Simulation and Testbed Evaluation for Optimizing Energy

Consumption in Ad Hoc Networks based on OLSR Protocol

Santiago González, Pau Arce and Juan Carlos Guerri

Institute of Telecommunications and Multimedia Applications (iTEAM), Universitat Politècnica de València,

Camino de Vera s/n CP.46022, Valencia, Spain

Keywords: Ad Hoc Networks, OLSR, Energy Consumption, Overhearing Effect, Testbed Evaluation.

Abstract: This paper presents a proposal to optimize energy consumption in ad hoc networks based on the OLSR

protocol. This approach focuses on the set up of routes with less congestion level and higher energy

capacity. Therefore, in addition to the remaining energy of nodes, a new metric is introduced, the strategic

value, which reports the importance of a specific node in the network based on the numbers of neighbors it

has. In order to obtain valuable results, the evaluation was performed in a simulation environment (NS3)

and on a real testbed. In that sense, an actual ad hoc network was implemented using embedded devices

(Raspberry Pi). Results show a decrease in energy consumption, especially in zones with the highest device

density, as well as an increase of the time of operation for nodes with higher amount of neighbors.

Additionally, the performed evaluation shows a positive effect in the quality of traffic flows, avoiding route

breakages and packet losses.

1 INTRODUCTION

Ad hoc networks represent an alternative in order to

implement new schemes of communication, for

example the concept of opportunistic

communication (Giordano 2014). In particular, such

approach stands out the possibility to take advantage

of the high density of mobile devices in order to set

up wireless links through a collaborative

mechanism, as described in (Tehrani et al. 2014) and

(Do et al. 2012).

Also, ad hoc networks present a real potential for

the implementation of services focused on smart

cityes, especially for the capture and fast

dissemination of data in urban zones (Bellavista et

al. 2013).

In spite of the advantages described, there are a

set of challenges attached to an ad hoc scenario.

Mainly, energy limitation in mobile devices due

to the use of batteries is a significant factor for the

design and implementation in real enviroments.

In that sense, the transmission of multimedia

traffic as well as the increase in the data rate

achieved using recent standards, such as IEEE

802.11n/ac, result in a higher traffic load and,

therefore, higher demand of energy consumption on

devices.

Moreover, wireless medium is another factor that

deserves to be analized in regard to energy

consumption, due to the operation of a radio

interface. A wireless card analyzes the power level

of detected signals in order to change to receiving

mode, or starts a transmission process if the medium

is available and there are packets to transmit. Such

mode of operation causes that a node located in an

interference zone changes to a receiving state due to

the detection of signals with a power level higher

than the threshold, even if it is not the target of data.

This effect is named overhearing and contributes

to the increase of energy consumption.

In this paper, we have focused on mechanisms at

routing level in order to optimize the energy

expenditure, specifically in Ad hoc networks based

on the OLSR protocol. In particular, OLSR

implements a mechanism to disseminate the routing

information over special nodes named MPR (Multi

Point Relay). These nodes are selected in a process

which consitently analyzes the availability to carry

out the MPR function (willingness), the number of

nodes within its connectivity area (reachability), and

the simmetry of the links with neighbour nodes

(Clausen and P. Jacquet 2003). However, the

original standard does not consider energy

limitations.

González, S., Arce, P. and Guerri, J.

Simulation and Testbed Evaluation for Optimizing Energy Consumption in Ad Hoc Networks based on OLSR Protocol.

DOI: 10.5220/0005955301290136

In Proceedings of the 13th International Joint Conference on e-Business and Telecommunications (ICETE 2016) - Volume 6: WINSYS, pages 129-136

ISBN: 978-989-758-196-0

129

In this sense, this paper presents a new approach

in order to optimize the enegy consumption by

means of analyzing energy capacity as well as the

distribution of the nodes in the network. Thus,

energy expenditure in zones with higher device

density is reduced while the lifetime of strategic

nodes (i.e. nodes with higher connectivity, with

higher number of neighbours) is increased. The

presented proposal is called, therefore, Strategic

OLSR (S-OLSR).

The evaluation was performed, on the one hand,

in a simulation environment (NS3) (Ns3-project

2014) and, on the other hand, over a real testbed

formed by a set of ten ad hoc nodes. The testbed

was implemented using Raspberry Pi. The results

prove that the proposed scheme contributes to

reduce the energy expenditure and, additionally,

causes a positive impact on the quality of

transmitted traffic.

The paper is organized as follows. Section 2

presents related work. Section 3 describes the

mechanism of optimization proposed. The

evaluation and results in the simulation environment

is presented in Section 4. Section 5 describes the

testbed implementation as well as the experiments

performed and results. Finally, Section 6 presents

conclusions and future lines of work.

2 RELATED WORK

This section presents a classification of related work

regarding energy optimizations implemented over

the OLSR protocol.

2.1 EMPR-OLSR

A first set of solutions propose to include the

residual energy level as metric in order to select

MPR nodes. In this paper, this methodology is

referred as Energy MPR - OLSR (EMPR-OLSR).

As an example, (Wardi et al. 2011) presents a

proposal that uses this approach based on a minimun

energy threshold as a condition to consider a node

for being part of the set of MPR candidates.

A modification is proposed in (Fatima Lakrami

and Najib Elkamoun 2012), in which the willingness

parameter is defined as a function of the residual

energy and, consequently, the normal process of

selection described in the standard is used.

However, these mechanisms cause an increment

in the number of nodes selected as MPR to reach the

nodes at more than one hop. Therefore, it generates a

flooding increase of Topology Control (TC)

messages. An analysis of this effect is described in

(Mahfoudh and Minet 2008).

2.2 ER-OLSR

The second related methodology takes into account

the energy capacity of nodes for routing

computation. Consequently, these mechanisms have

been called Energy Routing – OLSR (ER-OLSR). In

this sense, (Toh 2001) proposes a cost function in

order to evaluate the energy expenditure related to a

route between transmitter and receiver nodes, and

then select the path with less cost. The main

disadvantage of such strategy is that the evaluation

is performed globally on a route and, therefore,

intermediate nodes with critical energy levels may

not be taken into account.

In order to avoid such effect, (Adoni and Joshi

2012) proposes to set the value corresponding to the

node with less residual energy in a path as the route

cost. Additionally, (Rango et al. 2008) and (De

Rango and Fotino 2009) describe schemes

conceptually similar to (Adoni and Joshi 2012), in

this case introducing a metric called MDR

(Minimum Drain Rate) that provides an estimation

about the lifetime of a node.

Nevertheless, the main drawback of these

proposals is the liability to increase the number of

hops due to an evaluation focused on a single node.

In that sense, an analysis in each node along the

route represents a better indicator about the real

energy conditions in the network.

2.3 EA-OLSR

Finally, the third mechanism described in previous

works proposes the application of the metric of

residual energy for both MPR nodes selection and

route computation, simultaneously. These proposals

have been called Energy Aware – OLSR (EA-

OLSR) in this paper.

Regarding this mechanism, (Kunz 2008) uses the

residual energy in order to select MPR nodes and

additionally, a weigth is assigned in each link

accordign to the energy capacity in the node that

forwards the traffic.

A variation of this approach is presented in

(Machado et al. 2013), which introduces the usage

of the ETX (Expected Transmission Count) metric

for route computation. Such a metric allows to

evaluate the reliability of a link and, consequently,

an indirect management of the energy capacity can

be performed.

WINSYS 2016 - International Conference on Wireless Networks and Mobile Systems

130

3 STRATEGIC OLSR (S-OLSR)

In this paper, we introduce a new metric called

Strategic Value (SV). This metric consists in an

indicator about the number of neighbor nodes within

an interference zone. Consequently, the strategic

value is related to the relative location of a node and,

therefore, provides information about the

distribution of nodes in the network.

The SV is obtained from the information

collected during the exchange of hello messages

(OLSR protocol) and, therefore, it is updated

according to the configured hello interval. Then, the

SV is used in routing computation using information

from every hop, in order to find intermediate nodes

with less strategic value. Additionally, this process is

complemented with an energy analysis. The

proposed scheme selects the nodes with an energy

capacity equal or greater than 90% compared with

the residual energy available in a competitor node.

This is performed in order to set up routes with

less overhearing effect and suitable energy capacity.

Moreover, the tolerance configured guarantees

the priority for the routes selected. The presented

approach based on OLSR protocol, that takes into

account both energy capacity and strategic value of

nodes, is called Strategic OLSR (S-OLSR).

Figure 1: Descriptive diagram of S-OLSR operation and

interference zones for nodes: (a) Node A, (b) Node B, (c)

Node C, (d) Route analysis.

Figure 1 describes a diagram showing how the

proposal works. The routing protocol is going to set

up a route between nodes src and dst. Thus, the

coverage areas depicted in Figures 1(a), (b) and (c)

indicate that nodes A, B and C, respectively, are

candidates to be forwarder nodes in order to find the

shorter route to the target. The residual energy, as

well as the SV metric for these candidate nodes are

also depicted in the diagram. Additionally, there are

a set of neighbor nodes inside the interference zone

of each candidate node. Examples of SV and

remaining energy values were defined in order to

describe the operation of the proposal.

Then, according to the scenario, node B, which

has the lower strategic value and more residual

energy, becomes the best option to forward packets

towards the destination node (dst), as can be seen in

Figure 1(d). Especially, it can be observed that the

overhearing effect is lower (7 nodes within the

interference area) when using node B as forwarder

node than when using node A or C (10 nodes and 11

nodes within the interference area, respectively).

In order to carry out the assessment, the proposal

was implemented using the NS3 simulator,

performing the required modifications to the

standard protocol (OLSR RFC3626). Mainly, the

strategic value and the residual energy information

were included inside the header of hello messages.

Regarding the residual energy, this information

is provided by the physical layer in a cross-layer

operation. Finally, the metrics are evaluated for

routing computation as described above. The

pseudo-code for the routing modification is

presented in Table I. Moreover, next section

describes the methodology used for the evaluation of

the proposal in the simulation environment.

Table 1: Algorithm S-OLSR: Operation on each node n

for routing computation.

Required: TargetNode (tn), Neighbor-Set Nodes (N),

StrategicValue (SV), Residual Energy (Er), HopNumber

(h), Address (add), EnergyTolerance (

Initialize:(add

table

, SV

table,

table

= null; h = 2;

=0.9)

1:Begin RoutingTableComputation

while (RoutingFinished = false )

for (n = 0 ; n < N ; n++)

if (add

=NextHopToadd

&(add

table

=null||(SV

< . SV

table

& Er

≥

α*(Er

table

)))) then

5: add

table

= add

;

6: SV

table

= SV

;

7: Er

table

= Er

;

endif

endfor

10:

if add

table

= add

then

11: RoutingFinished = true;

12:

else

13: h++;

14:

endif

15:

endwhile

16:

end

Simulation and Testbed Evaluation for Optimizing Energy Consumption in Ad Hoc Networks based on OLSR Protocol

131

4 SIMULATION EVALUATION

4.1 Methodology

Figure 2 presents a diagram of the process for the

evaluation in the simulation environment.

Figure 2: Evaluation methodology for the proposal in the

simulation environment.

First, it is worth mentioning that a video

sequence has been used as traffic flow, mainly due

to the significant load of traffic that this kind of data

involves but also because of the increasing demand

of multimedia contents among mobile users

nowadays. As can be seen in Figure 2, the process

starts from a raw file (.yuv), and it is encoded

(H264) and packaged afterwards (MP4). Finally, the

trace file containing information about size and

timestamp of video packets is obtained. This process

is performed by means of FFmpeg (FFmpeg.org

2016) and Evalvid (Klaue et al. 2003). This

descriptor file, as well as the parameters specified in

the Table 2, are used to set up the simulation in the

NS3 environment. Furthermore, Figure 3 presents

the scenario of evaluation. As can be seen, the node

0 is the source of the traffic and node 1 is the

receiver. This scenario was designed in order to

compare the pattern of energy consumption for the

mechanism ER-OLSR and the standard (RFC3626)

versus the proposed scheme (S-OLSR).

Table 2: Simulation Parameters.

Parameter Value

Mac Protocol 802.11g

Rate 54 Mbps

Rx Sensitivity -76 dBm

Tx Power Level 0 dBm

Interference Range 30m

Intensity Consumption (mA)

(Intel7260 802.11a/b/g/n)

Tx:606; Rx:485;

Idle:75

Traffic Video 300s; 100 repetitions

Video Bitrate (Average) 300 kbps

Initial Node Energy 10000 J

Figure 3: Scenario designed for the evaluation of S-OLSR.

4.2 Results

Results are shown in Figure 4. Energy expenditure

works as a clear indicator of which nodes belong to

the routes used for data transmission. As can be

seen, the standard OLSR protocol causes higher

energy consumption on node 5, which is the node

with the highest strategic value in the scenario.

0123456789

Node

Energy Consumption (%)

Strategic Value

SV OLSR-RFC 3626 ER-OLSR S-OLSR

Figure 4: Comparison of energy consumption pattern and

the strategic value of the nodes.

This high strategic value entails that a greater

number of nodes are affected by the overhearing

effect, specifically, nodes 2, 3, 6, 8 and 9, which are

located inside the interference area of node 5. This

behavior is due to the use of the number of hops as

the single metric for the routing computation.

Consequently, the probability of selecting nodes

with a greater number of links is higher. Regarding

ER-OLSR, results show a reduction of the energy

expenditure on each node. However, the pattern of

consumption is similar to the original protocol. Such

behavior is mainly due to the lack of analysis about

the distribution of the nodes in the scenario. In

WINSYS 2016 - International Conference on Wireless Networks and Mobile Systems

132

regard to S-OLSR, the routing computation analyzes

the residual energy and, additionally, the number of

nodes inside the interference area by means of the

SV metric. Consequently, this mechanism assigns

priority to routes with less overhearing effect.

Therefore, results describe a significant

modification in the pattern of energy consumption.

Specifically, Figure 4 shows an increase in the

energy expenditure on the nodes 2 and 4 (with less

strategic value), compared with the competitors,

nodes 3 and 5. Regarding node 5, the difference

achieved is not remarkable due to the critical

location of this node in the scenario, i.e. it is inside

the interference area of all potential routes toward

the destination node. Nevertheless, the priority in the

selection of nodes with less strategic value has lead

to a significant decrease of the energy consumption

in nodes 3, 6, 7 and 9. Furthermore, Figure 5 shows

an analysis of the distribution of the energy

consumption on the scenario. As can be seen, S-

OLSR reduces the energy expenditure for all zones

defined on the scenario (Figure 3). In particular,

zone 2, which is characterized by the highest density

of nodes, presents the most significant difference,

53% (S-OLSR) compared with 59% (ER-OLSR)

and 64% (standard OLSR). The advantage achieved

is observed even globally, (including src and dst

nodes). In this case, results are 54% (S-OLSR), 58%

(ER-OLSR) y 62% (standard OLSR).

Zone1 Zone2 Zone3 Network

Energy Consumption (%)

OLSR-RFC 3626 ER-OLSR S-OLSR

Figure 5: Comparison of average energy consumption for

defined zones and for the whole network.

5 TESTBED EVALUATION

5.1 Ad Hoc Node Implementation

For the testbed evaluation, we implemented a set of

10 ad hoc nodes over embedded devices with Linux

(Raspberry Pi B+) (Ada 2015). The functional

diagram of a node is presented in Figure 6.

Figure 6: Functional diagram of an ad hoc node

implemented over a Raspberry Pi B+.

Ad hoc network configuration is performed in

the block Network/Synchronization. Additionally,

we installed the OLSR protocol (olsrd daemon) as

well as a NTP client to synchronize the nodes during

startup. The selection of the wireless card was

carried out taking into account the prior

experimentation with several models. The main

constraints were the compatibility of drivers with the

development platform and the operation of the card

in a real ad hoc communication mode. Taking into

account the results of the tests, Awuso36nh card

(linux driver rt2800 /chipset RT3070) (AlfaNetwork

2015) was selected. Finally, the nodes are powered

using a power bank of 10000mAh. At user level

several free distribution tools have been installed,

such as the mp4trace tool for video transmission,

available in the Evalvid package, tcpdump

(Tcpdum.org 2016) and tcpstat for capture and

analysis of traffic. The transmitter node stores the set

of videos which will be used in the test.

Additionally, the current sensor INA219

(Adafruit 2015) has been incorporated in each node,

in order to assess the level of average power

consumption demanded by the wireless card. The

sensor is handled by a set of Python libraries

developed by Adafruit. The communication is

carried out is via the I2C bus (Inter - Integrated

Circuit) in the GPIO pins (General Purpose Input /

Output). Figure 7 shows the physical ad hoc node

implemented and the components. The main goal of

the testbed is to help configure the scenario from the

Figure 3 for the evaluation of S-OLSR.

Simulation and Testbed Evaluation for Optimizing Energy Consumption in Ad Hoc Networks based on OLSR Protocol

133

Figure 7: Description of Ad hoc node components.

5.2 S-OLSR Implementation

Beyond the simulations, S-OLSR was implemented

on real ad hoc nodes. Figure 8 shows the functional

diagram for the performed development.

Figure 8: Functional diagram for S-OLSR implementation.

As can be seen, the first step is to measure the

energy consumption (EC) demanded by the wireless

card. For this purpose, the current sensor is

controlled by a Python script, which captures current

samples from the wireless card during a time

interval and later processes the samples in order to

compute the percentage of energy consumption.

Finally, this energy value is stored and updated

at the same rate as the hello interval does (2s) within

the OLSR protocol. In regard to modifications on the

routing protocol, we used the routing daemon olsrd-

0.9.0.2 (Olsr.org 2016) as starting point. In order to

include new metrics (SV and EC), the reserved

fields in the header of the hello message was used so

that modifications to the original protocol have kept

to a minimum. As aforementioned, the value of

energy consumption is introduced from the

information provided by the current sensor. The SV

metric is deducted from the number of nodes at one

hop observed in the neighbor table and also included

inside the hello message. Therefore, this slight

modification of the hello message allows the

exchange of the new metrics without altering the

original fields and thus, maintaining backward

compatibility. Moreover, routing computation has

been modified in order to take into account these

new metrics. The routes, previously determined by

the SPF process (Short Path First), give now higher

priority to the nodes with less strategic value and

less energy consumption to be selected as next hop

and included in the routing tables. Also, it is worth

clarifying that the energy metric used was the energy

consumption, instead of the residual energy, due to

the fact that obtaining the samples of current from

the wireless card using the sensor was simpler than

inferring the remaining battery.

5.3 Results

Figure 9 presents the set up for the experiment in the

laboratory environment. The scenario depicted in

Figure 3 was implemented. In order to replicate node

connectivity, layer-2 filters in each node have been

configured, providing connectivity only among

nodes according the scheme in Figure 3. The main

parameters used for the testbed are described in

Table 3.

Figure 9: Testbed evaluation in the laboratory

environment.

Table 3: Testbed parameters.

Parameter Value

Mac Protocol 802.11g

Rate 54 Mbps

Rx Sensitivity -76 dBm

Anntena Gain 5dBi

Tx Power Level 0 dBm

Hello and TC Interval 2s ; 5s

Intensity Consumption

(mA)

Tx:409;

Rx:204;offline:100mA

Video Traffic 60s -10 repetitions

Video Bitrate (Average) 300kbps

Node Energy (mA) 4.17mAh

An initial energy capacity has been defined

(4.17mAh) for every node in the network, excluding

N5, which is the node with the highest strategic

value. Intentionally, the initial energy capacity of

WINSYS 2016 - International Conference on Wireless Networks and Mobile Systems

134

node 5 was set to 50% (2.08mAh) in order to

evaluate the case when node 5 consumes all the

remaining energy before the experiment ends.

Additionally, the Python script will disable the

wireless card (switch to offline state) when the

energy consumption increases to 90% of the

capacity, which equates to 10% of residual energy,

and emulates the power-saving state. The traffic

used corresponds to 1min of the “Big Buck Bunny”

video sequence. The evaluation consists in the

comparison of S-OLSR with the standard protocol.

First, an analysis about the routes selected by

each mechanism was performed. Figure 10 shows

results about the throughput in each node. As can be

seen, the OLSR protocol (standard) leads the traffic

through routes defined by node 3, either nodes 5 or

6, and finally node 9. On the other hand, S-OLSR

estimates the better route through nodes 2, 4 and 8.

This is the operation expected for the proposal,

due to the less strategic value of such nodes.

Therefore, in this case the traffic flow avoids

node 5 (N5) because it is the node with the highest

energy restriction in the configured scenario.

120

150

180

210

240

270

300

330

23456789

Intermediate Node

Average Throughput (Kbps)

OLSR S-OLSR

Figure 10: Throughput at each intermediate node.

Additionally, Figure 11 shows the throughput

measured on the receiver node (N1).

0,2

0,4

0,6

0,8

0 5 10 15 20 25 30 35 40 45 50 55 60

Time (s)

roug

put

(Mb

)

OLSR S-OLSR

Figure 11: Throughput at the receiver node (N1).

As can be observed, the behavior of the standard

protocol causes the interruption of the traffic (from

28s to 35s), due to the full depletion of energy in

node 5, while the new route is recovered through

node 6, as can be inferred from Figure 10. In regard

to S-OLSR, the traffic flow is continuous during the

whole experiment. As can be seen in Figure 12,

S-OLSR presents higher reception rate (97%)

compared with the standard OLSR (82%). In this

sense, Figure 13 shows higher average PSNR

(38dB) versus the original OLSR (34dB).

100

Average Received Packets (%)

OLSR S-OLSR

Figure 12: Comparison of average packet reception rate:

Standard OLSR and S-OLSR.

PSNR (dB)

OLSR S-OLSR

Figure 13: Comparison of average PSNR: Standard OLSR

and S-OLSR.

Finally, another experiment was carried out in

order to characterize the energy depletion profile on

node 5. In this case, in addition to the video traffic

from N0 toward N1, a background flow (400Kbps)

from N3 to N9 was configured. This background

traffic tries to emulate ongoing connections from

other nodes in the network that, while not being

routed through node 5, interfere severely on energy

consumption due to the overhearing effect. The

initial energy capacity for node 5 was set up to 80%.

Again, the critical threshold of residual energy is

set up to 10% in order to disable the wireless card

and emulate a power-saving state. Results are shown

in Figure 14. As can be observed, the energy

depletion is most remarkable with the standard

protocol. In particular, the critical energy level takes

place approximately at 43s. The change in the slope

next to the critical value is due to lower consumption

demanded by the wireless card when is disabled

(100mA, Table 3). On the other hand, S-OLSR

presents a slower decrease of the residual energy on

node 5. Thus, the time of operation is extended to 55

seconds, corresponding to an increase of 20% in the

interval defined for evaluation (60 seconds).

Simulation and Testbed Evaluation for Optimizing Energy Consumption in Ad Hoc Networks based on OLSR Protocol

135

Figure 14: Comparison of energy depletion on the node

with highest strategic value in the scenario (N5).

6 CONCLUSIONS

In this paper we performed a thorough analysis

regarding energy optimization in ad hoc networks

and propose a new routing approach based on the

OLSR protocol. Specifically, our proposal (S-

OLSR), in addition to the energy metric, includes

the analysis of node distribution in order to set up

routes characterized by a less level of congestion.

Also, this approach aims at decreasing the power

consumption on nodes with higher amount of

neighbors, since they are likely to be the most

strategic nodes to maintain the whole network

connectivity. The evaluation performed on the

simulation environment shows clear changes in the

pattern of energy expenditure using S-OLSR. The

most significant difference is achieved on the zone

with higher node density. Specifically, results show

a reduction in the energy consumption of 6% and

11% in comparison with the ER-OLSR mechanism

and the standard OLSR protocol, respectively.

Moreover, results from the real testbed show the

expected behavior of the proposal. The routes are set

up through nodes with less strategic value, which

contribute to extend the lifetime of the node with the

highest number of links (N5), even when the traffic

load is increased in the network.

ACKNOWLEDGEMENTS

This paper was performed with the support of the

National Secretary of Higher Education, Science,

Technology and Innovation–Ecuador Government

(scholarship 195-2012) and the Thematic Network:

RiegoNets (514RT0486)-CYTED.

REFERENCES

Ada, Lady, 2015. Introducing the Raspberry Pi2-Model B.

Adafruit, 2015. INA219CurrentSensor. Datasheet.

Adoni, K.A. and Joshi, R.D., 2012. Optimization of

Energy Consumption for OLSR Routing Protocol in

MANET. IJWMN, 4(1), pp.251–262.

AlfaNetwork, 2015. AWUS036NH 802.11n/b/g USB

Adapter Technical Specifications.

Bellavista, P. et al., 2013. Convergence of MANET and

WSN in IoT urban scenarios. IEEE Sensors Journal,

13(10), pp.3558–3567.

Clausen, T. and P. Jacquet, 2003. Optimized Link State

Routing Protocol (OLSR). RFC:3626.

Do, N.M., Hsu, C.H. and Venkatasubramanian, N., 2012.

HybCAST: Rich content dissemination in hybrid

cellular and 802.11 ad hoc networks. In SRDS,

pp.352–361.

Fatima Lakrami and Najib Elkamoun, 2012. Mobility and

QOS Management in OLSR Routing Protocol.

IJCNWMC, 2(4), pp.17–26.

FFmpeg.org, 2016. FFmpeg. Documentation. Available at:

https://www.ffmpeg.org/ffmpeg.html.

Giordano, M.C. and S., 2014. Mobile ad hoc networking:

milestones, challenges, and new research directions.

IEEE Communications Magazine, 52(1), pp.85–96.

Intel7260 802.11a/b/g/n, 2013. Technical Specifications.

Klaue, J., Rathke, B. and Wolisz, A., 2003. EvalVid - A

Framework for Video Transmission and Quality

Evaluation. In TOOLS, pp.255–272.

Kunz, T., 2008. Energy-Efficient Variations of OLSR. In

IWCMC.

Machado, D.L.P., Carrano, R.C. and Saade, D.C.M., 2013.

Analysis of Energy Efficient OLSR Extensions and

OLSR- ETX Energetic Optimization Proposal. In PE-

WASUN.

Mahfoudh, S. and Minet, P., 2008. An Energy Efficient

Routing Based on OLSR in Wireless Ad Hoc and

Sensor Networks. In AINAW. pp.1253–1259.

Ns3-project, 2014. ns-3 Model Library. Available at:

https://www.nsnam.org/documentation/.

Olsr.org, 2016. Olsr project. Available at:

http://www.olsr.org/mediawiki/index.php/Releases.

De Rango, F. and Fotino, M., 2009. Energy efficient

OLSR performance evaluation under energy aware

metrics.In SPECTS, pp.193–198.

Rango, F. De, Fotino, M. and Marano, S., 2008. EE-

OLSR: Energy Efficient OLSR routing protocol for

Mobile ad-hoc Networks. MILCOM IEEE.

Tcpdum.org, 2016. TCPDUMP and LIBCAP. Available

at: http://www.tcpdump.org/.

Tehrani, M.N., Uysal, M. and Yanikomeroglu, H., 2014.

Device-to-Device Communication in 5G Cellular

Networks Challenges, Solutions, and Future

Directions . IEEE Communications Magazine, 52(5).

Toh, C.-K., 2001. Maximum battery life routing to support

ubiquitous mobile computing in wireless ad hoc

networks. IEEE Communications Magazine, 39(6).

Wardi et al., 2011. Residual energy-based OLSR in mobile

ad hoc networks. 2011, ICMT , pp.3214–3217.

WINSYS 2016 - International Conference on Wireless Networks and Mobile Systems

136