Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

Power Consumption and Energy Efficiency of 802.11 Network Devices P.M. Asuquo, E. Edet & I.K. Akpabio Department of Computer Engineering University of Uyo Uyo, Akwa Ibom State, Nigeria. [email protected], [email protected], [email protected]

ABSTRACT This paper examines the power consumption of IEEE 802.11standards with distance variations. The effectiveness of each standard is analysed over a small wireless network. Experiments carried out to analyse power consumed by 802.11 standards show that energy consumed during transmission of data is higher than that consumed during reception. Results from experiments carried out show that the power consumption of 802.11n is more than that of 802.11b/g in both transmitting and receiving modes. The results also show that 802.11n standards are more effective especially for data transfer where it doubles the transfer rate of b/g in most cases. To achieve minimum power consumption, the appropriate data rate based on the available standard should be chosen for network implementation. However it is also observed that more power is consumed by the laptop by transmitting and receiving of packets. Keywords – Energy Consumption of ICT Devices, Access Points, Wireless Networks, 802.11 Networks African Journal of Computing & ICT Reference Format: P.M. Asuquo, E. Edet & I.K. Akpabio (2013). EWiSECS: Power Consumption And Energy Efficiency of 802.11 Network Devices. Afr J. of Comp & ICTs. Vol 6, No. 5. Pp 35-.42.

1. INTRODUCTION It is noted that Information and Communication Technology (ICT) devices and wireless networks also contribute greatly to the power wastage [12]. Recently, it has been observed that ICT devices typically consume more than 20% of energy in some organisations and in other cases up to 70% [6]. Furthermore, [8]points out that between 2000 and 2009, the amount of IT devices in British households increased from 30 million to 65 million, bringing about an increase in energy consumption by these devices. The ICT devices consume 25% of electricity in commercial offices and this tremendous power consumption by these devices contributes about 2 to 2.5% of global carbon emissions [6]. As a result, the carbon emission by the ICT industry is on the verge of exceeding that of the aviation industry [10]. Globally, about 6 to 10% of the world’s energy is consumed by the ICT industry and that the energy consumption of these devices is doubling every five years at an alarming rate. On the causes of these incessant wastages of energy, different Researchers such as [5] and [9] agree that in a wireless network that apart from the transmission, the energy consumed during the idle mode and receiving mode are roughly the same as shown in figure 1 below. They also suggest that repeated transmission of a packets and receiving of unwanted packets by a node consume energy unnecessarily.

Also, [2] indicate that desktop computers and cathode ray tube monitors consume the largest amount of energy while greater portion of this energy is also consumed by the radio interface of sensors [11]. Furthermore, idle listening; overhearing, control packet overhead and collision are the major reason of energy wastage [1].In order to reduce these energy wastages, several steps applied by different researchers in the likely fields include the one proposed by [12] which is adjusting transmission range at each node and power aware routing which utilises the closest route for sending information to the base station. [2] points out that desktop computers and monitors consume the largest amount of energy and consider the replacement of the old cathode ray tube monitors with the liquid crystal display screen monitor as a helpful step. [10] Further highlights the use of centralized power management software which can automatically switch off equipment when it is not in use, migrating to high speed optical networks, hardware consolidation and hardware design optimization can also solve the problem [4]. MAC technology can help to save energy by achieving a good topology and adjusting sleep periods of sensors and the route technology which identifies the best path of transmission that will consume less energy, emphasizing on the route technologies (Flat and Hierarchical) [5] & [9].

35

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

2. METHODOLOGY

Figure 1: Energy consumed by different parts of a node [5] The Kruskal’s Algorithm shows that the total power consumed in a wireless network can be minimized by carefully choosing specific nodes to route information to neighboring nodes based on how close they are to each other and [3] work is similar to quorum-based loadsharing control protocol (QLSCP) proposed by [7] which selects suitable communication nodes, adjusts the service load of critical nodes and undergoes adaptive sleep management and also a reduction in the total number of transmissions made and retransmission usually caused by frame error and collision by implementing Extended Kalman Filter (EKF) is a nice panacea [11]. In [1], this can be reduced if nodes are put to sleep when other nodes communicate, collision avoidance mechanisms and cluster routing which they indicate is more energy efficient than direct and multi-hop routing. Finally since this cannot be eliminated completely, smart ICT meters can also help to reduce energy consumption to an extent by providing feedback on how much energy has been consumed and when wastage of energy has occurred. The other sections of the paper are organized as follows. Next section 2 discusses the methodology used while Section 3 presents the result and analysis of the experiments while section 4 will be the evaluation followed by conclusion, acknowledgment and references.

2.1 The System Design / System Description As shown below in figure 2, the design is made up of two laptops and one access point. The Cisco Linksys WAP610N dual-band wireless N access point with output power of 15mW is connected to Vaio mini notebook personal computer running windows 7 operating system via Straight-through to help to configure the access point for establishing connection between the with the other laptop. The other laptop which is HP pavilion dm4 notebook running on windows 7 with an i5 processor was connected wirelessly to the AP and the power consumption of this device was measured at different approximate distance from the AP while transmitting and receiving a 215mb file with Extension lead cable (10m) used for distance variation. Two plug-in mains Power meters were used for measuring the power consumption rated in Watt of the AP and the HP pavilion dm4 notebook personal computer respectively. The configuration stage indicated below describes how the access point was configured using 802.11n with 2.4GHz and 20MHz channel width. This configuration was done in the same manner for configuring other bands while varying the parameters. The SSID of the access point is given as Morpeth and it is enabled to broadcast. The Ekahau Site Survey software used for site survey shows that the D002 environment typically has over 200 access point’s or adhoc networks operating within its range. This means co-channel interference is bound to occur which is why the channel standard was set to auto channel which automatically detects which channel to use to minimise channel interference from other networks. The access point and the two wireless laptops in the network are configured with static IP addresses namely 192.168.1.1, 192.168.1.2 and 192.168.1.3 all in the 255.255.255.0 subnet.

Morpeth: 192.168.1.1/24

192.168.1. 2/24

192.168.1. 3/24

192.168.1.2/24

192.168.1.3/24 Figure 2: System Design

36

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

2.2 Experimental Environment / Implementation The three locations chosen for these experiments in the Ellison Building at Northumbria University were the D002 laboratory in block D in Ellison building with little human presence as the experiments were carried out at night time between the hours of 10 pm and 5am due to the fact that the laboratory is used for lectures and self-study during the day. The laboratory contains tables, chairs, desktop computers, scanners, a board, cabinets and an anechoic chamber and it is measured 111.85 square meters (msq). The first 4 stages of each experiment were carried out in this location. The second location is the corridor which leads to the D001 (Academic office) and the D002 and D003 laboratories. The corridor area measures 27.14msq and it was where the fifth stage of each experiment was carried out. The third location is the reception area which leads to the corridor D001 (Academic office), D002 and D003 laboratories and other locations in the building and measured 81.78msq. The sixth stage of each experiment was carried out in this location. The experiments are implemented to analyse the energy consumption in a small wireless network.

192.168.1.1/24

Varied

constant (3feet)

Transmitting/Receiving 192.168.1.2/24

192.168.1.3/24

Figure 3: Positioning of Equipment’s While Transmitting and Receiving Data This is achieved by measuring the energy consumed by the access point and one of the two computers in the network while undergoing a file transfer of 215 megabyte at approximately 5feet, 10feet, 15feet, 20feet, at a point in the D002 corridor approximately 15feet from the access point but in this case had a wall in between and lastly at a position in the reception leading to the D002 lab which was approximately 29feet. Each experiment is conducted over a period of 2 minutes to get an accurate result. The power consumption of the access point was taken down within in the first minute and that of the computer within the second minute for each experiment. The average of the power consumption range was then used to get the final power consumption tables. After this the power consumed on each side (the access point and laptop) is taken down and noted, the successful data transfer from the total of the 215 megabyte file transfer is also taken, to indicate a more efficient wireless standard that consumes the least amount of energy while performing efficiently over a period of time. The 802.11n with 20MHz and 40MHz and 802.11g/b with only 20MHz channel width were used to compare the power consumption and data transfer rate of the wireless network. As shown in figure 3 above, the measurements were taken from both the transmitting and receiving laptops and on both the transmitting and receiving interfaces of the AP with the transmitting device becoming the receiving device for the remaining sets of experiment. The positioning of equipment during the experiments shows that for all experiments one of the laptops was static at approximately 3 feet from the access point while the distance between the second laptop and the access point was varied.

37

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

3. RESULT AND ANALYSIS In order to analyze the experimental results better, several tables as shown below are used and different interpretations are provided for the results obtained. Charts are also used to convey more meaning to the tabular information. Table 1: File Transfer Comparison while transmitting Distance

Successful

file Successful file transfer

Successful file transfer

Successful file transfer

(feet)

transfer

802.11n 802.11n 2.4GHz 40MHz

802.11g 2.4GHz 20MHz

802.11b 2.4GHz 20MHz

2.4GHz

20MHz (MB)

(MB)

(MB)

(MB) 5

103

115

60.2

16.8

10

103

115

60.2

16.8

15

103

110

60.2

16.8

20

100

110

60.2

16.8

Corridor

62.6

110

60.2

16.8

50.5

60

60.2

16.8

(15) Reception (29) From the table above the 802.11n wireless network configured to use the 40MHz wide channel out performs that using the 20MHz channel based on the successful files transferred, this because the 40MHz channel allows for the passage of more amount of data, since it consists of two 20MHz adjacent channels. The 802.11g network performs less efficiently and has a constant file transfer at each stage. Finally the 802.11b wireless network has the least performance as expected since it possesses low data rates. Table 2: Access Point Power Consumption Comparison while transmitting Distance

802.11n

(2.4GHz) 802.11n

(2.4GHz) 802.11g

(feet)

(20MHz)

Access (40MHz) Access point

point

power power

(20MHz)

(2.4GHz) 802.11b AP power (20MHz)

consumption consumption (W)

(2.4GHz) Access point

power consumption (W)

consumption (W)

(W)

5

5.4

5.5

5.2

5.1

10

5.5

5.6

5.2

5.15

15

5.5

5.6

5.2

5.2

20

5.6

5.6

5.3

5.2

Corridor

5.9

5.7

5.2

5.3

5.9

5.8

5.4

5.2

(15) Reception(29)

From the table, it is noticed that 802.11n configured with both 20MHz and 40MHz consume almost the same amount of energy in average but 802.11n with 40MHz is slightly more efficient. The 802.11g network also consumes lower amount of energy during transmission with low performance while the 802.11b network with the least performance in terms of data rates has the least energy consumption.

38

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

Table 3: Laptop Power Consumption Comparison while transmitting Distance

802.11n

(feet)

(20MHz)

(2.4GHz) 802.11n

(2.4GHz) 802.11g

Laptop (40MHz)

Laptop (20MHz)

(2.4GHz) 802.11b Laptop power (20MHz)

power

power

consumption consumption (W)

consumption (W)

(W)

5

20.15

20.05

19.55

16.8

10

20.35

20.2

20.1

16.85

15

20.2

20.3

19.9

17.1

20

20.6

20.45

19.65

17.3

Corridor(15)

19.15

20.25

19.7

16.95

Reception(29)

19.3

19.35

19.4

17

(2.4GHz) Laptop

power consumption (W)

From the table above, the power consumption of the laptop when configured to use 802.11n networks consumes almost the same amount energy. It was also observed that the power consumption of the 802.11g network is nearly constant, it only has a slight increase at stage two. The 802.11b network has the lowest amount of energy consumption. Table 4: File Transfer Comparison while receiving Distance

Successful

(feet)

transfer 2.4GHz

file Successful

file Successful file transfer

Successful file transfer

802.11n transfer

802.11n 802.11g 2.4GHz 20MHz

802.11b 2.4GHz 20MHz

20MHz 2.4GHz

40MHz (MB)

(MB)

(MB)

(MB)

5

151

120

70

16.8

10

151

120

132

16.8

15

45.7

117

132

16.8

20

45.7

105

67.4

19.2

Corridor(15)

45.7

105

67.4

19.2

Reception(29)

45.7

80

60.1

16.8

Here, the first two stages of the 802.11n network configured to use 20MHz channel has the highest received file transfer for all experiments. The file transfer from the third stage onwards remains constant. The 802.11n 40MHz network file transfer reduces progressively as the distance increases due to degradation in signal strength. The first stage of the 802.11g network has a low file transfer compared to the next two stages due to delays in the reception of data. At the remaining stages, the file transfer reduces because of increase in the distance and obstacles between transmitter and receiver. The 802.11b network in this case received the lowest number of files for all networks.

39

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

Table 5: Access Point Power Consumption Comparison while receiving Distance (feet) 802.11n (2.4GHz) 802.11n (2.4GHz) 802.11g (2.4GHz) (20MHz)

Access (40MHz)

Access point (20MHz)

point

power power consumption (W)

consumption (W)

point

802.11b

(2.4GHz)

Access (20MHz)

Access

power point

power

consumption (W)

consumption (W)

5

5.3

5.3

5.3

5.2

10

5.3

5.4

5.25

5.2

15

5.4

5.4

5.2

5.2

20

5.4

5.4

5.25

5.3

Corridor(15)

5.3

5.5

5.2

5.2

Reception (29)

5.1

5.5

5.1

5.1

The power of the access points during reception is slightly less than that consumed during transmission for the 802.11n networks. As in the case of the 802.11g/b networks, the power for the transmitting/receiving stages does not indicate much difference.

Table 6: Laptop Power Consumption Comparison while receiving Distance (feet) 802.11n 802.11n (2.4GHz) 802.11g

(2.4GHz)

802.11b

(2.4GHz)

(2.4GHz)

(40MHz) Laptop power

(20MHz) Laptop power

(20MHz) Laptop power

(20MHz)

consumption (W)

consumption (W)

consumption (W)

Laptop power consumption (W) 5

14.95

15.85

16.9

14.45

10

15.25

16.1

17

14.9

15

15.9

16.2

17.25

15.3

20

16.05

16.25

16.9

14.85

Corridor (15)

16

16.1

17

14.6

Reception(29)

14.85

16

16.9

14.95

The table above shows that the laptop power consumption increases slightly with some fluctuations. At the last stage of the 802.11n network utilising the 20MHz channel the power consumption reduces due to delays and degradation in the signal strength.

40

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

4. CONCLUSION Based on the experiments conducted and the results obtained, the power consumption of the 802.11n networks is more than that of the 802.11b/g networks in both transmitting and receiving modes, although these results show that there is not much difference. It has been observed that the 802.11n networks are more effective, especially for data transfer where in most cases it doubles that of the 802.11b/g networks, but the power consumption in general is small, and it can be said that in the active mode for example transmitting the 802.11n will consume more and perform more effectively while in the inactive mode it will be better to use the 802.11b/g networks. In IEEE 802.11 wireless network, higher data rate produces higher throughput, but also higher energy consumption. Therefore, to achieve minimum energy consumption by these devices, appropriate data rate based on the available standards should be chosen for network implementation.

[4]

[5]

ACKNOWLEDGEMENT The author would like to thank the Staff of Northumbria Network (Labs D002 & D003), School of Computing, Engineering and Information Sciences, Northumbria University, Newcastle Upon Tyne, UK.

[6]

[7] REFERENCES [1] Aarthy, M. P. A. M. D. (2011). Energy conservation at node level using a wake-up scheme in WirelessSensor Networks Emerging Trends in Electrical and ComputerTechnology (ICETECT),2011 [Online].Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&r number=5760286[Accessed3 November 2011 [2] Chilamkurti, N., Zeadally, S. & Mentiplay, F. (2009). Green Networking for major components of ICT Systems. Available: http://downloads.hindawi.com/journals/wcn/200 9/656785.pdf [Accessed 15 September 2011]. [3] Das, A. D., S. (2011). Power conservation in Wireless Sensor Networks: A graph- theoretic approach Information Sciences and Systems (CISS), 2011 45th Annual Conference on [Online]. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp= &a rnumber=5766123 [Accessed 1 October 2011].

[8] [9]

[10]

41

Despin, C., Labeau, F., Labelle, R., Cheriet, M., Thibeault, C., Gagnon, F., Leon-Garcia, A., Cherkaoui,O., St. Arnaud, B., Mcneill, J., Lemineux,Y. &Lemay, M. (2011). Leveraging green Communications for Carbon Emission Reductions: Techniques,testbeds, and emergingcarbon footprint standards. Communications Magazine, IEEE [Online], 49. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp = &a rnumber=5978422[Accessed 1 October 2011]. Han, Y., Li, H. & Qiu, J. (2011). The Analysis and Summary about Energy saving Technologies of wirelesssensor network, Electronic andMechanicalEngineering and InformationTechnology (EMEIT) 2011International Conference on[Online], 2. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp = &a rnumber=6023235[Accessed 5 October 2011]. Hodges, R. W., W. (2009). Go Green in ICT. Available: http://www.greenit.net/whygreenit.html [Accessed 11 September 2011]. Huang, C. M., Ku, H. H. & Kung, H. Y. (2009). Efficient power-consumption-based load-sharingtopology control protocolfor harsh environments in wirelesssensornetworks.Communications,IE T [Online], 3.Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp = &a rnumber=4912240[Accessed29October 2011]. IT., G. (2010). The green IT review [Online]. Available: http://www.thegreenitreview.com/ 19 August 2011]. Panchabhai, A. M. B., C. W. . (2004). Node hibernation protocols for conservation of energy in wireless sensor networks,Military Communications Conference, 2004.MILCOM 2004. IEEE [Online], 3. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp =&arnumber=1495190 [Accessed 28th September 2011]. Riaz, M. T., Gutierrez, J. M. & Pedersen, J. M. (2009). Strategies for the next generation green ICT Infrastructure. Applied Sciences in Biomedical and Communication Technologies, 2009. ISABEL 2009. 2nd International Symposiumon [Online]. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp =rnumber=5373604 [Accessed 3 October 2011]

Vol 6. No. 5, December 2013 African Journal of Computing & ICT © 2013 Afr J Comp & ICT – All Rights Reserved - ISSN 2006-1781 www.ajocict.net

[11] Song, C., Sharif, H. & Nuli, K. (2005). Study of an adaptive frame size predictor to enhance energy conservationin wireless sensor networks SelectedAreas in Communications, IEEE Journal on [Online], 23. Available Online at http://ieeexplore.ieee.org/stamp/stamp.jsp?tp =&arnumber=1391037[Accessed28 September 2011]. [12] Xue, C. J., Xing, G., Yuan, Z., Shao, Z. & Sha, E. (2009). Joint Sleep Scheduling and Mode Assignment in Wireless CyberPhysical Systems DistributedComputing Systems Workshops, 2009. ICDCSWorkshops '09. 29th IEEEInternational Conference on[Online]. Available: http://ieeexplore.ieee.org/stamp/stamp.jsp?tp =&a rnumber=5158825[Accessed23 November 2011].

Author’s Brief

Philip Asuquo is a Network Engineer and a lecturer in Department of Computer Engineering, University of Uyo, Nigeria. He obtained a B.Eng Computer Engineering at the University of Uyo, Akwa Ibom, Nigeria in 2008 and a Masters Degree in Computer Network Technology at Northumbria University, Newcastle Upon Tyne, United Kingdom in 2011. His research focuses on technologies for increased spectral efficiency and heterogeneous network deployments to improve energy efficiency and dynamic resources allocation and sharing capabilities. He can be reached by phone on +234 8023863314 and through E-mail [email protected]

Itoro Akpabio is a Lecturer in the Department of Electrical/Electronics and Computer Engineering, University of Uyo, Uyo, Nigeria where she has carried out teaching, research and consultancy on information system security (ISS) for both government and private institutions. She completed a B.Sc in Electronics and Electrical Engineering at Obafemi Awolowo University. Osun State, Nigeria. She then undertook an M.Sc in Management and Information Systems (MIS) at the University of Manchester, UK. Her research interests lies in the area of Information and Communication technologies ranging from its introduction to the implementation of ISS theories with a focus on the mitigation of insider attacks. She has also collaborated actively with other researches in other disciplines of Electrical/Electronics and Computer Engineering on better techniques for improving the power consumption of communication devices. She can be reached by email on [email protected] and through the mobile number +2348034011162.

42