Performance Analysis of Hybrid Broadband Access Technologies Using PLC and Wi-Fi

1



Performance analysis of hybrid broadband access technologies using PLC and Wi-Fi

Oscar Andres Gonzalez, Juraj Urminsky, Miguel Calvo, Leandro de Haro

Abstract— While Power Line Communications deployment is gaining momentum higher capacity is being demanded. Actually one possible solution to obtain an improved high speed broadband access can be the use of a hybrid technology in an indoor environment. This paper shows the advantages that could be obtained if the power line communication is joined with the Wi-Fi technology.

Index Terms— PLC, Wi-Fi, Throughput, Hybrid access.

I.INTRODUCTION

I

n the last years it can be observed that the PLC technology is consolidating as an option to obtain a high speed broadband access with a minimum investment in infrastructure cabling. Actually, this is one of the main problems that the telecommunications operators are trying to solve to obtain benefits in shorter periods of time and it looks that the power line communications can be a very good option to overcome this economics situation.

Nowadays, big telecommunications companies are thinking in the possibility of developing a new high speed broadband access based in the joint of different transmission technologies. These technologies could be the PLC technology extended with wireless access allowing the use of mobile terminals to the end users so they can connect ubiquitously to the external data or internal control home network . This new environment where different technologies can join to provide a better broadband access has motivated this work in which we will be show a comparative study of the performance of access technologies like PLC, wireless and their combination .

The paper is organized as follows. The measurement procedure is presented in II, obtained measurements and performance analysis is presented in III and conclusions are drawn in IV.

II.Measurement procedure

A.Transmission characteristics measurement procedure of a PLC technology communication system

The evaluation procedure of a PLC communication system follows. The parameters that will be used in the evaluation of the system will be the signal throughput, transmitted and received packet statistics (lost packets rate and error packets rate) and the time delay between transmitter and receiver.

The throughput will be measured in Kbytes/second and represents the number of UDP packets transmitted by the power line channel.

The transmission procedure is based in the generation of UDP packets from a transmitter computer operating with Linux towards a receiver computer that operates in Windows. It is required to program in Linux a socket between the two computers to generate the UDP packets that will be counted later. In the transmitter PC we use the “tcpdump” program as a protocol analyzer that allows to know the number of transmitted packets. In the receiver PC we use “Ethereal” as protocol analyzer. This program allows us to observe the protocol stack of every frame that is received, which allows as to extract conclusions about the performance of the communications systems we are analyzing.

Ethereal shows a real-time graph of the throughput vs. time allowing to observe the channel performance in real time and how subsystems change their communications parameters to adapt themselves to the variable transfer function of the power line channel.

It is important to obtain also statistics of the packets that are sent/received. The packets parameters measured are the following:

a) Sent packets: the number is obtained with the “tcpdump” protocol analyzer in the transmitter PC.

b) Total received packets: this number is obtained with the “Ethereal” analyzer protocol, using a filter packet to extract only UDP packets and discarding other packets like ICMP, SNMP, etc….

c) Erroneous received packets: erroneous packets are measured using Ethereal by filtering the UDP packets whose checksum field is erroneous.

d) Lost packets: this number is calculated subtracting the number of sent packets minus total received packets

e) Correct received packets: this number is calculated subtracting the number of total received packets minus erroneous received packets.

Finally, the time delay is estimated using the “ping“ application, based in ICMP protocol. This application allows to know the delay time of every ICMP packet and, moreover, it can be known the average time delay if it considers all packets transmitted since the beginning of the application. Through to this parameter, conclusions will be obtained about the channel characteristics and about the communications system architecture.

B.Comparison between the performance of a non hybrid (Wi-Fi) and hybrid (PLC and Wi-Fi) broadband access technology

The main objective of this work is to make a comparison between a non hybrid and hybrid access technology. We understand hybrid access like the ability to join two or more different access technologies like wireless or cable (UTP, fiber, coaxial, etc…) to increase the broadband speed and the quality of service that actually telecommunications operators are providing to the end users.

Although the power line technology provides access to the data network in whatever part of the house, the use of terminals with mobile applications like PDAs or laptops force us to make compatible different technologies that always have been used in very different environments.

In this paper, conclusions are obtained on the added value of a hybrid access technology compared with the use of a single technology (power line or wireless).

Figure 1 shows the setup for the measurement of throughput, statistics of packets and time delay using a hybrid access with HomePlug 1.0 bridges or DS2 ‘s first generation 45Mbps technology modems and Wi-Fi.

Fig 1.Measurement setup for hybrid access with HomePlug 1.0/DS2’s first generation 45Mbps technology and Wi-Fi

In the wired part of the path, PLC access technologies are used, between the transmitter and the wireless access point through two bridges/modems from HomePlug 1.0/DS2’s first generation 45Mbps technology respectively. When using DS2 technology an intermediate element named Head-End that has control functions in the communication between the two extremes is also required . In this trial, it will use the CPE-HE-CPE topology ,that’s to say,a transmitted signal whose destination address is in the PLC network must cross the Head-End, that will process and send it towards the end terminal.

In the wireless part of the path a radio technology based on 802.11b/g is used to connect the radio access point to the Wi-Fi modem of the mobile computer.

One of the most important characteristics for the joint use of these two access technologies is the Ethernet interface that makes both technologies compatible. This common interface (through an RJ-45 cable) helps to integrate two different subsystems whose communication channels are totally different.

In figure 2 the measurement procedure for the 802.11b/g access technology is presented. The quality of the wireless link will depend on the number and shape of the elements interrupting the signal between the transmitter and receiver and the attenuation that they produce to the radio signal in the 2.4 GHz frequency band. The connection between transmitter and access point will uses an Ethernet cable.

Fig 2. Measurement procedure using an non-hybrid technology based in 802.11b/g

III.Results of the performance analysis of a hybrid access technology

The non hybrid and hybrid access performance in three different environments: a laboratory, a four-floor residential house and a one-floor apartment have been analyzed. Only DS2’s first generation 45 Mbps technology equipments were finally used due to the lack of compatibility between 802.11b/g and HomePlug 1.0 [1], [2] using the “PowerLine Ethernet Bridge” when the channel conditions are strongly degraded. The solution to this problem is to work with an “802.11x –HomePlug 1.0 protocol adapter“ from HomePlug Alliance that actually are available commercially.

A.Measurements in laboratory

Firstly, measurements were made in a laboratory, a place with long free space and too many connected equipments in the electric network that can produce an important noise level in the power line. In figure 3 the different locations into the laboratory where these measurements were made is shown. In this case, the PLC path is very short because the transmitter (point 2), the receiver modem (point 1) and Head-End (point 1) are very near from the access point (point 1).Four measurement points (A,B,C,D) have been chosen for this trial in a university’s laboratory, which are separated from the transmitter by crystal and plastic walls.

Fig 3.Measurements points for performance analysis of an hybrid access technology in the laboratory

In figure 4 it can be seen the throughput measured in location D. The non hybrid access is provided between the access point (point 1) and the receiver mobile (location D). It can be observed that throughput is similar for the hybrid and non-hybrid access. In this case the number of lost and erroneous packets was zero and the only noticeable difference is a greater variance in the throughput parameter in the case of hybrid access due to the short PLC path before arriving to the access point.[3] . Throughput mean value in these measurements is 40000 bytes/sec.

In general, the performance of both technologies is very similar in all locations and the differences are minimum.

Fig 4.Throughput measured in laboratory in location D with hybrid access (up) and non-hybrid access (down)

B.Measurements in a four-floor residential house

Figure 5 shows the physical distribution of the floors and the locations of the points where the measurements were made.

The house environment is more quiet than in the laboratory and the noise measured levels are lower. This house was empty, without connected appliances to the electric network and without furniture elements inside, so the multipath effect in the air is notably reduced. It is interesting to observe the performance of both access technologies when the communication is realized between different floors.

Fig 5.Distribution of location points in the four-floor residential house

Figure 6 (up) shows the measurements performed in the same floor (point F to G), where F-G is the PLC path, the access point is situated in G, the Head-End equipment is situated in the third floor (E point) and mobile receiver is in F. Lower part of the figure is the throughput between F-G using Wi-Fi access. It can conclude that the performance is very similar in both cases although the time delay in the hybrid access is 95 msec while in the non-hybrid access is 2 msec. This higher time delay is due to the longer path when the signal must cross the HE and return to the point G. In these measurements UDP packets have been sent from transmitter during a 2 minutes time period. Another difference noticed was the lost packets rate in the hybrid access (9.5%) compared with 4.5% in the non-hybrid access during this two minutes. The cause of this difference was the worst conditions of the mixed access communication channel. Throughput mean value in these measurements is 115000 bytes/sec.

Fig. 6.Throughput measurement with a PLC path between F-G, the access point in G, HE in E and receiver in F

Figure 7 (up) shows the throughput performance in the hybrid access when the PLC path is F-G, HE is situated in D point, access point is G and mobile receiver is in D. In the down part of the figure we show the throughput in the non-hybrid case when the Wi-Fi transmitter is in the G point and the Wi-Fi receiver is in point D. In this situation it can seen that the average value throughput is very similar in both cases although the its variance is greater for the hybrid situation. This is produced by the worst conditions of the PLC path as compared to the non-hybrid situation.

Fig. 7. Measured Throughput with a PLC path between F-G, the access point in G, HE in D and receiver in D

The packet lost rate is 11,87% while in non-hybrid situation is 12%, which demonstrates that the performance of both systems is very similar. Time delay in the first case is 54 msec and in the second case is 1 msec. Throughput mean value in these two situations is 80000 bytes/sec.

Figure 8 (up) shows the throughput performance of the hybrid system when the PLC path is F-H, HE is situated in E point, access point is in H and mobile receiver is in G..

Fig. 8.Throughput measurement with a PLC path between F-H, the access point in H,HE in E and receiver in G

The lower part of the figure shows the throughput when only the Wi-Fi link is used to transmit from G to H. It can be observed that the throughput mean value in the hybrid case (80000 bytes/sec) has a better performance than in the non-hybrid case (45000 bytes/sec). During 2-minutes transmission time period the packet lost rate was 24,11% and in the non-hybrid case 22%, which demonstrates that if we use only Wi-Fi the number of discarded packets will be lower and the number of send packets will be also lower. The time delay in the hybrid case is 86 msec and in the non-hybrid case is 3 msec

Figure 9 shows the throughput performance when the PLC path goes from point A to point C, HE is situated in B, access point is in C and the mobile receiver is in J .In this situation, we can not use only Wi-Fi because the signal arriving to the four floor is very weak. The packet lost rate is 45% and the time delay is 137 msec. Throughput mean value is 65000 bytes/sec. This implies that the channel conditions are not very good so it will be necessary to adequate the type of applications to be used in this environment. It is demonstrated that an efficient use of the hybrid access allows to improve the broadband access in this four-floor house. In the broadband access design it is very important to locate correctly along the house the intermediate elements like HE, repeaters and regenerators to obtain a high signal quality in all parts of the house.

Fig. 9.Throughput measurement with a PLC path between A-C, the access point in C,HE in B and receiver in J

C.Measurements in a one-floor apartment

Measurements were also made in an inhabited apartment environment. Compared with the residential situation we have now a greater number of electrical appliances connected into the electrical network, a higher number of persons living in the floor and a number of in place furniture.

In figure 10 it can see the measurements points used in this new environment.

Fig.10.Measurements points in the one-floor apartment

Figure 11 (up) shows the measurements when the PLC path is B-A, the access point is in A, HE is situated in L and the mobile receiver is in point B. In the lower part of the figure we are transmitting from point B to point A using only Wi-Fi. It can be observed that the average throughput is the same in the two situations and the variance is very similar too. During the 2-minutes transmission the packet loss rate with hybrid access was 61,2%, very similar to58,9% in the hybrid access situation. The time delay in the first case was 70 msec and in the second case was 2 msec. The longer path that signal must cross to the receiver point through the HE is the main cause of this considerable delay difference. Throughput mean value in the hybrid situation (up) is 75000 bytes/sec and in the non-hybrid situation (down) is 70000 bytes/sec.

Fig. 11.Throughput measurement with a PLC path between B-A, the access point in A, HE in L and receiver in B

Figure 12 (up) shows the measured throughput when the PLC path is B-E, the access point is in E, HE is situated in L and the mobile receiver is in point B. In the lower part of the figure we are transmitting from the point B to point E using only Wi-Fi.

It can be observed that the performance is very similar in both situations although the hybrid case shows a 100 seconds time period where the modem has adapted from a initial worse to a better final channel condition. Throughput mean value in the hybrid situation during the last 60 seconds (up) is 70000 bytes/sec and in the non-hybrid situation (down) is 70000 bytes/sec.

The packet lost rate in the hybrid case was 63% and in the Wi-Fi case is 43,5%. Time delay in the first case is 75 msec and in the second case is 1 msec.

Fig. 12.Throughput measurement with a PLC path between B-E, the access point in E, HE in L and receiver in B

IV.Conclusion

The performed measurements demonstrates that throughput performance of the hybrid and non-hybrid access is very similar if the communication is made in the same floor or between two floors. If the communication is made between more than two floors measurements shown that hybrid technology can improve the signal quality in locations that can not be covered with Wi-Fi. It is very important to plan correctly the physical location of the intermediate elements like repeaters and regenerators to achieve a high quality service.

ACKNOWLEDGEMENTS: This work was supported by the European Commission 6FP project OPERA. Help from Iberdrola ( and DS2 ( partners is greatly appreciated.

References

[1]Yu ju Lin, “A comparative performance study of Wireless and Power Line Networks”,IEEE Communications Magazine, April 2003, pp. 54-63.

[2]Myoung-Hee Young ,”MAC throughput analysis of Homeplug 1.0”, IEEE Communications Letters,vol.9,no. 2,February 2005

[3]Matthias Gotz, ”Power Line Channel Characteristics and their effect on Communication System Design”, IEEE Communications Magazine; April 2004

Manuscript received March 15, 2005. (Write the date on which you submitted your paper for review.)

Oscar Andres Gonzalez is with the Signals Systems and Radiocommunications Department in Polytechnic University of Madrid., e-mail: oscargon@ alumnos.etsit.upm.es.