publications, Lda.)

tecture are presented in Section 2, while Section 3 de-

scribes brieﬂy the HWMP, deﬁned in IEEE 802.11s

Standard. Section 4 overviews several existing rout-

ing metrics in WMNs, while in Section 5 the proposed

energy - aware routing metric is presented. Simula-

tion results are discussed in Section 6. Finally, Sec-

tion 7 concludes the paper.

2 THE IEEE 802.11s

ARCHITECTURE

The basic component of an IEEE 802.11s architec-

ture is the mesh station (STA), i.e an autonomous sta-

tion that implements the mesh functionalities, such

as formation of the mesh BSS, path selection, and

forwarding. The mesh facilities are available to the

mesh STAs that belong to a Mesh Basic Service Set

(MBSS).

Mesh STAs within the MBSS can communicate

with each other either directly, or through other mesh

STAs. Communication with nodes outside the MBSS

can be achieved through the Distribution System

(DS).

A mesh STA that can also provide access to one

or more DSs, via the wireless medium (WM) for the

MBSS is called mesh gate. Once an MBSS contains

a mesh gate that connects it to the IEEE 802.11 DS,

the MBSS can be integrated with other infrastructure

Basic Service Sets (BSSs) too, given that their Access

Points (APs) connect to the same DS.

When a MBSS accesses the IEEE 802.11 DS

through its mesh gate, the MBSS can be integrated

with a non-IEEE-802.11 LAN. Whereas the por-

tal integrates the IEEE 802.11 architecture with a

non-IEEE-802.11 LAN, the mesh gate integrates the

MBSS with the IEEE 802.11 DS.

Figure 1 depicts an IEEE 802.11s Architecture.

3 THE HYBRID WIRELESS

MESH PROTOCOL (HWMP)

The default routing protocol deﬁned in IEEE 802.11s

Speciﬁcation is the HWMP, which is operated on the

Data Link Layer and therefore it uses the MAC ad-

dresses for routing. Also, the speciﬁcation deﬁnes the

ALM as the default metric used by the HWMP for

path selection. The ALM reﬂects the total channel re-

sources consumed by the transmission of a frame over

a particular link and is given by the following equa-

tion.

O+

1− e

(1)

where, O and B

are constants listed in Table I and the

input parameters r and e

are the data rate in Mbps

and the frame error rate for the test frame size B

re-

spectively.

Table 1: Airtime Cost Constants.

Parameter Recommended

Value

Description

O Varies de-

pending on

the PHY

Channel access over-

head, which includes

frame headers, train-

ing sequences, access

protocol frames, etc.

8192 Number of bits in test

frame

HWMP operates in two different routing modes:

the “on - demand” and the “tree-based proactive”

modes.

In the “on - demand” mode, whenevera mesh STA

needs to establish a routing path to a destination mesh

STA, it broadcasts a Path Request (PREQ) message to

all its neighbors having speciﬁed the sequence num-

ber, the address of the destination mesh STA, and the

ALM metric initialized to the initial value of the ac-

tive path selection metric. Upon receiving a PREQ,

if the node that received the PREQ is not the destina-

tion mesh STA acts as follows: First, it creates a path

to the source mesh STA if no path exists, or it up-

dates its current path in two cases: 1) if the sequence

number of the PREQ is greater than the last one; 2) if

the sequence number is the same with its current path

but propagates a better metric for the path. Then, it

calculates its routing metric and updates the PREQ

with the cumulative metric. The updated PREQ is

forwarded to all its neighbors. This procedure is re-

peated until the PREQ reaches its destination. When

the destination mesh STA receives the PREQ, it cre-

ates or updates its routing path to the source node and

it sends unicast Path Reply (PREP) message back to

the source mesh STA.

In the “tree-based proactive” mode, there are two

mechanisms, namely, the proactive PREQ and the

proactive RANN (Root ANNouncement). The proac-

tive PREQ mechanism creates paths from the mesh

STAs to the root, using only group-addressed com-

munication. The RANN mechanism creates paths be-

tween the root and each mesh STA using acknowl-

edged communication.

It should be noted that on-demand and proactive

modes can be used concurrently, because the proac-

tive modes are extensions of the on-demand mode.

WINSYS2013-InternationalConferenceonWirelessInformationNetworksandSystems

216

4 WMNS’ ROUTING METRICS

Although the ALM takes into account the transmis-

sion rate, as well as, the transmission error rate,

since it does not deﬁne any load mechanism it can

route trafﬁc to congested areas (Islam et al., 2010).

Also, the fact that the ALM is unaware of intra-ﬂow

interference has a signiﬁcant effect on the network

performance in Multi-Channel Multi-Radio WMNs

(Bin Ngadi et al., 2012). Furthermore, issues such

as energy and security are not considered. However,

although extensive research carried out for the design

of new metrics very few works have focused on the

path selection mechanism for IEEE 802.11s WMNs.

One of the ﬁrst metrics designed for WMNs is the

Expected Transmission Count (ETX) metric (Camp-

ista et al., 2008) that calculates the expected num-

ber of transmissions required to successfully send

a packet over the link, including retransmissions.

However, ETX does not take into account either the

packet size or the bandwidth of the links. There-

fore, it increases the control overhead, which results

in low network performance. In order to confront the

aboveshortcomings, the Expected Transmission Time

(ETT) (Draves et al., 2004) has been introduced that

considers also and the throughput into its calculation.

Koksal and Balakrishnan (2006) also proposed two

variations of the ETX metric that estimate the losses

by means of the bit error probability: the modiﬁed

Expected Number of Transmissions (mETX), and the

Effective Number of Transmissions (ENT).

Mhatre et al. (2007) proposed the Expected

Throughput (ETP) metric, which measures the ex-

pected throughput of the link. Also, Passos et al.

(2006) proposed the Minimum Loss (ML) metric that

calculates the delivery ratio and selects the route with

the lowest end - to - end loss probability by multiply-

ing the delivery ratios of the links across the path. Es-

pecially, for IEEE 802.11s WMNs Islam et al. (2010)

proposed the Expected Forwarding Time (EFT). EFT

calculates the end - to - end delay that a packet ex-

periences on its way to the destination and selects the

path that has the lowest end - to - end delay to route

packets.

Furthermore, for multi-channels WMNs also sev-

eral metrics have been proposed. Draves et al. (2004)

proposed the Weighted Cumulative ETT (WCETT)

metric that tries to ﬁnd paths with less intra-ﬂow in-

terference and channel diversity. However, WCETT

does not guarantee the selection of the shortest path

and does not take into account inter - ﬂow interfer-

ence. Thus, the selected routing paths may suffer

from congestion. Another metric, that aims to limit

the interference levels on a WMN, is the Metric of In-

terference and Channel-switching (MIC) (Yang et al.,

2005). MIC is based on the ETT metric and takes

into account the inter - ﬂow interference by calculat-

ing the number of nodes that interfere. Subramanian

et al. (2006) proposed the iAWARE metric that com-

bines the interference ratio (IR) with the ETT metric.

More speciﬁc, the interference experienced by links

is given by the fraction of ETT, calculated for a link,

to the interference ratio of the same link. The Signal

to Noise Ratio (SNR) and Signal to Interference and

Noise Ratio (SINR) are used for the calculation of the

interference variations.

All the metrics mentioned above do not consider

energy constraints. The Expected Transmission En-

ergy (ETE) (Jin et al., 2011) was introduced for im-

plementation in wireless mesh sensor networks (WM-

SNs). The aforementioned metric takes into account

the energy distribution in the network aiming to ex-

amine the routes that are selected when ETE metric is

in use. In addition, a threshold has been introduced

aiming in avoiding the calculation of the metric by

nodes with residual energy below the threshold. The

calculation of the metric is given below.

′



∑

i=1

init

100E



1− e

∑

i=1

init

(2)

Where, E

is the residual energy of node i after the

completion of the transmission, E

is the energy con-

sumption of node i, n is the number of nodes that con-

sist the network and n

is the number of nodes along

the route that has been selected. The O

and O

are

constants used in older versions of the 802.11s Speci-

ﬁcation to describe the channel and the protocol over-

head, respectively.

Table 2 gathers the above metrics that have been

proposed for use in WMNs.

5 THE PROPOSED METRIC

The proposed metric takes into account the residual

energy of a node by calculating the total energy con-

sumed by a node wheneverit is in one of the following

states:

• IDLE: During this state the nodeis idle.

• CCA

BUSY: During this state the node is busy.

• TX: During this state the node only transmits

packets.

• RX: During this state the node only receivespack-

ets.

• SWITCHING: During this state the node switches

from one channel to another.

EnergyAwareRoutinginIEEE802.11sWirelessMeshNetworks

217

Table 2: Existing Routing Metrics.

Metric Path Selection Criterion

EFT end - to - end delay for all routing

paths

ETX forward and reverse delivery ratios

of the link

ETT forward and reverse delivery ratios

of the link, throughput

ML packet delivery ratio, end-to-end

loss probability

WCETT end - to - end delay and channel di-

versity

MIC forward and reverse delivery ratios

of the link, throughput, inter -ﬂow

interference

ETP expected throughput of the link

mETX losses by means of the bit error

probability

ENT losses by means of the bit error

probability

iAWARE intrerference ratio, forward and re-

verse delivery ratios of the link,

throughput

ETE transmission rate, transmission er-

ror rate, energy consumption rate of

nodes in the network

More speciﬁcally, for each node n

the residual en-

ergy R(n

) is computed as:

R(n

) = E

current

) − E

con

) (3)

where E

current

) and E

con

) denote the current en-

ergy of the node n

and the energy consumed by the

node n

, respectively. At the beginning, the current

energy is set to the initial energy of the node.

Since the consumed energy of each node n

de-

pends on the state that the node is (i.e. IDLE, TX,

RX, CCA

BUSY, SWITCHING), we use the follow-

ing equation to determine its consumed energy:

con

) = Current(n

) ∗Voltage(n

) ∗ Duration (4)

where Current(n

) is the current in Ampere and de-

pends on the state in which the node n

is, Voltage(n

)

voltage is the supply voltage in Volts and Duration is

the interval that passed since the last energy update.

For each node n

an energy cost function C(n

)is

assigned that is given by:

C(n

) =

init

)

R(n

)

(5)

Thus, the total energy cost for a route p from

source node n

to destination node n

, is given by:

∑

∈ p, n

6= n

C(n

) (6)

The selected route l will be the one that satisﬁes

the following property:

= min{E

: p ∈ V} (7)

where V is the set of all the possible routes.

6 SIMULATION RESULTS

The proposed metric (denoted as energy in the fol-

lowing ﬁgures) was evaluated using the simulation

software ns-3 (http://www.nsnam.org) and its perfor-

mance was compared only against the ALM (denoted

as airtime in the following ﬁgures) since important

implementation details, as well as, simulation param-

eters for the ETE metric are not given. Also the

authors consider that a node consumes energy only

when it transmits or receives data .

We set up our simulation by constructing a grid

topology with N x N mesh STAs. The distance be-

tween two neighboring mesh STAs in the grid was

set to 120 m. In our simulations we generated four

UDP trafﬁc ﬂows among 4 nodes in the network. Fig-

ure 2 shows an illustrative example for the case of a

3 x 3 grid topology. Each simulation run consisted

of 10 iterations, each having the same pair of sender-

destination. Details concerning the simulation param-

eters are given in Table 3.

Figure 2: A 3 x 3 grid topology.

Table 3: Simulation Parameters.

Parameter Value

N 2, 3, 4, 5, 6 and 8

Initial Energy (E

init

) 33 Joules

Interval 0,1 sec

Data rate 150 kbps

Packet size 1024 bytes

Propagation loss model log-distance

Transmission power 18 dbm

Simulation time 25000 sec

WINSYS2013-InternationalConferenceonWirelessInformationNetworksandSystems

218

The residual energy of a node is calculated by the

energy that consumes whenever it is in one of the

states described above. The supply voltage of each

node is set to 3 Volts, while the Duration is set to 1sec.

For each state different current has been set. Table 4

summarizes the current of a node for each state.

Table 4: Energy consumption for each state.

State Current (Ampere)

IDLE 0.00426

CCA BUSY 0.00426

TX 0.0174

RX 0.0194

SWITCHING 0.00426

We used three different quantitative measures

(network lifetime, delay and the Packet Delivery

Fraction (PDF)) to compare the performance of the

three routing metrics. It should be noted that the re-

active mode of the HWMP protocol was used for path

selection, which implies the absence of a root node

in the mesh network used in our simulation scenarios.

Also, as the network lifetime we consider the time

that the ﬁrst node of the network ran out of energy.

100

110

120

130

140

150

160

170

0 10 20 30 40 50 60 70

Lifetime (min)

# Nodes

Network Lifetime(min)

Energy

Airtime

Figure 3: Lifetime vs # Nodes.

Figure 3 illustrates the network lifetime for each

metric that was applied in the HWMP. As Figure 3

depicts, the proposed metric outperformsthe ALM re-

garding the network lifetime. When the ALM metric

is applied in the HWMP, energy constraints are not

taken into account for the forwarding of data. Thus,

the energy of the nodes along that path is depleted at

shorter time.

In terms of the total delay that the network experi-

ences, as shown in ﬁgure 4, the ALM metric achieves

lower delay in comparison with the proposed one.

Since the proposed metric considers the energy con-

straints it should select longer routes than ALM, lead-

ing to a slightly increase in terms of delay.

The last measure that was used for the compari-

0.005

0.01

0.015

0.02

0.025

0.03

0.035

0 10 20 30 40 50 60 70

Delay (sec)

# Nodes

Total Delay(sec)

Energy

Airtime

Figure 4: Total Delay vs # Nodes.

100

0 10 20 30 40 50 60 70

pdf (%)

# Nodes

Packet Delivery Fraction(%)

Energy

Airtime

Figure 5: Packet Delivery Fraction vs # Nodes.

son of the three metrics is the Packet Delivery Frac-

tion (PDF), which denotes the percentage of the suc-

cessfully delivered packets. As shown in ﬁgure 5, the

PDFs for both merics are quite identical.

7 CONCLUSIONS

In this paper, we have proposed a new routing met-

ric that was incorporated with the HWMP, in place

of the ALM, and applied in a WMN. The proposed

metric calculates an energy cost function by taking

into account the residual energy of nodes in the net-

work and selects the path that minimizes the afore-

mentioned cost function. Simulation results showed

that the proposed metric prolongs the lifetime of a

WMN in comparison with the ALM, while causing

a little higher total delay in the network.

ACKNOWLEDGEMENTS

This work was carried out during the tenure of an

ERCIM ”Alain Bensoussan” Fellowship Programme.

EnergyAwareRoutinginIEEE802.11sWirelessMeshNetworks

219

The research leading to these results has received

funding from the European Union Seventh Frame-

work Programme (FP7 2007-2013) under grant agree-

ment no 246016.

This work is also partly supported by UPRC (Uni-

versity of Piraeus Research Center).

REFERENCES

802.11s 2011/D12.0 (2011). Ieee standard for information

technology – telecommunications and information ex-

change between systems – local and metropolitan

area networks–speciﬁc requirements part 11: Wire-

less lan medium access control (mac) and physical

layer (phy) speciﬁcations amendment 10: Mesh net-

working. IEEE Std 802.11s-2011 (Amendment to

IEEE Std 802.11-2007 as amended by IEEE 802.11k-

2008, IEEE 802.11r-2008, IEEE 802.11y-2008, IEEE

802.11w-2009, IEEE 802.11n-2009, IEEE 802.11p-

2010, IEEE 802.11z-2010, IEEE 802.11v-2011, and

IEEE 802.11u-2011), pages 1–372.

Bin Ngadi, M., Ali, S., Abdullah, A., and Khokhar, R.

(2012). A taxonomy of cross layer routing metrics for

wireless mesh networks. EURASIP Journal on Wire-

less Communications and Networking, 2012(1):1–16.

Campista, M., Esposito, P., Moraes, I., Costa, L., Duarte,

O., Passos, D., de Albuquerque, C., Saade, D., and

Rubinstein, M. (2008). Routing metrics and protocols

for wireless mesh networks. Network, IEEE, 22(1):6–

12.

Draves, R., Padhye, J., and Zill, B. (2004). Routing in

multi-radio, multi-hop wireless mesh networks. In

MobiCom ’04, 10th annual international conference

on Mobile computing and networking, pages 114–128.

ACM.

Islam, M., Alam, M., Hamid, M., Hong, C., and Lee, S.

(2010). EFT: a High Throughput Routing Metric for

IEEE 802.11s Wireless Mesh Networks. Annals of

Telecommunications, 65(5-6):247–262.

Jin, Y., Miao, H., Ge, Q., and Zhou, C. (2011). Ex-

pected transmission energy route metric for wireless

mesh sensor networks. International Journal of Digi-

tal Multimedia Broadcasting, 2011.

Koksal, C. E. and Balakrishnan, H. (2006). Quality-aware

routing metrics for time-varying wireless mesh net-

works. Selected Areas in Communications, IEEE

Journal on, 24(11):1984–1994.

Mhatre, V., Lundgren, H., and Diot, C. (2007). Mac-aware

routing in wireless mesh networks. In WONS ’07,

Fourth Annual Conference on Wireless on Demand

Network Systems and Services, pages 46–49. IEEE.

Nandiraju, N., Nandiraju, D., Santhanam, L., He, B., Wang,

J., and Agrawal, D. (2007). Wireless Mesh Networks:

Current Challenges and Future Directions of Web-In-

The-Sky. Wireless Communications, IEEE, 14(4):79–

89.

Passos, D., Teixeira, D. V., Muchaluat-Saade, D. C., Mag-

alh˜aes, L. C. S., and Albuquerque, C. V. (2006). Mesh

network performance measurements. In ISMICT ’06,

5th International Information and Telecommunication

Technologies Symposium.

Shyy, D. (2006). Military usage scenario and ieee 802.11s

mesh networking standard. In Military Communica-

tions Conference, 2006. MILCOM 2006., pages 1–7.

IEEE.

Subramanian, A., Buddhikot, M., and Miller, S. (2006). In-

terference aware routing in multi-radio wireless mesh

networks. In WiMesh ’06, 2nd IEEE Workshop on

Wireless Mesh Networks, pages 55–63. IEEE.

Yang, Y., Wang, J., and Kravets, R. (2005). Designing rout-

ing metrics for mesh networks. In WiMesh’05, 1st

IEEE Workshop on Wireless Mesh Networks. IEEE.

WINSYS2013-InternationalConferenceonWirelessInformationNetworksandSystems

220