Sei sulla pagina 1di 5

Scalability of DVB-H Deployment on Existing Wireless Infrastructure

Aurelian Bria
Royal Institute of Technology, Sweden Email: aurelian@kth.se

David G mez-Barquero o
Polytechnic University of Valencia, Spain Email: dagobar@iteam.upv.es

Abstract This article investigates the scalability of DVBH deployment on existing wireless infrastructure for digital broadcasting and cellular networks. Three deployment strategies are investigated: when only the broadcasting and the cellular infrastructure are used separately, and the case when both can be used jointly. The outcomes show that, in general, the investment needed for increasing the service capacity by switching to higher order modulation, or for increasing the coverage level over a given service area, lead to a higher annualized cost per Mb/s.

employed for digital Radio/TV broadcasting and cellular systems. We approach the problem from the perspective of both broadcasting and cellular operators, that hold infrastructure assets. The measure of interest is the cost per Mb/s, as a function of the transmission powers and number of cellular sites needed, when the network migrate from low to high service capacity and/or coverage. This study is motivated by the results the authors obtained in a preliminary investigation presented in [4]. The conclusions were that deployment of DVB-H networks with indoor coverage is a very challenging task, requiring very large transmission powers and additional sites. The envisaged requirement for DVB-H channel capacity of 10 Mb/s (per 8 MHz RF channel), as suggested in [2], is difcult to achieve in realistic environments, especially indoor. Re-using existing broadcasting sites for DVB-H is not enough to achieve the same coverage level as for DVB-T services. Favorable propagation conditions of DVB-T (e.g. xed rooftop antennas, Lineof-Sight) lead to large coverage radius of the broadcasting towers (e.g. up to 70 km and more), but the same does not hold for portable reception characterized by Non-Line of Sight, multi-path and Doppler impairments, strong propagation loss (especially for indoor reception), and poor receiving antenna gain. Re-using cellular sites for complementary transmitters can lower the transmission powers required at the broadcasting sites, but this implies a cost proportional with their number. Moreover, the required number of sites needed is large, since the cell radius of the cellular network should be short enough (e.g. less than 3 km) in order to avoid excessive transmission power (e.g. over 30 dBW EIRP). Obviously the number of additional sites needed depends on the allowed maximum transmission power and antenna height. The results obtained in [4] are in line with public DVB-H implementation guidelines, like [5], where it is stated that DVB-H networks must be more dense and should use more robust transmission modes than DVB-T networks to achieve satisfactory coverage levels. The rest of the paper is organized as follows: Section II introduces the basis of DVB-H deployment and identies the major bottlenecks, Section III discusses the cost of implementation, Section IV describes the system model used for numerical evaluations, Section V presents the results and Section VI consists of the main conclusions.

I. I NTRODUCTION TV in the mobile is the hot topic of today. 3G cellular networks have already started to offer streaming media services, as videoclips from sport events or TV programs, but it proves to be a very challenging and resource demanding task. The alternative technology to cellular networks is digital broadcasting. In Europe, DVB-T (Digital Video Broadcasting Terrestrial) is the standard for terrestrial digital TV broadcasting. Although the ability of DVB-T to serve mobile terminals (using diversity antenna receivers) has been proven, DVB-T is not practical for handheld terminals (small, lightweight, portable, single antenna reception and especially batterypowered). DVB-H (Digital Video Broadcasting - Handheld) [1], is a recent extension of DVB-T to reach mobile and handheld terminals. DVB-H is backwards compatible with DVB-T, and the main features introduced are: a) time slicing - reduces the average power consumption of the terminal and enable smooth and seamless handovers b) the 4K mode - improves Single Frequency Network (SFN) planning exibility, and c) MPE-FEC (Multi Protocol Encapsulation - Forward Error Correction) - an optional coding scheme at the data link layer to improve the robustness. A detailed description of these features can be found in [2] and [3]. DVB-H is generating signicant interest among the wireless operators, since it allows the transmission of IP mass multimedia content to mobile phones at very high data rates. Broadcasting operators will probably retain a leading role since they can easily introduce the technology in their current networks (e.g. DVB-H can co-exist with DVB-T in the same multiplex). Furthermore, spectrum assigned for DVB-H services coincides with broadcasting frequency bands. As DVB-H is intended for mobile users, it presents interest for cellular operators as well. They may consider deploying DVB-H on their existing sites. The goal of this paper is to evaluate the scalability of DVBH deployment on the existing wireless infrastructure, presently

II. DVB-H P LANNING AND D EPLOYMENT Planning of the DVB-H transmission mode is based on the choice of four parameters: FFT size, relative Guard Interval (GI), modulation and Coding Rate (CR). The FFT size (8K, 4K, or 2K) is related to the number of OFDM sub-carriers, their inter-carrier distance and the useful symbol period. The choice of the FFT size has no impact on the capacity, but on the trade-off between mobile reception (maximum speed) and SFN cell sizes. The 2K is the most suitable for mobile reception, whereas the 8K gives the largest SFN cell size. The relative guard interval can have a duration of 1/4, 1/8, 1/16, or 1/32 of the useful symbol duration and together with the FFT sizes gives the absolute guard interval value. The choice of the relative guard interval represents a trade-off between capacity and SFN cell size. Modulation and coding rate determines the minimum required Carrier-to-Noise Ratio (CNR) and, together with the relative guard interval, the channel capacity. An important enhancement for DVB-H network planning is the MPE-FEC scheme, which complements the physical layer FEC of DVB-T. First conclusions on the MPE-FEC performance can be found in [6] and in [7]. The outcome of these studies show that use of MPE-FEC makes the DVB-H service availability almost independent of the speed of the terminal, obtaining a similar performance as mobile DVB-T receivers with antenna diversity. However, the use of MPE-FEC for pedestrian users does not turn in any benet, since only an approximated reduction of 2 dB in the CNR requirement can be achieved at the expense of 25% parity overhead, [8]. Broadcasters may choose out of the three different network congurations envisaged for DVB-H: Existing DVB-T network sharing the multiplex with DVB-H. Hierarchical transmission of DVB-T (low priority stream) and DVB-H (high priority stream) in the same multiplex. DVB-H dedicated network. The main problem of sharing a multiplex with DVB-T and DVB-H is that the transmission mode is unique, since both systems are mixed in a single transport stream. As DVBT is intended primarily for stationary terminals with rooftop antennas and large SFN cell sizes (usually 8K mode), the DVB-H service availability will be considerably reduced for moving and indoor terminals. This conguration is the most immediate implementation of DVB-H. Another option is the use of the hierarchical modulation, so DVB-T and DVB-H are transmitted in separated transport streams. However, its less robust than some standard modulations and the maximum capacity for DVB-H is only 5.5 Mb/s [6]. The third option, a dedicated network, can take advantage of the physical layer extensions of DVB-H, the 4K mode and in-depth interleaving. The most important benet is the possibility to plan the network for the desired coverage and capacity without any constraint imposed by DVB-T. The drawback is that additional core network equipment and a separate frequency channel are needed. Cellular operators may go alone for a dedicated DVB-H

network, forming dense SFNs, by using all or only a part of their existing sites. A combination of cellular and broadcasting infrastructure is also possible. Cellular sites can be employed as a complement for the Radio/TV broadcasting towers, acting not only as gap-llers, but also as synchronized transmitters. In the following subsections we describe each implementation strategy focusing on the required infrastructure investment, achievable service capacity, and coverage. A. DVB-H on the existing Broadcasting Sites Broadcasters might start offering DVB-H services by using shared DVB-T/H networks, although they will reach very poor indoor coverage levels of DVB-H, since most DVB-T networks are planned for rooftop reception. Thus, broadcasters will sooner or later deploy dedicated DVB-H networks in order to increase the coverage, and also the capacity, by using a more robust transmission mode. However, it is not possible to trade-off enough capacity to achieve satisfactory indoor coverage [4]. Due to these reasons, either the transmission power of the broadcasting sites should be very large, or additional sites have to be deployed. Many digital TV sites will not be able to run over todays transmission powers until the analogue switch-off (to avoid interference with the analogue TV coverage). Another limitation is imposed by the electromagnetic exposure limits specied by the current international regulation [9], which is more or less restricting the EIRP (Equivalent Isotropic Radiated Power) of the big TV towers to around 60 dBW (1MWatt). The option of deploying additional sites is appealing only if the cost of this investment can be recovered by the revenue stream generated by the new service. B. Cellular DVB-H In an SFN many receiving locations are served by more than one transmitter, introducing redundancy in the transmitted signal and improving the coverage, especially when portable indoor reception is required. Previous investigations into the performance of large SFNs have shown that self-interference issues are not problematic, e.g. [10]. Results in [4] show that with this conguration it is possible to achieve indoor coverage by re-using a large number of cellular sites at moderate EIRP levels (30 dBW, upper limit imposed by [9] and interference issues to the cellular base stations, especially GSM in 900 MHz). The drawback is that the cost of this deployment is proportional to the number of cellular sites used. C. Hybrid Cellular-Broadcasting Infrastructure Deployment of DVB-H on both types of infrastructure brings the benet of the re-using existing cellular sites for low power transmitters working together with the high power transmitters at the broadcasting sites in an SFN. Both indoor and outdoor coverage may come at a lower cost and in a more scalable manner, by avoiding investments in new sites or in EIRP increase at the Radio/TV broadcasting tower. The drawback is that cellular sites were chosen under a different design process and their position cannot be changed. Results

in [4] are promising and show that for indoor coverage, a 10 dB difference in broadcasting site power (from 50 to 60 dBW) can be compensated by employing 30 cellular sites at 30 dBW. In the rest of the paper we investigate the relationships among transmission power (EIRP), achieved coverage, and number of cellular sites for the presented strategies for the deployment of dedicated DVB-H networks on existing wireless infrastructure when increasing the service capacity and the coverage level over a given service area. III. D EPLOYMENT C OST AND S CALABILITY I SSUES This section introduces simple cost models for each of the mentioned DVB-H deployment strategies. Numerical results obtained from simulations are used to evaluate the cost per Mb/s of the service capacity. We are interested in if this cost decreases, remains constant or increases, when the service capacity and coverage level are increased over a given service area. The economic success of the DVB-H networks will be partly determined by the dynamics of the cost per Mb/s. An ideal situation would be to experience economies of scale (the value of cost per Mb/s decreases when more capacity is provided over the service area). In the following, all the costs are expressed as annualized costs (i.e. cost per year). The annualized costs for a broadcasting site can be written: Cb (R) = C0b + C1b (R) + C2b Pb (R)

(antennas, cables, transmitter, etc.), and other installation costs [11]. Since the EIRP cannot be as high as at the broadcasting TV towers, its impact on the total costs is considered to be negligible. In a hybrid cellular-broadcasting network, the number of cellular sites needed depends on both the capacity, R, and the EIRP at the broadcasting site Pb . The annualized cost of operation is a modied sum of (1) and (2): Ch (R) = C0b + C0c + C1b (R) + C2b Pb + + [C1c (R) + C2c ] n(R, Pb ) (3) A quick look on the described models let us notice several helpful aspects. For the case of DVB-H deployment on one site (e.g. shared with DVB-T) the cost is more or less proportional with the EIRP, if we assume that C1b does not vary too much with the capacity. For cellular deployment, the cost is clearly proportional with the number of employed sites. In order to obtain the cost per Mb/s the above costs should be divided by R - the provided DVB-H service capacity. As such, the relevant performance measures that are directly affecting the cost gures and scalability of deployment are: Pb /R and n/R. IV. S YSTEM M ODEL For the investigations we have assumed a scenario corresponding to a medium size city, where the TV tower is situated close to the city-center. Our system model consists of a hexagonal service area of 25 km radius, a broadcasting tower situated in the middle of it, and cellular sites that are placed following a homogeneous hexagonal lattice of cells. We have assumed the specic 4K mode and a relative Guard Interval (GI) 1/4. This gives the same absolute guard interval of 112 s and equivalent guard interval distance of 33.6 km, than most DVB-T networks (8K mode with GI 1/8). Three different modulations (QPSK, 16-QAM and 64-QAM) and two coding rates (1/2 and 2/3) have been considered. The coverage targets are 70%, 90% and 95%. DVB-T/H service capacity (Mb/s) and CNR requirement for good portable reception for different modulation and coding rates (MCR) [12] are listed in the Table I. Two reception conditions have been considered: pedestrian DVB-H indoor and outdoor. The performance measure of interest is the Signal-toInterference plus Noise Ratio (SINR) [4]. A receiving location is considered covered if its SINR fulls the CNR requirement . The interference model takes into account only selfinterference from the network and no external interferences. Note that the TV tower does not create any self-interference in the service area, since the service area radius is smaller than the equivalent guard interval distance. The self-interference, if exists, is only due to the cellular sites. The value considered for the cell radius of the cellular network is 2.5 km, giving 109 sites in the service area. The antenna height is 150 m for the TV broadcasting tower and 35 m for the cellular sites. Ideal omni-directional antennas have been assumed at all sites. The EIRP for DVB-H at the cellular sites have been optimistically assumed to be 30 dBW. In reality

(1)

C0b accounts for discounted cost of equipment, site rental, maintenance, licence, etc. C1b represents the cost of transmission capacity (backbone network) that can be over leased lines, optical ber or microwave links. This cost is usually an increasing function of the capacity, R[b/s]. C2b is the cost per EIRP unit (Watt), and accounts mainly for discounted cost of power ampliers and antenna system and annual expenses for electricity. Pb is the EIRP of the broadcasting site. At low EIRP levels (up to 40 dBW, for example) the rst two cost components are dominant. On the other hand, for high EIRP levels as 50 or 60 dBW, the annualized equipment cost (number of power amplication modules, larger antenna) and running costs (electricity and maintenance) are dominant. For example, the annual electricity bill for a 50 dBW EIRP transmitting site is not negligible. For a cellular DVB-H implementation the annualized cost of operation is: Cc (R) = C0c + C1c (R) n(R) + C2c n(R)

(2)

C0c accounts for core network cost (e.g SFN related, MUX, etc.). C1c is similar to C1B . C2c accounts for site rental, maintenance, electricity, licence, etc. n is the number of cellular sites employed in SFN. The dominant part of expenses will be represented by transport network, additional power supply needed, equipment

TABLE I MCR QPSK 1/2 QPSK 2/3 16-QAM 1/2 16-QAM 2/3 64-QAM 1/2 64-QAM 2/3 CNR(dB) GI 1/4 11.5 14.7 17.2 20.9 22.1 25.5 4.98 6.64 9.95 13.27 14.93 19.91 Bit Rate (Mb/s) GI 1/8 GI 1/16 5.53 7.37 11.06 14.75 16.59 22.12 5.85 7.81 11.71 15.61 17.56 23.42

20
64QAM 2/3

(27.3)

(155.9)

(361.1)

18

GI 1/32 6.03 8.04 12.06 16.09 18.10 24.13


Capacity [Mb/s] 14
16QAM 2/3
(81.1)

16
64QAM 1/2
(95.0)

12 10 8
(46.1)

16QAM 1/2

this may be lower due to possible interference caused to the cellular system, especially GSM in 900 MHz. Link budget values used are the ones proposed in [13] for DVB-T coverage planning. An omni-directional antenna with a gain of -5 dBi is assumed for the DVB-H terminals. The operating frequency is 700 MHz. Since only pedestrian users are targeted, MPE-FEC is not considered. The shadow fading has been implemented by means of three uncorrelated lognormal distributions, namely: outdoor, building and microscale, with standard deviation values of 5.5 dB, 6 dB and 3 dB respectively. Indoor users enjoy three types of shadowing, whereas building shadowing is not considered for outdoor users. The building shadowing follows a truncated distribution, so its sum with a mean building penetration loss of 7 dB never becomes a gain. The micro-scale shadowing accounts for the fast fading and depends only on the users position, and has the same value for all links between the user and all sites. On the other hand, the outdoor and building shadowing are computed independently for each path. No spatial correlation has been considered in our study. The path loss model given in the recommendation ITU-R P.1546 (curves for the 50% percentile of the time) [14] has been used for the broadcasting tower as recommended in [13]. As the curves are valid for a receiver antenna height of 10 m, a height path loss correction factor of 12 dB has been used to account for 1.5 m receiver antenna height. The sub-urban Okumura-Hata path loss model has been used for the cellular sites. V. N UMERICAL R ESULTS Fig. 1 shows the achievable capacity as a function of the EIRP for the case when only the broadcasting tower is employed, for indoor and outdoor reception and the coverage targets considered. Table II contains the values of the ratio Pb /R [Watts/(b/s)] for the points of Fig. 1. As the ratio Pb /R increases when switching to higher modulation and coding rate, it means that extra Mb/s will come at higher cost if the same coverage requirement is kept. Keeping the same capacity but increasing coverage leads again to a signicant increase of the ratio Pb /R. Moreover, there is almost an order of magnitude difference in cost when the system is dimensioned for indoor reception, compared with outdoor. Note that the difference in EIRP between outdoor and indoor reception is larger for higher coverage targets. This is due to the effect of the building shadowing, that makes more difcult to

QPSK 2/3

(38.9)

6
QPSK 1/2
(24.8) (4.3) (57.5)

Outdoor 70% Outdoor 90% Outdoor 95% Indoor 70% Indoor 90% Indoor 95% 65 70

4 35

40

45 50 55 60 EIRP [dBW] from Broadcasting Tower

Fig. 1. DVB-H service capacity vs. EIRP under different coverage requirements, for one broadcasting tower. Service area radius is 25 km. TABLE II MCR Coverage Target 95% 90% Outd. Ind. Out. Ind. 4.8 7.6 9.0 15.8 18.5 30.3 57.5 90.1 106.9 187.8 220.1 361.1 2.6 4.0 4.7 8.3 9.8 16.0 24.8 38.9 46.1 81.1 95.0 155.9

70% Outd. Ind. 0.7 1.1 1.3 2.2 2.6 4.3 4.3 6.8 8.1 14.2 16.6 27.3

QPSK 1/2 QPSK 2/3 16-QAM 1/2 16-QAM 2/3 64-QAM 1/2 64-QAM 2/3

achieve coverage targets above 80%. We point out that DVB-T deployment is not scalable either, but as the EIRPs required for rooftop reception of DVB-T are considerable lower than for indoor reception of DVB-H (approximately three orders of magnitude [4]), the costs of having high capacity (64-QAM) are affordable. Some examples of the achievable service capacity as a function of the number of employed cellular sites are shown in Fig. 2, for different EIRP levels at the TV broadcasting site, coverage targets, and indoor/outdoor reception. Note how the slopes are higher for outdoor than for indoor reception. The collection of values corresponding to the number of cellular sites per Mb/s, n/R, are presented in Table III. Values in Table III show that, in general, we experience an increasing number of sites per Mb/s when we upgrade the network to higher capacity (by selecting a higher order modulation) or coverage. However, there are few exceptions. For QPSK 2/3 and 16-QAM 1/2 the values are almost the same, making the latter a better choice due to higher capacity. Almost a similar situation can be notice between 16-QAM 2/3 and 64-QAM 1/2. One conclusion may be that 1/2 coding rate is preferable to 2/3.

20 18 16 14 Capacity [Mb/s]
16QAM 2/3 64QAM 2/3

from one MCR to another is not guaranteed that the initial set of sites is kept in operation. VI. C ONCLUSIONS We have investigated the scalability of DVB-H deployment by re-using existing wireless infrastructure. Our numerical results show that when using only the broadcasting tower the cost of increasing capacity or coverage comes at very large costs. Cellular sites can be a good complement to traditional broadcasting towers, especially for indoor reception planning. However, a cost-efcient DVB-H deployment will not be achieved if high coverage and capacity levels are targeted. Efforts should be focused mostly on implementing adaptive applications based on data caching in the user terminal, and consider the possibility of using the cellular network to provide point-to-point retransmissions. This may replace the need for high coverage and high capacity values, at least for non realtime content. ACKNOWLEDGMENT The authors would like to thank to Generalitat de Valencia (Spain) for the nancial support given to David during this study.

64QAM 1/2

12 10 8 6 4 2 0 0 20 40
QPSK 2/3 QPSK 1/2 16QAM 1/2

EIRPBC 50dBW, outdoor 70% EIRPBC 50dBW, outdoor 90% Only Cellular, outdoor 70% Only Cellular, outdoor 90% EIRPBC 50dBW, indoor 70% EIRPBC 50dBW, indoor 90% Only Cellular, indoor 70% Only Cellular, indoor 90% 100

60 80 Number of Cellular Sites

Fig. 2. DVB-H service capacity vs. number of employed cellular sites for different area coverage targets and EIRP from broadcasting tower. Service area radius is 25 km. TABLE III Pb (dBW) Cov. (%) 70 90 0 95 70 90 40 95 70 90 50 95 i/o i o i o i o i o i o i o i o i o i o QPSK 1/2 2/3 3.0 1.4 5.0 2.4 6.4 3.0 1.2 0 2.8 0 4.0 1.2 0 0 0 0 1.6 0 3.3 1.7 5.4 2.6 6.9 3.0 1.5 0 3.6 0.9 4.8 1.5 0 0 0 0 2.6 0 16-QAM 1/2 2/3 3.0 1.4 4.9 2.2 6.1 2.8 1.7 0 3.6 1.1 4.8 1.6 0 0 1.8 0 3.0 0 3.6 1.7 6.0 2.6 7.7 3.2 2.5 0.7 5.0 1.6 6.4 2.3 0.7 0 2.9 0 4.5 0.8 64-QAM 1/2 2/3 3.8 1.7 6.4 2.7 3.3 2.7 0.7 5.3 1.9 6.8 2.4 0.9 0 3.4 0 5.1 0.9 4.4 2.0 3.1 4.0 3.4 1.1 2.4 3.1 1.9 0 4.8 1.0 1.8

R EFERENCES
[1] ETSI, EN 302 304 v1.1.1, Digital Video Broadcasting (DVB); Transmission System for Handheld Terminals (DVB-H), October, 2004. [2] M. Kornfeld and U. Reimers, DVB-H, The Emerging Standard for Mobile Data Communication, EBU Technical Review, January, 2005. [3] G. Faria, DVB-H to Deliver Digital TV to Handheld Terminals, IBC 2004 Conference, Amsterdam, Netherlands, 2004. [4] D. G mez-Barquero and A. Bria, Feasibility of DVB-H Deployment on o Existing Wireless Infrastructure, International Workshop on Convergent Technologies (IWCT), Oulu, Finland, June 2005. [5] DigiTAG Handbook, Television on a handheld receiver - broadcasting with DVB-H, March, 2005. http://www.digitag.org/ DVBHandbook.pdf [6] B. Mazieres, DVB-H Experimental Field Trials in Metz, First Conclusions, Multiradio Multimedia Communications Workshop, Berlin, Germany 2005. [7] ETSI, TR 102 401 v1.1.1, Digital Video Broadcasting (DVB); Transmission to Handheld Terminals (DVB-H); Validation Task Force Report, May, 2005. [8] ETSI, TR 102 377 v1.1.1, Digital Video Broadcasting (DVB); DVB-H Implementation Guidelines, Feb., 2005. [9] ICNIRP Guidelines, Guidelines for Limiting Exposure to Time-Varying Electric, Magnetic, and Electromagnetic Fields (up to 300 GHz), Health Physics Vol. 74, No. 4, pp. 494-522, 1998. [10] A. Ligeti, Single Frequency Network Planning, Ph.D. dissertation, Royal Institute of Technology, Stockholm, Sweden 1999. a [11] K. Johansson, A. Furusk r, P. Karlsson, and J. Zander, Relation Between Cost Structure and Base Station Characteristics in Cellular Systems, IEEE 15th PIMRC Conference, Barcelona, Spain, 2004. [12] R. Schramm, DVB-T C/N Values for Portable Single and Diversity Reception, EBU Technical Review, April, 2004. [13] ITU, Digital Terrestrial Television Broadcasting Handbook, Digital Terrestrial Television Broadcasting in the VHF/UHF Bands, Edition 2002 - v1.01. [14] Recommendation ITU-R P.1546: Method for Point-to-Area Predictions for Terrestrial Services in the Frequency Range 30 MHz to 3000 MHz, 2003.

It should be pointed out that the minimum variation of the n/R corresponds to the case when only cellular sites are employed (0 dBW at the broadcasting tower). However, the absolute cost of this implementation might be higher than for hybrid situations. Note that n/R remains almost constant from QPSK 1/2 to 16-QAM 1/2 for indoor reception, whereas for outdoor reception the cellular network is almost scalable up to 64-QAM 1/2. Results in Table III are optimistic in the sense that the number of sites necessary for certain capacity and coverage is the minimum possible under our assumptions. When switching

Potrebbero piacerti anche