Umts900 Full Report

Market Study for UMTS900
A report to GSMA
February 2007
Project Number CLW28 Version V1.1
Ovum Consulting, Cardinal Tower, 12 Farringdon Road, London EC1M 3HS
Telephone +44 (0) 20 7551 9000 Facsimile +44 (0) 20 7551 9090/1 www.ovumconsulting.com
Contents
1 2 3 4 5 6 7 Executive summary ..................................................................1 Introduction.............................................................................3 Report findings ........................................................................4 Deployment analysis ..............................................................20 Regulatory analysis ................................................................48 Handset and chipset analysis..................................................73 Economic analysis ..................................................................77
Annex A Model description ...................................................................93 Annex B Operator benchmarks ........................................................... 105 Annex C GSM-900 and E-GSM bands.................................................... 112
UMTS 900 Market Study Final Report V1.1
MARKET STUDY UMTS 900

A report to GSMA
1 Executive summary
GSMA is investigating the issues that can impact the use of the 900MHz frequency band for UMTS services. This report was commissioned to assess the benefits and limitations of UMTS900 deployment. The report indicates that UMTS900 provides between 44% (in urban areas) and 119% (rural areas) increased coverage per Node-B compared with UMTS2100. This is primarily due to the propagation characteristics of the lower frequency band and leads directly to lower capex and increased mobility benefits, providing a new option, with greater service capability, for operators who may wish to replace their GSM networks. To maximise the benefits of UMTS900 through lower equipment and device costs, and greater certainty of outcome there needs to be international harmonisation of the 900MHz band to allow UMTS900 use, which is one of the most used bands in the world. The analysis for this report considered potential capex reductions (UMTS900 compared with UMTS2100) for a typical mix of geography types in 4 world regions. The case for Sub Saharan Africa shows that cumulative capex reductions of 41% may be possible over a five year period if UMTS900 is used rather than UMTS2100. Other regions are Western Europe - 40%, Middle East - 36% and Asia Pacific 32%. Slightly lower reductions are indicated if a more typical mix of UMTS900 and UMTS2100 is deployed. If revenues over a five year period are also considered, then a simple analysis indicates NPV improvements ranging from 40% - 55% in the Middle East and SubSaharan Africa respectively (with investment in UMTS900 to provide the same coverage as would be provided with UMTS2100). If the cost savings are reinvested to enable the operator to reach a larger customer base by extending geographic coverage, then the NPV improvements of 39% - 105% are indicated in Western Europe and Asia Pac. Considering specific countries to illustrate the benefits indicated by our high-level modelling: Finland could provide UMTS900 coverage for $700M less than it might cost to deploy UMTS2100; in Saudi Arabia it would be a $2.1bn reduction; South Africa $500M and Sri Lanka $24M. There are several factors important for the successful introduction of UMTS900. Co-ordinated policy to refarm 900MHz spectrum, which is currently primarily used for GSM, is required to allow UMTS. Some NRAs have already initiated discussions and the EC has indicated that the GSM900MHz and GSM1800MHz bands will be approved for UMTS900/1800 use in September/October 2007. The co-ordination will need to address the interest of GSM users, interference between bands and the potential for cross border interference. NRA spectrum policy on pricing and roaming has also to be considered.

A report to GSMA
The success of UMTS900 introduction will also require economies of scale to help drive down handset and device prices. Chipset and handset vendors will not push product into immature markets, the markets will have to be stimulated by market pull. This in turn will be greater if it is possible to introduce services in multiple markets in parallel. Our modelling indicates that UMTS900 can generate cost reductions of up to 40% in capex and 30% in overall costs when compared to a baseline case scenario of deployment using UMTS at 2100MHz. These lower costs are primarily due to the radio propagation characteristics in the lower band which provide greater reach of UMTS900 and improved in-building coverage. At a qualitative level, UMTS900 can bring economic benefits, especially to less developed countries. Capital investments would be stimulated and improved communications can stimulate productivity and job creation and business working practices.

A report to GSMA
2 Introduction
There is a growing interest in deploying UMTS in the 900MHz frequency band in order to reduce the cost of coverage for mobile communications services, especially into rural areas. This may allow an operator to economically roll out higher data rate services to a larger percentage of the rural population. In addition there is the need to increase coverage indoors which is easier to achieve in the lower frequency bands. This report was commissioned by GSMA and QUALCOMM to help their investigation into the issues that may impact the deployment of UMTS900 and to better understand how any transition may be achieved over time. The objectives were to investigate the key factors around the UMTS900 business case, assess the benefits and limitations and evaluate the deployment scenarios that could meet operator needs for service deployment. The report provides: A summary of the findings of the study in Chapter 3 In Chapter 4, an assessment of the reduction in costs when deploying a 3G/UMTS network when an operator has access to 900MHz as well as 2.1 GHz spectrum, in 4 different regional markets. An assessment of how any reduction in deployment costs may help business case NPV, and the impact of extending 3G deployments beyond limits possible at 2.1 GHz (due to reduced deployment costs) is also provided An analysis of the regulatory situation in each region in Chapter 5, together with analysis of the issues related to refarming of 900MHz spectrum Chapter 6, which looks at the status of handset and chipset supply and the impact of their availability on the development of UMTS900 services A qualitative economic analysis of the benefits of introducing UMTS900, in Chapter 7 The annexes to the report provide: A detailed overview of the network dimensioning and cost modelling methodology Case studies of Telstra and Cingular two operators that have deployed UMTS 850.

A report to GSMA
3 Report findings
This section of the report provides a summary of the main findings of the study.
3.1 Demand and network cost analysis

UTMS 900 provides increased coverage The propagation characteristics of the 900MHz spectrum allow UMTS900 base stations to cover a greater geographic area than UMTS2100. The relative coverage area increase is shown in Figure 3.1. The results are based on the assumptions used in the modelling and, given the wide variations between national and international regions, the results should be considered to be indicative only.
Figure 3.1 Percentage increase in coverage area

Frequency Percentage increase in coverage area per Node-B (km 2) Dense Urban 900MHz vs. 2100MHz 87% Urban 44% Suburban 60% Rural 119%
Source: Ovum model The reason for the differences in coverage area per Node-B in dense urban and rural areas is due primarily to lower signal attenuation, which improves reach and in-building coverage. In more densely populated regions a combination of UMTS900 and UMTS2100 may be needed to provide capacity as well as coverage. The difference in coverage areas for each Node-B will directly affect the number required to serve a given geographic area and consequently the Capex costs. UMTS900 improves mobility management The improved Node-B coverage of UMTS900 technology reduces the number of cells in a region. This should decrease the potential number of gaps between cells, which will help overcome handover problems and thus improve the customer experience. Network capex efficiencies and enhanced capabilities Whilst each case is operator and network specific, UMTS900 offers a new option for operators facing a GSM network equipment replacement cycle. As well as UMTS900 requiring lower up-front investments than UMTS2100 it also enhances network efficiencies and the networks capability to handle both voice and data traffic. Additionally, the increased coverage capabilities of UMTS900 allow high speed data

A report to GSMA
and more efficient voice traffic handling, with reduced Capex, when compared to UMTS2100. Harmonisation of the 900MHz band The 900MHz band is one of the most used bands in the world. GSM operators all over Europe, Africa and Asia use this band extensively, which makes it one of the most harmonised bands in the world. All operators using the 900MHz have started with GSM services and most of them have already acquired 3G licenses at 2.1 GHz. This business evolution makes UMTS900 a most attractive option for operators and a likely follow-up technology in the 900MHz band. In order for operators to be able to utilise UMTS technology in the 900MHz band two prerequisites are required regarding spectrum availability: the NRA must allow the deployment of UMTS in the 900MHz band re-farming of 900MHz spectrum band is needed in order for the operators to be able to deploy 5MHz (WCDMA) carriers. For those operators having largely deployed GSM in the 900 MHz band 2 x 10MHz is an indicative amount of spectrum required to roll out GSM and UMTS in parallel. In some countries where 900 MHz is still available, operators may start UMTS900 with 2x5 MHz of clean spectrum.
If these prerequisites are met and if both GSM and UMTS technologies can be deployed in the 900MHz band then this harmonisation of spectrum will improve the flexibility of mobile operators to use spectrum efficiently (through network and frequency planning process) to meet coverage and capacity requirements. Demand analysis and network costs The demand analysis model was developed to provide a high level view of service demand based on scenarios for each of the four major world regions considered in the study (W. Europe, Asia Pacific, Sub-Saharan Africa, Middle East). For each world region Capex and Opex costs were modelled to take account of demand, subscriber base and network coverage. The radio access network was dimensioned to provide coverage and capacity, the core network was determined and network capital costs derived for each type of deployment area (rural, suburban, urban and dense urban). Opex was derived using industry norms for capex/ opex ratios for network related opex and non network-related opex. Figures 3.2 and 3.3 show the 5 year cumulative network capex costs for UMTS900 only, and combined UMTS900 and UMTS2100 deployments respectively when compared to a network with only UMTS2100. In the UMTS900 case, the Sub Saharan Africa case shows the highest cumulative capex reductions of 41% followed by Western Europe with 40%, Middle East with 36% and Asia Pacific with 32%. In the case of UMTS900 and UMTS2100, the Sub Saharan Africa case shows the highest cumulative capex reductions of 37% followed by Middle East with 31%, Western Europe with 27% and Asia Pacific with 16%.

A report to GSMA
These results reflect a modelled scenario and there are many factors which will influence the actual capex reduction benefits in real-world networks. Although the case for UMTS900 seems to be the better of the two sets of results, it must be recognised that in many cases a combined UMTS900 and UMTS2100 network will be preferred, in order to meet market conditions for coverage and capacity. In the regions examined it is to be expected that coverage and not capacity will be the main driver for network deployment. This is mainly due to the expectation that high-speed data service coverage and usage will be limited in most of the emerging markets due to cost of deployment. It is for this reason that the UMTS900 only case shows higher capex savings compared to both the UMTS2100 only case and the combined UMTS900 and UMTS2100 case. If analysis is performed on a country level with more detailed data on end-users usage profiles and traffic split then it is expected, especially in W. European developed markets, that an overlay solution of UMTS900 for coverage issues combined with a deployment of UMTS2100 networks for capacity issues in hot-spots will be the optimal one. Detailed analysis is not covered in this report and results shown in subsequent graphs show the magnitude of capex savings expected when operators use UMTS900 as a stand-alone or as a combined technological option. Figure 3.2 Cumulative Capex costs for UMTS900 only as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex: UMTS900 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Ba seline scenario: 2.1 GHz only
Asia-Pac
W. Europe
Sub-Saharan Africa
Low Demand
Medium Demand
High Demand
Source: Ovum

A report to GSMA
Figure 3.3 Cumulative Capex costs for UMTS900 and UMTS2100 as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex: UMTS900 and UMTS2100
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Baseline scenario: 2.1 GHz only
Asia-Pac
W. Europe
Sub-Saharan Africa
Low Demand
Medium Demand
High Demand
Source: Ovum Figure 3.4 presents a similar assessment, but broken down by geographic country area. In the case of UMTS900 only, dense urban environment case shows the highest cumulative capex reduction in the range of 37%-46%, followed by rural environment (range of 33%-46%), suburban environment (range of 26%-34%) and finally urban environment (range of 20%-36%) for all of the regions examined. The Sub Saharan Africa case shows for all types of environments (except urban) the highest cumulative capex reduction.

A report to GSMA
Figure 3.4 Cumulative Capex costs per type of environment for UMTS900 only as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex for different environments - UMTS900 only 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% W. Europe AsiaPac Dense Urban Urban Middle East Suburban Rural Africa
Source: Ovum Figures 3.5 and 3.6 show the 5 year cumulative network capex and opex costs for UMTS900 only, and combined UMTS900 and UMTS2100 deployments respectively when compared to a network with only UMTS2100. Figure 3.5 Total Capex and Opex costs for UMTS900 only as a percentage of Total Capex and Opex costs for UMTS2100 only
Total Costs: UMTS900 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac Low Demand
W. Europe
Sub-Saharan Africa
Medium Demand
High Demand
Source: Ovum

A report to GSMA
Figure 3.6 Total Capex and Opex costs for UMTS900 and UMTS2100 as a percentage of Total Capex and Opex costs for UMTS2100 only
Total Costs: UMTS900 and UMTS2100
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac Low Demand
W. Europe
Sub-Saharan Africa
Medium Demand
High Demand
Source: Ovum In the case of UMTS900 only, the Sub Saharan Africa case shows the highest total Capex and Opex costs reductions, when compared to UMTS2100 only, of 38%. The other regional total cost reductions are the Middle East with 32%, Asia Pacific with 17%, and Western Europe with 10%. In the case of combined UMTS900 and UMTS2100,deployments, the Sub Saharan Africa case shows the highest total Capex and Opex costs reductions of 33% followed by Middle East with 27%, Asia Pacific with 8% and Western Europe with 7%. Initial NPV Analysis By means of a straightforward cash flow calculation, a simple NPV analysis was prepared for all of the cases and regions examined. Revenues were derived by calculating total service usage and associated prices for all of the services examined (voice, SMS, data) along with their evolution over time for all four regions. Figure 3.7 presents NPV percentage improvements for UMTS900 deployment compared with the case of UMTS2100 only and for combined UMTS900 and UMTS2100 compared with UMTS2100 only, for all four regions. The main conclusions drawn are that UMTS900 only presents greater opportunities for NPV improvements over a 5-years period than combined UMTS900 and UMTS2100 deployments. The Sub-Saharan Africa and Middle East regions present the greatest opportunities for NPV improvements due to the reduced capex and opex costs incurred.

A report to GSMA
10
Figure 3.7 NPV percentage improvements (medium demand scenario) over UMTS2100 only case for both UMTS900 only and UMTS900 and UMTS2100 cases
NPV improvements over UMTS2100 only (*)
120% 100% 80% 60% 40% 20% 0% W. Europe AsiaPac UMTS900 only
(*) refers to Medium Demand scenario
ME
Africa
UMTS900 and UMTS2100
Source: Ovum model Revised NPV analysis The assumptions for inputs leading into the initial NPV analysis (such as WCDMA penetration, geographic coverage) have remained the same in order to calculate, on an equal basis, potential capex reductions through the introduction of UMTS900 technology. However, capex reductions can be re-assessed to reflect a more realistic situation, in which the operator is able to take advantage of the cost reductions to fund further capex, to attract more customers and hence increase the number of connections or increase geographic coverage. By adjusting the model to take account of changes to the number of WCDMA connections and WCDMA geographic coverage the NPV calculations were revised. The results are shown in Figure 3.8. The main reasons for greater NPV improvements in W Europe and Asia Pacific are that the absolute number of additional subscribers is higher in these two regions (due to higher potential addressable market) which contributes to higher expected revenues. Additionally, the capex spending per subscriber required to increase coverage/capacity requirements is lower due to landscape characteristics and existing WCDMA penetration it is easier to provide additional capacity in existing areas than to greatly increase coverage in rural areas with no coverage. There is potential for virtuous circles to develop: the lower the cost of deployment of mobile infrastructure, the greater the coverage - leading to more users and greater national benefit more users will drive a greater the volumes of handsets and devices, leading to a wider range of products becoming available and, when economies of scale cut in, leading to lower device costs.

A report to GSMA
11
Figure 3.8 Updated NPV percentage improvements over UMTS2100 only

NPV improvements over UMTS2100 only
120%
100% 80%
60% 40%
20%
0% W. Europe Asia Pacific UMTS900 only Middle East UMTS900 & 2100 Sub-Saharan Africa
Source: Ovum model
3.2 Case studies examples

In this section, a brief analysis of the UMTS900 benefits on a country level is presented. One countries in each of the four regions has been examined. For each of the countries examined the following conclusions can be derived: a) When the same population and geographic requirements are met in both scenarios - UMTS2100 only and UMTS900 only: As expected there are direct financial savings for operators in the UMTS900 case due to reduced capex spending, which is in the range of 25% to 40%. However, apart from the direct capex reduction benefits for the operators there are some also indirect benefits due to the reduced number of cell sites needed to be deployed; time and effort savings and reduced environmental and health impact. b) When the same cumulative capex spending (over 5-years) is assumed for both scenarios - UMTS2100 only and UMTS900 only: In this case, there is an increase in population and geographical coverage when deploying UMTS900 only as opposed to the UMTS2100 only. The increase varying with each country. However, as noted in Section 7 of the report, it in the UMTS900 only case, countries can enjoy significant socioeconomic benefits as the 3G addressable market enlarges and more people have access to wireless broadband services. This is also discussed in Section 3.5 of the report. In cases presented below only a high level analysis has been performed. To derive more accurate results a more detailed approach which assesses the countrys market and geospatial characteristics would be needed. However, the analysis sets

A report to GSMA
12
the overall picture and provides some useful insights of the benefits that are expected with the deployment of UMTS900 networks.
3.2.1 Sri Lanka

Sri Lanka is a developing market with the first WCDMA subscribers expected by the second half of 2007. The same assumptions about deployment analysis are made as previously discussed and comparative outcomes between use of UMTS900 and UMTS2100 technology are derived. There is the opportunity for countries like Sri Lanka, which do not have existing UMTS2100, to start with UMTS900 if the regulatory path for its deployment is clear, and a there is reasonable handset and device availability. When deploying UMTS2100 technology (assuming 8% population coverage and 15% geographic coverage by 2011) a total cumulative capex (over 5 years) of around $126m is required to deploy approximately 1,100 sites to meet coverage and capacity requirements. In the case of UMTS900 technology, and with the same requirements, a cumulative capex spending of $94m is required to deploy approximately 750 sites. In the second case there is a 25% reduction in capex spending required, mainly driven by the reduced number of cell sites and other core and transport cost savings.. In the case where we assume that the operator will spend the same amount as in the case of UMTS2100 (i.e. $126m), over 5 years, but by deploying UMTS900 networks instead, then both population and geographic coverage can be significantly increased. By following a linear cost-oriented approach (capex/subscriber driver), it is estimated that by 2011 population coverage could increase by 50% from 8% to 12% and geographic coverage increases to 20% as compared to 15% previously.
3.2.2 Finland
In Finland, 3G penetration at the start of 2007 is almost 15% and is forecast by Ovum to rise to approximately 78% by 2011. The same assumptions about deployment analysis are made as previously discussed and comparative outcomes between use of UMTS900 and UMTS2100 technology are derived. When deploying UMTS2100 technology (assuming 78% population coverage and 75% geographic coverage by 2011) a total cumulative capex (over 5 years) of around $1.86bn is required to deploy approximately 11,000 sites to meet coverage and capacity requirements. In the case of UMTS900 technology, and with the same requirements, a cumulative capex spending of $1.1bn is required to deploy approximately 7,000 sites.

A report to GSMA
13
In the second case there is a 40% reduction in capex spending required, mainly driven by the reduced number of cell sites and other core and transport cost savings. In the case where we assume that the operator will spend the same amount as in the case of UMTS2100 (i.e. $1.8bn), over 5 years, but by deploying UMTS900 networks instead, then both population and geographic coverage can be significantly increased. By following a linear cost-oriented approach (capex/subscriber driver), it is estimated that by 2011 population coverage could increase by 30% to reach almost 100% and geographic coverage increases to 95% as compared to 75% previously.
3.2.3 Saudi Arabia

In Saudi Arabia, the first WCDMA (3G) networks are expected to be launched in 2007. Ovum forecasts that population penetration will rise to reach 45% by 2011. The same assumptions about deployment analysis are made as previously discussed and comparative outcomes between use of UMTS900 and UMTS2100 technology are derived. When deploying UMTS2100 technology (assuming 45% population coverage and 35% geographic coverage by 2011) a total cumulative capex (over 5 years) of around $6.3bn is required to deploy approximately 30,000 sites to meet coverage and capacity requirements. In the case of UMTS900 technology, and with the same requirements, a cumulative capex spending of $4.2bn is required to deploy approximately 21,000 sites. In the second case there is a 37% reduction in capex spending required, mainly driven by the reduced number of cell sites and other core and transport cost savings. In the case where we assume that the operator will spend the same amount as in the case of UMTS2100 (i.e. $6.3bn), over 5 years, but by deploying UMTS900 networks instead, then both population and geographic coverage can be significantly increased. By following a linear cost-oriented approach (capex/subscriber driver), it is estimated that by 2011 population coverage could increase by 50% to reach almost 65% and geographic coverage increases to 50% as compared to 35% previously.
3.2.4 South Africa

In South Africa, 3G penetration is gradually increasing and is forecast by Ovum to rise from around 2% to approximately 25% by 2011. The same assumptions about deployment analysis are made as previously discussed and comparative outcomes between use of UMTS900 and UMTS2100 technology are derived. When deploying UMTS2100 technology (assuming 25% population coverage and 20% geographic coverage by 2011) a total cumulative capex (over 5 years) of

A report to GSMA
14
around $1.4bn is required to deploy approximately 19,000 sites to meet coverage and capacity requirements. In the case of UMTS900 technology, and with the same requirements, a cumulative capex spending of $0.9bn is required to deploy approximately 12,000 sites. In the second case there is a 35% reduction in capex spending required, mainly driven by the reduced number of cell sites and other core and transport cost savings. In the case where we assume that the operator will spend the same amount as in the case of UMTS2100 (i.e. $1.4bn), over 5 years, but by deploying UMTS900 networks instead, then both population and geographic coverage can be significantly increased. By following a linear cost-oriented approach (capex/subscriber driver), it is estimated that by 2011 population coverage could increase by 40% to reach almost 35% and geographic coverage increases to 30% as compared to 20% previously.
3.3 Key regulatory issues

Spectrum availability Some NRAs, especially in developed countries, have already initiated discussions about new 2G spectrum management policies, the re-farming of GSM bands and the processes to be followed for the migration of GSM services to UMTS services. Spectrum allocation is made more difficult by the requirement to allocate UMTS spectrum in 5MHz blocks. It is essential to ensure that both GSM and UMTS services can continue to operate in a parallel and well-balanced manner as GSM networks will continue to be the core business for the majority of mobile operators for many years. Sharing of 900MHz spectrum NRAs will have to access the impact of allowing the GSM and UMTS services to coexist. Interference effects among the services operating within the same band need to be minimised, and co-ordination with adjacent countries will also be necessary. In Europe, this is being addressed in CEPT/ECC reports 82 and 96. GSM licence renewal In the majority of the cases second generation licences are technology specific allowing only GSM services to be used in the designated bands. Although the refarming of the GSM bands allows the co-existence of second and third generation mobile services, such arrangements may raise concerns and competition issues among the mobile industry.

A report to GSMA
15
Impact of UMTS900 on existing GSM and UMTS operations The full use of 900MHz spectrum for the provision of 3G services in the future may be a barrier for new entrants. The licence holders of 900MHz band may be able to deploy UMTS900 in current GSM spectrum in some countries without applying for a separate 3G licence. Furthermore, holders of current 3G licences may believe that their spectrum will be undervalued as a result of additional spectrum becoming available for 3G. It is therefore important that before altering the current market conditions or the conditions upon which the licences were granted, NRAs have to consider future implications. Spectrum pricing or national roaming NRAs may have to develop new pricing formulas for spectrum fees. In cases where 2G operators who are not 3G licence holders are allowed to use their spectrum for 3G services, NRAs need to consider whether these operators should pay a levy as an additional payment for the extra usage of the spectrum. Mobile operators need to know what is available and its cost in order to define their strategies and invest accordingly. Vendors and handsets manufacturers need also a clear view when deciding to invest in new equipment. Alternatively, NRAs may provide a regulatory framework for national roaming agreements to compensate the higher network costs of the operators without 900MHz spectrum. The impact of EU directives and national developments The regulatory framework of the EU directives is a significant factor which could speed-up or delay the deployment of UMTS900 networks. Most of the NRAs are reluctant to move towards the UMTS900 adaptation due to these directives. However, at the European level there are processes in progress to develop the regulatory framework for the deployment of 3G networks within the 900MHz and 1800MHz bands. The regulatory situation in a range of countries is discussed in Section 5 of the report, some countries specify the technology that can be used, others are technology neutral. In summary: Western Europe/EU: Second generation licences are technology specific, and there is no free spectrum in the 900MHz band. There is strong interest in UMTS900 and regulators in France, UK, Germany and Finland take a technology neutral approach with regard to GSM/3G. It is expected that demand from operators will accelerate regulatory developments Asia/Pacific: in the leading countries (in terms of mobile development), NRAs have either adopted a technology neutral policy or have proposed spectrum liberalisation policies. Whilst UMTS networks have a strong uptake, alternative solutions in the 900MHz are available (e.g. CDMA-2000) Middle East: In most countries, mobile licences are technology specific, but key regulators (e.g. Saudi Arabia, UAE) in the region are in favour of mobile

A report to GSMA
16
technology neutrality. UMTS network deployments are still immature and there is no current demand for UMTS900 licences and UMTS2100 is preferred Sub-Saharan Africa; UMTS networks deployments are immature with very slow uptake. 2G re-farming policies and UMTS900 are not considered as part of the short-term spectrum strategy. NRAs wish to safeguard GSM operations, due the national revenues which are derived
3.4 Handset and chipset analysis

All major chipset and handset manufacturers agreed that UMTS900 will happen. They are all either currently developing or planning to develop UMTS900 chipsets and handsets. Product evolution will start with UMTS2100/900 (dual band) devices before moving towards quad band. The vendors were unable to predict expected volumes as that depends on demand from operators. It is the success in rolling out UMTS900 services that will pull demand for devices to support UMTS900. Vendors stressed that operators will need to create the pull to ensure that the UMTS900 market is seeded with handsets / devices. This seeding is needed to ensure device availability before the network is enabled, without which customer take-up is bound to be limited. Manufacturers expect that between a few hundred thousands and up to a few million handsets will be produced by 2009. With these low quantities, volumerelated cost reductions are not expected. However, advances in chipset design, e.g. higher levels of integration, will be the main driver for cost reduction. Furthermore, chipset manufacturers do not in general expect cost reductions to be necessarily reflected in prices.
3.5 Economic analysis

The modelling for this study demonstrates that UMTS900 can effectively generate cost reductions of up to 40% in CAPEX and 30% in overall costs when compared to a baseline case scenario of deployment using UMTS at 2100MHz. These lower costs are primarily due to the radio propagation characteristics in the lower band which provide greater reach of UMTS900 and improved in-building coverage. In a study of this nature which assesses costs and demand at a macro level it is not possible to model the affect between capital costs, the price of services and the demand for services, however, some general economic principles were assessed. Impact of UMTS900 on prices It is considered unlikely that operators will pass on capex and opex savings from deploying UMTS900 to customers immediately. The major reasons for operators not to reduce their prices being that they are incurring capital expenditure for new infrastructure, which is funded by debt, and continued shareholder pressure to realise higher returns. Also, in a market with high mobile penetration and where

A report to GSMA
17
operators deploy more than one technology (e.g. GSM and UMTS in multiple bands), operators would not wish to undermine their established pricing structures and profitability. In the medium term, as investments in the GSM network have already been amortised and the declining contributions of the GSM platform represent a residual net present value which is low when compared with the potential NPV with an UMTS only network, we might expect a more accelerated replacement of GSM by UMTS900. When this happens, prices will tend to change reflecting the lower long run costs of UMTS. Impact of UMTS900 on demand for services UMTS900 will impact demand for services in three ways: Elasticity of demand to changes in price Expansion of the addressable market The effect of tariffs on service usage
Analysis of the UK market shows that a 10% change in average price per minute for mobile voice services results in a 12% increase in usage. This cannot be assumed to be a direct result of price elasticity, as one of the natural drivers of the demand for mobile services is fixedmobile substitution. This is not strongly driven by price of the mobile service and therefore we would not expect lower costs of UMTS900 deployment to influence this major driver. Analysis of data from Wireless Intelligence shows that there is greater elasticity of demand in developing countries. The main impact of UMTS900 on the demand for services is expected to come from developing countries and regions, which may create an incentive for operators to deploy in these regions. One valuable characteristic of UMTS900 is that it makes it possible to extend mobile coverage in developing countries as it becomes more economically viable to deploy UMTS services in suburban or rural areas where this was not possible previously. A detailed national study would be required to assess the value of such additional network roll-out to a nations economy. If lower costs are passed on in the form of lower prices, then this will expand the usage of services. If flat rate tariffs are introduced for data services, then this can be a major stimulator of demand. Impact on country economics There are many other ways in which UMTS900 can bring economic benefits. These are illustrated by examining how, according to economic growth theory, it might contribute to economic growth: Through the stimulation of capital investments By stimulating job creation, by virtue of improved communications

A report to GSMA
18
By improving productivity By stimulating change in business working practices.
Regional differences The major benefits from the introduction of UMTS900 are expected to vary by region. In Western Europe it is expected that economic benefits from the introduction of UMTS900 technology be from the long term benefits that the diffusion of mobile broadband wireless access and mobile internet will have in reshaping business practices. In the Middle East there is still room for increase in the penetration of services and capital investment is one key economic benefit. However, as oil-rich countries look into diversifying to tourism and financing, UMTS900 can be also instrumental in improving labour productivity. In Asia Pacific there is a mixed picture. In countries with high population density, the economic benefit that deployment of UMTS900 can yield is to speed-up the substitution of 2G to 3G thus stimulating the reshape of business models. In countries such as Australia and New Zealand, UMTS900 can also play an important role in bringing 3G to more remote areas improving economic activity of indigenous population and reducing transaction costs for farmers. In sub Saharan countries the case is clearly to bring prices down and to make mobile services accessible to a larger percentage of the population while also expanding the supply by extending the coverage to rural areas. Capital investment will influence growth in a higher degree that it does in other areas but UMTS900 has a much bigger role in helping people find employment or as an enabler of informal economic activities. Additional economic benefits The benefits of UMTS900 to the economy are not restricted to the economic growth that increased voice-centred mobile service penetration and usage prompts. Because of its native broadband data capabilities the introduction of UMTS900 has a much broader impact in the ICT sector as a whole. One direct consequence of the introduction of UMTS900 will be to speed up the penetration of 3G services in the different regions. UMTS900 will also impact Internet access. It will allow more individual users to access content in the Internet using web-browse enabled handsets or data cards. The positive effects will be felt particularly in rural areas where, in many occasions in developing countries, Internet access is not available because there are no fixed lines or broadband service available. It will also provide one more channel for enterprises to keep its workforce connected to the Internet even when on the move. In developing countries, the broader availability of Internet access for

A report to GSMA
19
enterprises may also stimulate that more companies decide to launch their own websites. All in all, UMTS900 can be instrumental in reducing the digital divide between developed and developing countries and urban and rural environments.

A report to GSMA
20
4 Deployment analysis
This section of the report analyses the following areas: Implications of UMTS900 technology: a summary of the impact that UMTS900 technology might have on a mobile operators current operations and of the opportunities that the technology provides Scenario analysis: describes the different deployment scenarios from which results and conclusions have been developed. The different scenarios allow the effect of different combinations of input parameters to be assessed Demand analysis: analysis of the affect of varying input parameters, such as penetration, usage profiles, ARPU levels and service pricing in each of the regions examined. Demand analysis has a direct impact both on costs (number of subscribers, bandwidth demand, etc.) and revenues (number of subscribers, services pricing, etc.) Capex and Opex analysis: provides a high level description of the Capex and Opex calculations. Full details of the model methodology are provided in Annex A Comparative analysis of the regions: compares the of key indicators, e.g. relative differences of Capex and Opex costs from the baseline scenario for each of the regions studied.
4.1 Implications of UMTS900 technology

There are many areas in a mobile operators network, both technical and operational, that could be affected by the introduction of UMTS900 technology. Five main areas are assessed: Network coverage capability Mobility management Network capabilities/ efficiencies and Capex Positioning against other access technologies Harmonisation of spectrum in the 900MHz band
4.1.1 Network coverage

The main benefit of UMTS deployment in the 900MHz band is lower signal attenuation when compared to the 2.1 GHz band. This directly impacts cell coverage and will also improve handover quality as explained below under mobility management. Figure 4.1 presents the cell radius for both UMTS900 and UMTS2100 technologies for four different types of environments: dense urban, urban, suburban, rural. The

A report to GSMA
21
cell radius for different environments is derived from standard 3GPP dimensioning models1 after first calculating the maximum allowed path loss in each environments for each of the two technologies (UMTS900 and UMTS2100) under consideration. Figure 4.1 Cell radius for different types of environments Frequency
Dense Urban
Cell radius (km)

Urban 1.62 1.35 Suburban 3.47 2.74 Rural 12.50 8.44
900MHz 2100MHz Source: Ovum model
1.03 0.75
By applying the hexagonal calculation1 the total area covered by each Node-B the coverage area per site is calculated, as shown in Figure 4.2. Figure 4.2 Coverage area for different types of environments and for both UMTS900 and UMTS2100 technologies Frequency
Dense Urban
Coverage area (km2)

Urban 5.15 3.57 Suburban 23.47 14.65 Rural 304.87 139.06
900MHz 2100MHz Source: Ovum model
2.06 1.10
Figure 4.3 Percentage increase in coverage area Frequency Percentage increase in coverage area per Node-B (km2)
Dense Urban Urban 44% Suburban 60% Rural 119%
900MHz vs. 2100MHz Source: Ovum model
87%
As it can be seen, there are significant variations in the maximum coverage area that can be achieved per Node-B between the two technologies, presented in percentage terms in Figure 4.3. This difference will directly affect the number of Node-Bs required to serve a given geographic area and consequently the Capex costs. It must be noted that the greatest differences in coverage area per Node-B are observed in dense urban and rural areas.
Okumura-Hata propagation models - described in Annex A

A report to GSMA
22
These differences are indicative of improvements which might be expected when deploying UMTS900 technology and apply only with specific input parameters, as explained in more detail in Annex A.
4.1.2 Mobility Management

Mobility management is an issue faced by mobile operators due to non-uniform coverage of Node-Bs, which can lead to coverage holes. This is related to limitations of coverage provided per Node-B and the associated costs, and can impact call handover, which can directly affect a customers service experience. The improved Node-B coverage of UMTS900 technology will help overcome handover problems and thus improve the customer experience. Whilst it cannot be quantified at this point, it is believed that UMTS900 will lead to better mobility management by avoiding coverage holes in the network by enabling overlap of adjoining Node-B coverage areas. Moreover, 2.1 GHz can be used to complement the 900MHz band by using overlay deployments in areas with high capacity demand using macro-cells and micro (pico)-cells configurations in hotspot areas such as train stations and stadiums.
4.1.3 Network capex efficiencies and enhanced capabilities

When considering the deployment of UMTS900 technology, it is important to consider the replacement cycle of GSM network equipment and the strategic options and choices that an operator might have to make during this replacement phase. If an operator has an operational GSM network equipment, which is amortised (especially on the access side as BTS account for around 65% of Capex), the operator has to take strategic replacement strategy decisions. One option is to replace existing GSM equipment with new GSM systems, possibly with enhanced capabilities; another option (if the operator holds a UMTS license) is to replace with UMTS equipment. By following the first option the operator will limit the capabilities and efficiencies of their networks and there is the possibility that the technology may become obsolete within 10-15 years through lack of vendors support. In the second option operators might be restrained by the up-front investments needed to deploy UMTS2100 networks. The decision is easier if the operator faces problems which would be eased by deploying a UMTS network, such as voice channel congestion, lack of frequency channels to support more and significant take-up of bandwidthhungry services.

A report to GSMA
23
Whilst each case is operator and network specific, UMTS900 offers a new option for operators facing a GSM network equipment replacement cycle, with lower up-front investments compared to the UMTS2100, at the same time as enhancing network efficiencies and capabilities to handle both voice and data traffic.
4.1.4 Positioning against other access technologies

Cellular operators are about to face competition in their traditional markets from other access technologies, the most prominent of these being Mobile WiMax 802.16e. The roadmap planned for Mobile WiMax is to offer all-IP high data speed wireless services with lower up-front Capex investments than can be achieved with UMTS. This may present a viable business case especially in developing (emerging) markets, where mobile penetration or geographic coverage is relatively low and geospatial characteristics (majority of land comprises of rural areas) require significant Capex investments. However, in these cases, UMTS900 technology, due to its increased coverage capabilities as discussed in this paper, could possibly also provide a viable solution since it could also offer high data speed rates and provide voice traffic handling efficiencies with reduced Capex investments, when compared to UMTS2100. It must be noted that the issues discussed above are case-specific, dependent also on regional and market characteristics. Different access technologies do not necessarily have to compete, and in many cases they can complement each other. UMTS900 technology provides an extra option for mobile operators to position themselves and pursue their strategic choices based on both market characteristics and competitive forces in their respective markets.
4.1.5 Harmonisation of spectrum in the 900MHz band

The 900MHz band is one of the most used bands in the world. GSM operators all over Europe, Africa and Asia use this band extensively, which makes it one of the most harmonised bands in the world. All operators using the 900MHz have started with GSM services and most of them have already acquired 3G licenses at 2.1 GHz. This business evolution makes UMTS900 a most attractive option for operators and a likely follow-up technology in the 900MHz band. In order for operators to be able to utilise UMTS technology in the 900MHz band two prerequisites are required regarding spectrum availability: a. the NRA must allow the deployment of UMTS in the 900MHz band.
b. re-farming of 900MHz spectrum band in order for the operators to be able to deploy 5MHz (WCDMA) carriers This is an important issue as not all GSM license holders may have enough spectrum to accommodate the 5MHz carriers required for the deployment of UMTS technology. In this case, spectrum restructuring may necessary in order for operators to be able to deploy UMTS900.

A report to GSMA
24
If these points are resolved and both GSM and UMTS technologies can be deployed in the 900MHz band then this harmonisation of spectrum will allow maximum flexibility for mobile operators to use spectrum efficiently (through network and frequency planning process) to meet coverage and capacity requirements.
4.2 Scenarios
Scenarios have been developed for each of the four regions being considered (W. Europe, Asia Pacific, Sub-Saharan Africa, Middle East) which are characteristic of typical countries in each region.
4.2.1 Region characteristics

The characteristics of each region differ and the following parameters are taken into consideration: a. Geo-types: In each region the land area for each geo-type (dense urban (indoors / hotspots), urban, sub-urban and rural) is modelled. The proportion of land in each of these four types differs greatly in each region b. Network coverage: UMTS coverage varied dependent on the type of environment and, importantly, between regions c. Population split: the population split between regional geo-types varies in each of the four regions. This factor affects network capacity drivers and profiles have been assigned to different population groups in each region
d. Demand analysis: the demand for service will vary in each of the four regions, and three parameters have high variation: the penetration of WCDMA technology; usage profiles and ARPU. The analysis does not include any variation of penetration for UMTS900 and UMTS2100 since UMTS900 handsets availability is not considered as a significant impediment to UMTS900 take-up. Within each usage profile we consider the annual service usage per user (as Minutes of Usage (MoU), number of SMS and MB of data) and these parameters vary by region. ARPU is directly impacted by both levels of usage and service pricing. Further details and analysis are provided in Section 4.3 e. Operational costs: which vary between each of the four regions, e.g. subscriber acquisition costs and average churn rates. For every region there are two input parameters that can be altered and thus lead to different results: Type of mobile operator Demand scenario

A report to GSMA
25
4.2.2 Type of mobile operator

The baseline scenario is the case in which an operator deploys UMTS2100 to serve a specific geographic coverage and service demand. The analysis assumes that the operator is only deploying UMTS with no obligations to also deploy a GSM network. The implications and interoperability issues between GSM and UMTS900 technologies are analysed in Section 4.5. The starting point of UMTS coverage and penetration differs between each of the four regions examined, and the analysis assumes that an, WCDMA deployments before start of the modelling period are UMTS2100, after which time operators have the flexibility to choose between different technologies. The baseline scenario is compared with two others; one where an operator deploys UMTS900 technology only and the other where an operator has access to both 900MHz and 2100MHz spectrum and deploys both UMTS900 and UMTS2100 technology. In the latter case, it has been assumed that the operator deploys UMTS900 technology in dense urban (indoors/hotspots) and rural areas to provide extended coverage and in-building coverage, and UMTS2100 technology in urban and sub-urban areas to benefit from increased capacity, as discussed in Section 4.1.
4.2.3 Demand scenarios

The demand parameters, such as WCDMA penetration, usage profiles and ARPU differ significantly between the regions. Within each region three different scenarios related to demand and usage are used reflecting high, base and low demand. The base scenario data are derived from Ovum and Wireless Intelligence forecasts and other sources such as the Economist Intelligence Unit and the World Bank. The adjustments for the high and low scenarios and discussion on demand data is provided in Section 4.3.
4.3 Demand analysis

4.3.1 Overview of cost model
Figure 4.4 presents an overview of the structure of the cost model. The demand analysis is based on forecasts of subscriber penetration and usage profiles (voice, SMS, data) over the examined period. It drives the network dimensioning component to provide bandwidth demand calculations and also the Opex costs and service revenues as explained below. A detailed description of the model cost (Capex and Opex) calculations and financials is provided in subsequent sections of the report.

A report to GSMA
26
Figure 4.4 Overview of model structure

Key:Input sheet Calculation sheet
Demand
Demand_Inputs Subscriber Calcs
Output sheet
Costs
Network Cost & Financials
Voice usage Calcs and Data Usage Calcs
Capex Calcs
Opex Calcs
Demand Summary
Cost Calcs
Bandwidth Demand
Financial Network dimensioning

Network Design Inputs Equipment Calcs D3 D2 D1
Depreciation schedules
Service Pricing
Revenue Calcs
RoI
(*) Note: Indicative (major) flows shown only
NPV
IRR
Source: Ovum
4.3.2 Demand analysis model

The demand analysis model provides a high level view of the possible revenue from the development of WCDMA in various world regions. It incorporates data from Wireless Intelligence and other sources for each country to provide a region based view for each parameter to model. The model revenue outputs are derived from voice minutes, SMS and data volumes and tariffs. The model does not include international call / SMS revenues or premium rate SMS and data services. All national data is treated at a flat rate, without special pricing for html, ringtones or any other content. Data sources The Wireless Intelligence service provided data for subscriber numbers and population per country, growth over the previous years by quarter per country, penetration of WCDMA in certain countries and average by region, as well as growth for these selected countries and average voice minutes per user in each of the major regions.

A report to GSMA
27
Information on the percentage of subscribers that use a certain feature and the frequency of usage was derived from various sources including press releases, third party studies and extrapolation between countries and regions. Pricing information was derived from the lowest data tariff available, which tend to be business tariffs, in selected countries. Data sources were complemented by reasoned assumptions where specific data sources were not available. Volume The data sources were used to determine the demand on spectrum bandwidth, based on the penetration of each feature, the number of active users of SMS, voice and data features and frequency of use. Voice is calculated in minutes of use per user, SMS per average messages sent per user and data as an average volume downloaded. Capacity drivers such as games, ringtones, and web browsing are all calculated only in terms of the amount of data traffic each will generate. Reasonable assumptions regarding consumer behaviour and how this might evolve with time have been made, based on experience in the mobile and fixed internet content industries, first hand observations and data available for selected countries. Growth per region The variable that most affects results derived from the model are related to the growth of mobile phone connections and in particular WCDMA connections within the wider mobile phone base. The growth assumptions in the model are based on current and forecast mobile phone penetration in each country compared to the countrys population and the relative penetration of WCDMA phones within this installed base. The growth in the last year and last quarter of available data were assessed in order determine whether the market penetration had reached saturation point or not, and the growth prospects towards market saturation. The data is aggregated into a region profile, which includes the average penetration of connections and WCDMA connections in the region the homogeneity of the region, spread, standard deviation, etc is also assessed. This region profile is the used to model growth, assuming that growth in the region based on average growth for recent periods. Because WCDMA has been recently introduced in some regions there is little information to indicate how growth will slow as it approaches a saturation point, in these case we have made assumptions that seemed reasonable in light of more mature markets such as Asia.

A report to GSMA
28
Regulation, cultural factors, speed of product development and the business environment also affect growth. As such we have modelled the high demand case to assume a 20% increase in medium demand forecasts (based on Wireless Intelligence data). The low demand scenario assumes a 20% decrease in medium demand forecasts (based on WI data). Most of the variation in results can be accounted for changes in the growth in WCDMA penetration in the Asia region. This is both due to the already high penetration relative to other regions in percentage terms and to the very large population in the area. As the annual growth is compounded a small change in the expected growth of WCDMA becomes even more significant. Other variables produce the results that would be expected, as for example lowering price by a given percentage reduces the ARPU by the same percentage, but applying these changes within reasonable ranges only modify the results slightly when compared to growth. Demand analysis outputs Figures 4.5 presents the projections for WCDMA connections for all three scenarios (low demand, base (medium) demand, high demand) and for all four regions examined, over a 5 year time period.

A report to GSMA
29
Figure 4.5 WCDMA connections forecast

W. Europe: WCDMA connections 300 250 200 150 100 50 0 2007
Millions
2008 Low Demand
2009 Medium Demand
2010 High Demand
2011
Asia Pacific: WCDMA connections 600 500 400 300 200 100 0 2007
Millions
2008 Low Demand
2009 Medium Demand
2010 High Demand
2011
Middle East: WCDMA connections 30 25 20 15 10 5 0 2007
Millions
2008 Low Demand
2009 Medium Demand
2010 High Demand
2011

A report to GSMA
30
25 20 Millions 15 10 5
Sub-Saharan Africa: WCDMA connections
0 2007
2008 Low Demand
2009 Medium Demand
2010 High Demand
2011
Source: Ovum Reasonable assumptions were made for the subscriber usage profiles for voice, SMS and data services for each of the four regions examined. These were based on Ovum and Wireless Intelligence data and forecasts. This demand data was fed into the revenue analysis, to produce the ARPU per subscriber forecasts shown in Figure 4.6 for all four regions examined. ARPU levels along with usage profiles per subscriber and overall WCDMA penetration then determine the services revenues calculation.
Figure 4.6 Annual ARPU ($) per subscriber for four regions examined
Annual ARPU ($) per subscriber
$900 $800 $700 $600 $500 $400 $300 $200 $100 $0 2007 W. Europe 2008 Asia Pacific 2009 Sub-Sah Africa 2010 Middle East 2011
Source: Ovum All these demand analysis outputs feed as inputs into Capex and Opex costs and services revenues calculations respectively and they will overall drive financial analysis.

A report to GSMA
31
4.4 Network capex analysis

4.4.1 Main assumptions
While conducting the costing analysis there are several assumptions that must be made. Below the main assumptions that have been adopted in order to derive results and conclusions are presented: All assumptions made like WCDMA penetration, usage profiles, population and geographic area coverage for a specific region have been kept the same for all different scenarios examined (i.e. UMTS900 only, UMTS2100 only etc). In this way comparisons are performed on same grounds. Sensitivity analysis of how these parameters may be affected are presented in different section No GSM rollout has been considered. Network dimensioning is focusing only on WCDMA technology. Specific implications of GSM and WCDMA collocation and coexistence are addressed in different sections Starting point of network dimensioning for the four regions examined is different in terms of WCDMA subscribers, population and geographic coverage etc. We assume that anything related to the period before the starting point is irrelevant. In this way both a paper thin network (case of Africa) and a more advanced and mature one (case of Western Europe) are examined WCDMA Release 99 has been considered for two reasons. Firstly, because network dimensioning has been performed considering voice as the main service offered in all regions examined (especially the ones of Middle East, Sub-Saharan Africa and Asia Pacific); thus associated quality parameters (as Eb/No, interference margin) are calculated mainly for voice by taking also into consideration some of the data services characteristics. Secondly, for simplicity reasons and in order to avoid complexity of assessing impact on the core network (moving from Release 99 to Release 5 or 6 entails significant Capex investment since architecture is different division of switch and control planes). However, qualitative analysis of how WCDMA Release 5 (HSDPA) and WCDMA Release 6 (HSUPA) on the access side could possibly impact results is provided. Moreover, it must be noted that this analysis does not by any means specify or recommend the use of different WCDMA Releases for different regions since this is considered to be completely operator specific choice All final results are based on comparative rather than absolute basis
4.4.2 Capex analysis

After following the methodology described in Annex A (Detailed model description), the volumes of network equipments required are calculated. Figure 4.7 provides an example on the type of network equipment that have been dimensioned in order to satisfy both coverage and capacity requirements on an annual basis.

A report to GSMA
32
Figure 4.7 Example of volumes of network equipment required after network dimensioning performed
Equipment volumes required TRX Node-B RNC MSC/VLR HLR SGSN GGSN 2007 354,741 118,247 924 18 13 6 3 2008 2009 37,404 38,406 12,468 12,802 98 101 Illustrative 18 31 12 21 11 21 5 9 2010 40,428 13,476 106 57 39 41 17 2011 51,321 17,107 134 106 73 80 32
Source: Ovum model Network equipments costs, or other related activities (i.e. sites acquisition / construction) are also taken into consideration in order to derive annual Capex costs. Figure 4.8 provides an example of the costs of network equipment or associated activities that have been considered. It must be noted that this list is by no means exhaustive and in real life Capex costs include other elements as well. However, list above comprises of the main access network elements. For the purposes of this study, which is deriving results on a rather comparative than absolute basis, analysis conducted is regarded as sufficient. Figure 4.8 Example of costs of network equipment and associated activities
Network elem ent unit costs (nom inal US $) TRX Node-B Node-B installation and site acquisition costs Node-B annual lease RNC MSC/VLR HLR SGSN GGSN 2007 15,000 75,000 30,000 20,000 1,000,000 2,000,000 1,500,000 750,000 750,000 2008 15,000 75,000 30,000 20,000 1,000,000 2,000,000 1,500,000 750,000 750,000 2009 2010 15,000 15,000 75,000 75,000 30,000 30,000 20,000 20,000 1,000,000 1,000,000 Illustrative 2,000,000 2,000,000 1,500,000 1,500,000 750,000 750,000 750,000 750,000 2011 15,000 75,000 30,000 20,000 1,000,000 2,000,000 1,500,000 750,000 750,000
Source: Ovum model Also, it must be noticed that special considerations have been taken for evolution of costs over time as well as any possible cost differences between the two technologies; however rather insignificant in both cases. By multiplying network elements volumes, and associated activities, with relevant costs the total Capex spending on an annual basis is thus calculated. Cumulative Capex is then derived by Capex calculations as described above.
4.4.3 Opex analysis

Opex costs are comprised of network related and non-network related costs.

A report to GSMA
33
Figure 4.9 summarises all Opex costs considered and detailed analysis follows in subsequent sections. Figure 4.9 Opex costs considered
Total Opex
Network related Opex
Non-network related Opex
OAM&P costs
Other network related Opex costs
Applicable to new subs
Applicable to all subs SGA costs Customer care costs
SAC costs
SIM card costs
Source: Ovum Network related Opex costs Network related costs are comprised by Operations, Administration, Maintenance and Provisioning (OAM&P) costs and other network related operational costs. OAM&P costs are directly driven by the number of network elements and a well know and adopted methodology has been followed2 in order to come up with relevant calculations. OAM&P costs have been calculated as a percentage of Cumulative Capex costs, as presented above, on an annual basis. The percentage used is 10% of Cumulative Capex costs per annum. Other network related Opex costs include transmission annual lease costs, sites and switches annual lease costs and other minor ones. Non-network related Opex costs Non-network related costs are mainly driven by the number of subscribers and consist of costs associated with acquisition of new subscribers and marketing and sales costs of all existing ones. More specific, non-network related Opex costs include Subscriber Acquisition Costs (SAC) (that consist of commissions, handsets subsidies, promotion packages etc.) and cost of SIM cards that are applicable for only all the new subscribers in the network (new additions plus churn subscribers) and sales and marketing costs (calculated as a percentage -in the range 10-12%of revenues) and customer care related costs (i.e. customer care, charging and billing) that are applicable for all subscribers in the network.
http://www.comlab.hut.fi/studies/3510/PDOTS_Smura_300306.pdf

A report to GSMA
34
It must be noted that in real life operators incur other significant Opex costs (as interconnection and roaming) that are not considered in this study. The main reasons for that is simplicity and comparative nature of analysis and final results.
4.5 Comparative analysis / Summary of the regions

As explained above, different scenarios have been assessed and the combination of results is numerous. In order to strengthen the impact of the messages derived we highlight and focus on the three most important ones which are: Cumulative Capex (over the course of 5 years) percentage difference from baseline scenario Cumulative Capex (over the course of 5 years) percentage difference from baseline scenario providing breakdown for four types of environments (dense urban, urban, suburban, rural) Total Capex and Opex costs (over the course of 5 years) percentage difference from baseline scenario
A more detailed presentation of results is further provided in Annex B.
4.5.1 Cumulative Capex

Cumulative Capex is calculated by adding Capex costs over the course of 5 years time period for the four regions and the three scenarios examined as presented above. Figures 4.10 and 4.11 present the Cumulative Capex costs for the UMTS900 only and UMTS900 and UMTS2100 cases as a percentage of the Cumulative Capex costs for the UMTS2100 only case, for all three demand scenarios.

A report to GSMA
35
Figure 4.10 Cumulative Capex costs for UMTS900 only as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex: UMTS900 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac
W. Europe
Sub-Saharan Africa
Low Demand
Medium Demand
High Demand
Source: Ovum Figure 4.11 Cumulative Capex costs for UMTS900 and UMTS2100 as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex: UMTS900 and UMTS2100
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac
W. Europe
Sub-Saharan Africa
Low Demand
Medium Demand
High Demand
Source: Ovum By studying the graphs above, main results derived are summarised below: Case of UMTS900 only The Sub Saharan Africa case shows the highest Cumulative Capex reductions of 41% followed by Western Europe with 40%, Middle East with 36% and Asia Pacific with 32%

A report to GSMA
36
Apart from the case of Sub Saharan Africa Cumulative Capex differences are decreased as demand grows (High Demand scenario). This is mainly due to the fact that the greater benefits of UMTS900 are mainly derived from coverage and not capacity related issues. In the case of Sub Saharan Africa, demand does not vary greatly over the three scenarios and due to landscape characteristics (majority of land considered as rural area ) coverage is always the main driver as far as Capex investments are concerned
Case of UMTS900 and UMTS2100 The Sub Saharan Africa case shows the highest Cumulative Capex reductions of 37% followed by Middle East with 31%, Western Europe with 27% and Asia Pacific with 16% Apart from the case of Sub Saharan Africa Cumulative Capex differences are decreased as demand grows (High Demand scenario). This is mainly due to the fact that the greater benefits of UMTS900 are mainly derived from coverage and not capacity related issues. In the case of Sub Saharan Africa, demand does not vary greatly over the three scenarios and due to landscape characteristics (majority of land considered as rural area ) coverage is always the main driver as far as Capex investments are concerned
It must be noted that overall results presented above refer to a rather simplistic network planning case. The real case of assessing impact of UMTS900 should be treated as case and more specifically market and operator specific. For example, a detailed overlay network of using UMTS900 macrocells combined with UMTS2100 microcells in heavy traffic profile hotspots has not been extensively examined. The results above should by no means disregard or downgrade the business case of deploying UMTS2100. Overall results might seem to be better in the case when using UMTS900 only technology; however this can be challenged for the following reasons: Scenarios and analysis provided do not necessarily capture real cases at the greatest extent (explained above) Operators might face spectrum availability limitations in the 900MHz band regarding the deployment of UMTS (especially in the case of an existing GSM900 operator). Frequency re-use pattern might thus be limited and availability of 2100MHz spectrum might be extremely useful in order to offer UMTS services in the market Operators might prefer to deploy UMTS2100 technology in dense urban and urban areas due to favourable adjacent cell interference (ACI) characteristics. This last point is especially important in dense populated areas in order to ensure adequate quality of service

A report to GSMA
37
For all the reasons provided above, results presented and associated conclusions should be carefully treated.
4.5.2 Cumulative Capex for four types of environments (dense urban, urban, suburban, rural)
A further breakdown of Cumulative Capex on type of environment basis is further provided. In order to calculate the breakdown of costs per type of environment the approach described below has been followed: a) Access network: Access network elements (as TRX, Node-B, RNC) and other associated (as sites construction/acquisition) costs are directly calculated per type of environment as described in sections 4.1 and 4.2 and Annex A b) Core network: Core network elements (as MSC/VLR, HLR, SGSN, GGSN etc) costs have been allocated to different types of environment based on the percentage of number of subscribers as of total forecasted per type of environment respectively. Therefore, percentage of subscribers as of total has been used as the driver for cost allocation of core network elements and provides a proxy of how core network elements Capex costs might be incurred and distributed per type of environment Figures 4.12 presents the Cumulative Capex costs per type of environment for UMTS900 only case as a percentage of Cumulative Capex costs for UMTS2100 only case and for all four regions examined. Further breakdown per type of demand scenario for all four regions is provided in Annex B.

A report to GSMA
38
Figure 4.12 Cumulative Capex costs per type of environment for UMTS900 only as a percentage of Cumulative Capex costs for UMTS2100 only
Cumulative Capex for different environments - UMTS900 only 100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% W. Europe AsiaPac Dense Urban Urban Middle East Suburban Rural Africa
Source: Ovum By studying the graph above, main results derived are summarised below: Case of UMTS900 only Dense urban environment case shows the highest Cumulative Capex reduction in the range of 37%-46%, followed by rural environment (range of 33%-46%), suburban environment (range of 26%-34%) and finally urban environment (range of 20%-36%) for all regions examined The Sub Saharan Africa case shows for all types of environments (apart from urban one) the highest Cumulative Capex reduction among all regions examined
4.5.3 Total Capex and Opex costs

Total Capex and Opex costs are calculated by adding Capex and Opex costs on an annual basis over the course of 5 years time period for the four regions and the three scenarios examined as presented above. Figures 4.13 and 4.14 present Total Capex and Opex costs for the UMTS900 only and UMTS900 and UMTS2100 cases as a percentage of Total Capex and Opex costs for the UMTS2100 only case, for all three demand scenarios.

A report to GSMA
39
Figure 4.13 Total Capex and Opex costs for UMTS900 only as a percentage of Total Capex and Opex costs for UMTS2100 only
Total Costs: UMTS900 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac Low Demand
W. Europe
Sub-Saharan Africa
Medium Demand
High Demand
Source: Ovum
Figure 4.14 Total Capex and Opex costs for UMTS900 and UMTS2100 as a percentage of Total Capex and Opex costs for UMTS2100 only
Total Costs: UMTS900 and UMTS2100
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% Middle East
Asia-Pac Low Demand
W. Europe
Sub-Saharan Africa
Medium Demand
High Demand
Source: Ovum By studying the graphs above, main results derived are summarised below:

A report to GSMA
40
Case of UMTS900 only The Sub Saharan Africa case shows the highest Total Capex and Opex costs reductions of 38% followed by Middle East with 32%, Asia Pacific with 17%, and Western Europe with 10% For all regions examined, Total Capex and Opex costs differences are decreased as demand grows (High Demand scenario). This is mainly due to the fact that the greater benefits of UMTS900 are mainly derived from coverage and not capacity related issues
Case of UMTS900 and UMTS2100 The Sub Saharan Africa case shows the highest Total Capex and Opex costs reductions of 33% followed by Middle East with 27%, Asia Pacific with 8% and Western Europe with 7% For all regions examined, Total Capex and Opex costs differences are decreased as demand grows (High Demand scenario). This is mainly due to the fact that the greater benefits of UMTS900 are mainly derived from coverage and not capacity related issues
4.5.4 Implications of upgrading to WCDMA Release 5 (HSDPA) and Release 6 (HSUPA)

As already explained, analysis so far is based on WCDMA Release 99. Reasons for that have been provided above. By upgrading access network to Release 5 and eventually Release 6 new capabilities are further introduced in the network. The case of upgrading core to Release 5 and 6 is not addressed due to increased complexity entailed. The areas where cost model is altered for the methodology described above are: Network dimensioning is now based on voice services at 64.4kbps and basic data services at 384kbps. This will allow for the case of Circuit Switched video calls and advanced relatively high-speed data services (wireless broadband) Maximum capacity per cell is consequently altered (spectral efficiency). Enhanced capabilities introduced in the network due to higher modulation and coding rates allow for the offering of higher speed rates per user Maximum coverage per cell is consequently decreased. The actual cost of providing higher data speed rates is decreased cell range per Node-B due to higher noise rise involved Impact on available spectrum bands should be also considered. Referring to spectral efficiency issue explained above Network element costs are changed. Access network elements costs are most possibly increased (though not significant variations expected) due to enhanced capabilities embedded in new Releases

A report to GSMA
41
All these factors are taken into consideration and the case of upgrading access to Release 5 and 6 is assessed.
4.6 NPV analysis

By means of a straightforward cash flow calculation, a simple NPV analysis was prepared for all of the cases and regions examined. The total costs have been derived as presented above in Section 4.5.3. Revenues are derived by calculating total service usage and associated prices for all of the services examined (voice, SMS, data) along with their evolution over time for all four regions, as presented below.
4.6.1 Revenue calculations

Cash flow is calculated for all different scenarios and cases. In order to derive NPV values, a rate of return (WACC) of 10% is used. Usually, WACC (pre-tax) rates adopted worldwide on behalf of NRAs in the mobile industry are found in the range of 10% to 16%. The lower end rate of 10% is in this study. NPV refers to a 5-years time period. As previously discussed, only comparative and not absolute results are presented in order to capture and reinforce messages derived. Figure 4.11 presents NPV percentage improvements, in the case of medium demand scenario, over the case of UMTS2100 only for both the cases of UMTS900 only and UMTS900 and UMTS2100 respectively and for all four regions. By studying Figure 4.15 the main conclusions drawn are: The case of UMTS900 only presents greater opportunities for NPV improvements over a 5-years period The cases of Sub-Saharan Africa and Middle East present greater opportunities for NPV improvements due to reduced Capex and Opex costs incurred (section 4.5.3- Total costs)
Figure 4.16 presents NPV improvements over UMTS2100 only case when compared to the UMTS900 only case for all three different demand scenarios and four regions examined.

A report to GSMA
42
Figure 4.15 NPV percentage improvements (medium demand scenario) over UMTS2100 only case for both UMTS900 only and UMTS900 and UMTS2100 cases
NPV improvements over UMTS2100 only (*)
120% 100% 80% 60% 40% 20% 0% W. Europe AsiaPac UMTS900 only
(*) refers to Medium Demand scenario
ME
Africa
UMTS900 and UMTS2100
Source: Ovum model
Figure 4.16 NPV percentage improvements over UMTS2100 only case for UMTS900 only case and three different demand scenarios
UMTS900 only: NPV improvements over UMTS2100 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% W. Europe AsiaPac
Medium Demand
ME
High Demand
Africa
Low demand
Source: Ovum model Figure 4.17 presents NPV improvements over UMTS2100 only case when compared to the UMTS900 and UMTS2100 case for all three different demand scenarios and four regions examined.

A report to GSMA
43
Figure 4.17 NPV percentage improvements over UMTS2100 only case for UMTS900 and UMTS2100 case and three different demand scenarios
UMTS900 and UMTS2100: NPV improvements over UMTS2100 only
100% 90% 80% 70% 60% 50% 40% 30% 20% 10% 0% W. Europe AsiaPac Low Demand ME High Demand Africa Medium Demand
Source: Ovum model
4.6.2 NPV scenarios

The analysis so far has assumed that different inputs (such as WCDMA penetration, geographic coverage) have remained the same in order to calculate, on an equal basis, potential capex reductions through the introduction of UMTS900 technology. The capex reductions derived in Section 4.5 can be re-assessed to reflect a more realistic situation, in which the operator is able to take advantage of the cost reductions to fund further capex, which might be to attract more customers and increase the number of connections or increase geographic coverage. Each of these is discussed below: WCDMA connections In both cases (UMTS900 only and UMTS900 & UMTS2100) Cumulative Capex requirements are reduced when compared to the case of UMTS2100 only. Mobile operators might well leverage this advantage by increasing the coverage reach of their network and hence enhance ability of serving more end-users. In order to calculate additional subscribers the following approach has been followed: Calculate average Capex spending per subscriber over the period of 5-years for each of the three cases (UMTS2100 only, UMTS900 only, UMTS900 and 2100) respectively and for the medium demand scenario Calculate additional end-users that can be accommodated by dividing the differences in Cumulative Capex between baseline case (UMTS2100 only) and the two examined ones (UMTS900 only and UMTS900 and 2100) by the

A report to GSMA
44
average Capex spending per subscriber for every case respectively. In this way we assume that baseline cases Cumulative Capex spending is available in two other cases as well and this can be used by the operators to accommodate more subscribers in their networks Additional subscribers numbers for every case are then considered. Figure 4.18 Figure 4.21 present the additional subscribers calculated, with the method described above, for all four regions examined. Figure 4.18 WCDMA subscribers for different cases and for all four regions
W.Europe: UMTS subscribers for different scenarios
400,000,000 350,000,000 300,000,000 250,000,000 200,000,000 150,000,000 100,000,000 50,000,000 0 Year 1
Year 2 UMTS2100 only
Year 3 UMTS900 only
Year 4 UMTS900&2100
Year 5
Source: Ovum model

A report to GSMA
45
Figure 4.19 WCDMA subscribers for different cases and for all four regions
Asia Pacific: UMTS subscribers for different scenarios

700,000,000 600,000,000 500,000,000 400,000,000 300,000,000 200,000,000 100,000,000 0 Year 1
Year 3 UMTS900 only
Year 4 UMTS900&2100
Year 5
Source: Ovum model Figure 4.20 WCDMA subscribers for different cases and for all four regions
Middle East: UMTS subscribers for different scenarios
40,000,000 35,000,000 30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0 Year 1
Year 3 UMTS900 only
Year 4 UMTS900&2100
Year 5
Source: Ovum model

A report to GSMA
46
Figure 4.21 WCDMA subscribers for different cases and for all four regions
Sub-Saharan Africa: UMTS subscribers for different scenarios
30,000,000 25,000,000 20,000,000 15,000,000 10,000,000 5,000,000 0 Year 1
Year 3 UMTS900 only UMTS900&2100
Year 4
Year 5
Source: Ovum model WCDMA geographic coverage Geographic coverage requirements have been consequently adjusted in order to ensure that additional subscribers can be served and network reach is higher. Figure 4.22 presents geographic coverage requirements followed in the original case followed and the updated ones for both cases (UMTS900 only and UMTS900 & UMTS2100) and all four regions examined.
Figure 4.22 WCDMA network geographic coverage requirements

Overall WCDMA geographic coverage
Original case Updated case for UMTS900 only Updated case for UMTS900 & 2100
W. Europe
84% 92% 88%
Asia Pacific
58% 73% 67%
Middle East
34% 53% 46%
Sub-Saharan Africa
11% 31% 23%
Source: Ovum model NPV scenarios calculation By adjusting the model to take account of these changes to the number of WCDMA connections and WCDMA geographic coverage NPV calculations have been updated. As shown in Figure 4.23, which presents for both cases (UMTS900 only and UMTS900 & UMTS2100) the updated NPV calculations/improvements compared to the UMTS2100 only case for the medium demand scenario.

A report to GSMA
47
Figure 4.23 Updated NPV percentage improvements over UMTS2100 only

NPV improvements over UMTS2100 only
120%
100% 80%
60% 40%
20%
0% W. Europe Asia Pacific UMTS900 only Middle East UMTS900 & 2100 Sub-Saharan Africa
Source: Ovum model The main points derived by studying Figure 4.23 are: UMTS900 only case presents the greatest opportunities regarding potential NPV improvements over the course of 5 years Asia-Pacific and W. Europe present the cases where greater opportunities regarding NPV improvements for UMTS900 technology (either as a stand-alone or combined with UMTS2100) are expected. The main reasons for that is on the one hand absolute number of additional subscribers is higher in these two regions (due to higher potential addressable market) contributing to higher expected revenues and on the other that additional Capex spending per subscriber required to increase coverage/capacity requirements is lower due to landscape characteristics and already experienced WCDMA penetration (easier to cover additional capacity than greatly increase coverage in rural areas)

A report to GSMA
48
5 Regulatory analysis
5.1 Introduction
One of the responsibilities placed on communications regulators is to ensure that the interests of consumers are served by creating an environment which promotes competition, and in the area of wireless communications the mandate includes ensuring the optimal use of the electromagnetic spectrum and ensuring that a wide range of communications services can be provided. The spectrum management policies followed by the National Regulatory Authorities (NRAs) together with any legal or regulatory framework defined by European and International bodies will play a critical role in the deployment of UMTS900 networks. The issue of whether to re-farm (i.e. reallocate) spectrum in the 900MHz bands which is currently assigned to GSM services to allow mobile network operators to operate UMTS services in the spectrum is being considered by a number of regulators. Within this section of the report: the key issues related to UMTS900 deployment and implications for NRAs are assessed, estimations of the time schedule for such development are provided, together with possible future spectrum strategies the regulatory position of NRAs in four geographical regions are examined, with regards to spectrum re-farming processes, technology-neutral/technologyspecific policies and UMTS900 deployment.
In the following sections, there are references to technologies developed under the ITU IMT-2000 framework. This is to encompass technologies used in regions where UMTS is not the only standard for 3G services (e.g. APAC). The positions, decisions and strategies of regulators are stipulated in terms of IMT-2000. Hence, any reference to IMT-2000 technologies also includes UMTS-based services.
5.1.1 What is Spectrum re-farming?

Spectrum re-farming is the process of redeploying spectrum from existing uses and reallocating it to others. When undertaken by regulatory authorities, it is to benefit consumers and service providers, to allow spectrum to be used more efficiently or allocated to another higher value or more important use. Existing spectrum uses, in particular those assigned with large blocks of spectrum, are invariably affected by such processes. Under this definition, re-farming of the existing 2G spectrum to 3G services refers to the re-designation of some or all of the currently assigned frequency bands for

A report to GSMA
49
GSM networks to UMTS networks. In particular, UMTS900 refers to the deployment of UMTS network which utilises frequencies in the 900MHz GSM band. Following this implementation, mobile network operators would allow to (re)use the allocated GSM spectrum which provides greater UMTS coverage than the current UMTS in the 2100MHz bands.
5.1.2 Technology-specific and technology-neutral spectrum management policies

There are two main spectrum management approaches to allocating frequency bands to current and future technologies: the technology-neutral and the technology specific approach, although hybrid versions of the two policies have been adapted by the NRAs. A technology-specific policy designates technologies to specific frequency bands, and dictates the use of the spectrum for these technologies. This approach has the disadvantage the as new technologies and services develop it can be challenging to optimally manage of the spectrum. To allow the spectrum to be allocated more effectively and increase spectrum efficiency, the regulators need to define and predict the future technologies which might be successfully deployed. The regulators usually do this in an open and transparent way, consulting the potential players and taking into account new markets and services. On the other hand, under a technology-neutral spectrum management policy, the regulator allocates spectrum flexibly without designating the technology to be used within frequency bands. Spectrum-licence holders are able to use the spectrum according to their market needs and spectrum demands. The most challenging issue with this approach is to ensure an interference free environment to other services, adjacent bands and/or other spectrum users. Although technology neutrality allows the deployment of new technologies within certain spectrum allocations, the actual usage of these allocations may depend not only on the availability of technology equipment to access certain frequency bands, but also on the international (or regional) regulatory framework. Neutrality can therefore be interpreted in different ways across the spectrum bands. For example, a technology-neutral frequency band could be used only for UMTS networks, allowing the use of TDD or FDD multiplexing techniques in the same band or the use of compatible radio equipment. In addition, the same policy could allow the deployment of only same generation mobile networks within the same band. For the purpose of this study, the term technology-neutral for mobile services indicates the use of the spectrum for GSM or/and any UMTS version standard. This is exactly what led to the definition of an IMT-2000 family at the International Telecommunications Union.

A report to GSMA
50
5.2 Key issues

Spectrum availability Some NRAs, especially in developed countries, have already initiated discussions about new 2G spectrum management policies, the re-farming of GSM bands and the processes to be followed for the migration of GSM services to UMTS services. Spectrum availability in the GSM-900 and E-GSM bands is one of the key issues to be addressed by the NRAs, since these bands are highly populated by current GSM operators. In many countries the deployment of UMTS900 networks will require the redevelopment of national frequency plans to ensure better spectrum utilisation and distribution. This is made more difficult by the requirement to allocate UMTS spectrum in 5MHz blocks. It is essential to ensure that both GSM and UMTS services can continue to operate in a parallel and well-balanced manner as GSM networks will continue to be the core business for the majority of mobile operators. The protection of GSM operation needs to be reassured, since the GSM bands will continue to be used for many years. Sharing of 900MHz spectrum Apart from the spectrum availability issue, technical considerations also need to be taken into account. NRAs will have to access the impact of allowing the GSM and UMTS services to co-exist. Some of the regulators are considering the UMTS900 deployment under certain technical criteria which will ensure interoperability and compatibility of both services. Interference effects among the services operating within the same band need to be minimised. In most countries, extensive coordination with adjacent countries is also an important task in order to ensure an interference-free environment on each side of the border. GSM licence renewal Another challenge NRAs will face is the renewal process of the current GSM licences. For the majority of the cases the second generation licences are technology specific allowing only GSM services to be used in the designated bands. Although the re-farming of the GSM bands allows the co-existence of second and third generation mobile services, such arrangements raise concerns among the mobile industry. Competition issues are crucial to be addressed especially in countries with UMTS greenfield operators. Mobile operators, who granted 3G licences under specific market conditions, will face a disadvantageous effect compared to other operators in case GSM-900 band will be made available for the provision of 3G services. Impact of UMTS900 on existing GSM and UMTS operations The full use of 900MHz spectrum for the provision of 3G services in the future may be a barrier for new entrants. The licence holders of 900MHz band may be able to deploy UMTS900 in current GSM spectrum in some countries without applying for a

A report to GSMA
51
separate 3G licence. Furthermore, holders of current 3G licences may believe that their spectrum will be undervalued as a result of additional spectrum becoming available for 3G. It is therefore important that before altering the current market conditions or the conditions upon which the licences were granted, NRAs have to consider future implications. Spectrum pricing or national roaming NRAs will have to develop new pricing formulas for spectrum fees. In cases where 2G operators who are not 3G licence holders are allowed to use their spectrum for 3G services, NRAs need to consider whether these operators should pay a levy as an additional payment for the extra usage of the spectrum. Mobile operators need to know what is available and its cost in order to define their strategies and invest accordingly. Vendors and handsets manufacturers need also a clear view when deciding to invest in new equipment. Alternatively, NRAs may provide a regulatory framework for national roaming agreements to compensate the slightly higher network costs of the operators without 900MHz spectrum. The impact of EU directives and national developments The regulatory framework of the EU directives is a significant factor which could speed-up or delay the deployment of UMTS900 networks. Most of the NRAs are reluctant to move towards the UMTS900 adaptation due to these directives. However, at the European level there are processes in progress regarding the formulation of the regulatory framework for the deployment of 3G networks within the 900 and 1800MHz. This is highlighted with the technical studies which the CEPT Electronic Communications Commission (ECC) has undertaken and the consultations which are in progress. In other continents, NRAs are willing to move towards spectrum-neutrality approaches and allow the deployment of UMTS networks in the 900MHz band under an international harmonisation process. In some countries, these issues have been debated and resolved: In Europe, France has adopted a spectrum liberalisation policy by allowing the reutilisation of GSM bands. The French regulator, considering UMTS coverage requirements, industry views and the European and international frameworks, will issue a transitional implementation plan in 2007-2008 to ensure the equitable distribution of GSM bands to mobile operators. France is the first country in Europe where the GSM licenses have been renewed and where the re-farming of 2G frequencies has been approved Finland will follow an approach similar to France by the end of 2007 In the UK, Ofcom has already indicated its intention to extend spectrum trading to the 2G and 3G bands in 2007 The Hong Kong regulator has issued a consultation on the renewal of mobile networks licences. It was decided that spectrum designated for second generation services could be used for 2G and 3G services. This plan will be

A report to GSMA
52
implemented in phases to allow operators to upgrade their networks and review their strategies In the Middle East, there is limited deployment of 3G/UMTS services. The NRAs, taking into consideration a rather slow uptake of UMTS deployment in the region, are in favour of technology specific approaches for mobile services. The core band of UMTS is the preferred option for allocating UMTS spectrum. However, CITC (the regulator in Saudi Arabia) is the exception, by allowing the operators to use either second or third generation mobile technology in any spectrum band (including the 900MHz) allocated for mobile services. Similarly, in sub-Saharan region, 3G/UMTS deployment is in its very early stage. Only one country has granted licences for UMTS deployment. With technology-specific approaches being widely adopted across these countries, deployment of UMTS900 networks will not be a concern for the NRAs for the next 3-4 years, since the NRAs wish to protect national GSM operations.
Although international harmonisation processes and a clear European regulatory environment will accelerate further developments most of the NRAs examined will move towards a technology neutral approach in order to allow UMTS networks to operate in the GSM bands. Technology neutrality and flexible access to spectrum bands has become a critical factor to allow new technologies to be introduced. This approach to spectrum management will eventually allow mobile operators to allocate their spectrum in a flexible way and increase spectrum utilisation. The operators will be able to optimise their network capacity according to traffic requirements, market trends and the introduction of new services. It is expected that GSM networks will progressively migrate to become IMT-2000/UMTS networks. The migration schedule will depend on market demand, national regulatory conditions and licensing schemes.
5.3 Regulatory position in different regions

5.3.1 Western Europe / EU
The European Regulatory framework To reserve frequency bands for mobile communications in the European Community the EU has produced specific technology driven directives for GSM900. The designation of specific technologies in certain frequency bands has however created inflexibilities, according to CEPT / ECC. Any modification or abrogation of the Directives results in a long administrative process, since any changes require a proposal from the EC to initiate a co-decision process (requiring agreement from the Council and the European Parliament). These changes are required when a market does not develop as expected or, in the case of 3G services, when a spectrum band might be opened up to include applications and services due to technological or market developments. In the case of GSM-900, the

A report to GSMA
53
spectrum harmonisation measures have had positive results, since they allowed spectrum availability in the pan-European market and operational re-usability on a multi-national basis. The GSM bands (880-915 MHz, 925-960 MHz, 1710-1785 MHz and 1805-1880 MHz) have been subject to several harmonisation measures taken at EU level or by the ECC3: Council Directive (87/372/EEC) and the related Council Recommendation (87/371/EEC), which came into force in 1987 ERC Decision(94)01 on the frequency bands to be designated for the coordinated introduction of the GSM digital pan-European communications system ERC Decision(95)03 on the frequency bands to be designated for the introduction of DCS 1800 ERC Decision(97)02 on the extended frequency bands to be used for the GSM Digital Pan-European Communication System
Today, these GSM bands are intensively used by GSM networks while at the same time UMTS networks are rapidly developing in the 2GHz band. The New Regulatory Framework for electronic communications infrastructure and associate services states that the EU-members have to ensure that Spectrum Management Authorities (SMAs) take the utmost care to make regulation technology neutral, i.e. they should neither impose the use of a particular type of technology nor discriminate against it. This does not preclude taking proportionate steps to promote specific services where justified4. An important point relates also to one of the conditions mentioned in the Authorisation Directive is the Designation of service or type of network or technology for which the rights of use for the frequency has been granted. Identifying particular technologies is therefore permitted by the Authorisation Directive as long as it is justified5. It is important to note that SMAs are not responsible to predict which technologies will serve the most markets and consumers. The industry itself, standardisation bodies and key market players will be responsible to identify technologies needs.
ECC (2006) - Cover note to draft ECC/DEC/(06)MM on designation of bands 880-915MHz,
925-960MHz, 1710-1785MHz and 1805-1880MHz for terrestrial IMT-2000/UMTS systems

4
ECC Report 80 (2006) Enhancing Harmonisation and Introducing Flexibility in the
Spectrum Regulatory Framework

5
ECC Report 80 (2006) Enhancing Harmonisation and Introducing Flexibility in the
Spectrum Regulatory Framework

A report to GSMA
54
In the case of IMT-2000 technologies, the ECC/CEPT report mentions that administrations sometimes need to balance between conflicting requirements in order to ensure effective use of spectrum. It is therefore necessary in each case to balance the principle of technology neutrality and the aim of effective and efficient spectrum use. The European Commission considers also the harmonisation of radio frequencies allocations as an important task which member states must promote to ensure effective and efficient spectrum use. The Commission, following a market-based approach to spectrum management within the European Union, has proposed that frequency bands used for terrestrial mobile communication services, such as GSM and 3G networks, should be under a tradeable and flexible in-use regime in order to ensure a smooth transition for markets and services relying on spectrum usage6. At meetings in January 2007, the EC indicated that the GSM900MHz and GSM1800MHz bands will be approved for UMTS900/1800 use in September/October 2007. In summary, for the migration of 2G services to 3G services or even other future mobile services, it is ECC/CEPTs opinion that the Directive has to be abrogated, with the process to start as soon as possible. The ECC has proposed the designation of the above GSM frequency bands (880-915 MHz, 925-960 MHz, 1710-1785 MHz and 1805-1880 MHz) for the terrestrial IMT-2000/UMTS systems, subject to market demand and national licensing schemes and that administrations shall take all necessary measures to ensure the protection of continuing GSM operations in these bands. Studies of the approach taken by some regulators are examined in more detail below. These cases present the opinions of NRAs and their views in regards to 2G spectrum re-farming and UMTS900 deployment, potential implications and proposed measures. During process of reviewing spectrum re-farming the regulators have invited operators and other market players to provide comments in consultation papers and submit proposals. Appendix A shows the GSM-900 and E-GSM frequency bands allocations in 5MHz blocks for the EU-15 countries. The table shows the frequency blocks allocated to operators and the expiration dates. The fact that the expiration dates and the precise portion of the band allocated may vary, these allocations should not prevent spectrum in the 900MHz band from being allocated for UMTS900 use. The situation in the UK There are four operators currently in the UK offering second generation (2G) mobile telephony services with licences to use bands of spectrum at 900MHz and
EC COM (2005)400 Communication from the Commission to the council, the European
Parliament and the European Economic and Social Committee and the Committee of the Regions

A report to GSMA
55
1800MHz. Ofcom, the UK NRA, awarded five licences for third generation (3G) services in 2000 and 3G services were commercially launched in 2003. The re-farming of 900MHz spectrum and its use by UMTS services is considered as part of Ofcoms general spectrum management strategy in regards to 2G spectrum liberalisation. As an example, part of the band is in the process of being reallocated7. Ofcom has already indicated its intention to extend spectrum trading to all of the 2G and 3G bands in 2007. In 2005 Ofcom published the Spectrum Framework Review: Implementation Plan document, which sets out the options for the release of spectrum in 2005 - 2008 and extends spectrum trading and liberalisation to mobile services in order to allow for a smooth transition and efficient spectrum management8. Ofcoms key proposals are: The removal of restrictions from licences that presently prevent the use of spectrum for the provision of mobile services, including 3G services and mobile services other than 3G The consideration of a transitional period before removing restrictions on the ability to use spectrum that is currently not used for mobile phone services to develop 3G services in the future, suggesting that this transitional period might last till 2007, in order to provide an appropriate notice period to the existing licensees The consideration of a potential extension of spectrum trading and liberalisation to the bands currently licensed for 2G and 3G mobile services. Restrictions on the provision of 3G services could only be removed where it is possible to do so under law and subject to interference constraints and international obligations.
Ofcom also identifies a range of considerations that need to be taken into account, and a range of alternative options. Main issues are: Technical constraints in relation to the removal of the restrictions The existence of European harmonisation measures affecting the use of the bands The circumstances of the 3G auction held in 2000. Through this auction the Government created a market structure for the provision of 3G services with five licences, of which the largest was reserved to a new entrant. It is possible that variations in the terms of the 2G licences held by existing 2G licensees
Ofcom published a consultation paper regarding the award of available spectrum: 872-
876MHz paired with 917-921MHz. Responses from stakeholders suggested that the band should be used for UMTS. Ofcom proposal is to auction the spectrum on a technology and application neutral basis as this will give the market the opportunity to make the decision as to which service the Spectrum Bands can be most successfully used for.
8
Ofcom (2005) Spectrum Framework Review: Implementation Plan

A report to GSMA
56
(for example to allow 3G services) might have a disadvantageous effect on the competitive position of this new entrant, compared to the other operators9 Potential constraints may arise also to the scope for new international obligations in the future. These might take the form of new measures that may have binding effect on the UK, and/or changes to existing measures that have new binding effects
Ofcom considers the following three Options to balance the considerations stated above, as part of its long term strategy. Option 1 is to defer the decision in relation to the extension of liberation to the existing 2G bands. Although this option might be an appropriate course of action, Ofcom considers also the associated risks with the potential benefits of an early liberalisation of 2G spectrum and the uncertainty imposed to the mobile industry being the main ones. Option 2 is to extend liberalisation by removing restrictions on existing licensees use of 2G spectrum. Alternatively, this option would be followed under one or none of the following conditions: Delay date of liberalisation of 2G spectrum in order to minimize any adverse effect on competition Impose an additional payment on the existing 2G licensees to counteract any adverse effect on competition Define other pre-conditions prior to the 2G spectrum liberalisation
Option 3 is to extend liberalisation by issuing new licences. In this case, the liberalised 2G bands would typically be used to provide 2G services in addition to other services than 2G. Under the extended licences the spectrum users would not be able to make immediate use of these licences unless the incumbent 2G licensee agrees for this use. Based on these options, Ofcom will be able to allow a selective liberalisation according to market demands, while at the same time will allow some time for future developments.
This is especially the case for Hi3G. In regards to the Spectrum Management Review, Three
UK commented that a different spectrum policy that would allow further 3G capacity to be made available would alter the original spectrum offering conditions, thereby eventually affecting 3G future business plans. Thus it is possible that if in the future some spectrum that can be used to provide substitute products is allocated using administrative pricing, then there is a significant danger that 3G operators who got a licence through a highly competitive auction process could be discriminated from certain providers who would have a lower cost base as a result of government policy, rather than as a result of efficiency.

A report to GSMA
57
Ofcoms next step is to conduct further economic and technical analysis of the issues raised by the possible extension of spectrum liberalisation to the existing 2G bands. Ofcom has not published a final decision on the issue, but it is believed that Ofcom will strongly consider the economic benefits of a liberalised regime for the use of spectrum and hence allow spectrum users i.e. operators to decide on the usage of 2G spectrum. The situation in France There are currently 3 mobile operators in France. GSM licences were allocated as follows: Orange and SFR got their licence in March 1991 for a period of 15 years their initial licences expired in March 2006 Bouygues Telecom got its licence in December 1994 for a period of 15 years its licence expires in December 2009.
The reallocation of the GSM licences in France is simply due to the fact that the two first 15-year licences allocated to mobile operators expired in March 2006. The conditions of renewal of the GSM licences have to be agreed 2 years before the expiration date. As a consequence, renewal conditions had to be decided by March 2004. Arcep, the French NRA, published a consultation paper which deals with the renewal of Bouygues Telecoms GSM authorisation in the 900 and 1800MHz bands, which expires on 2009. This is a similar process followed to the consultation held in 2003 when the licences of Orange France and SFR were renewed. In October 2006 Arcep launched a second consultation paper which introduces the re-utilisation of spectrum in the 900MHz and 1800MHz bands for the deployment and development of 3G networks. Orange France and SFRs GSM authorisations include the possible reuse of the 900MHz bands for UMTS. Both operators have notified Arcep of their desire to take advantage of this possibility. Arceps main considerations, as presented in the consultation paper are: UMTS coverage and capacity requirements In terms of UMTS deployment, SFR and Orange are required to provide 70% population coverage by the end of 2007 and 2008 respectively, which could result in a improved adoption of innovative and multimedia services. Arcep considers the wide-area roll-out of 3G as a future goal and important mission. If must be noted that with high population density in urban and sub-urban areas, 70% population coverage in France can be achieved with around 30% of the geographic territory being covered. The better propagation and building penetration characteristics of UMTS900 will allow increased geographic coverage to be achieved and provides an

A report to GSMA
58
opportunity for operators to increase their population and geographic coverage beyond the position in early 2007. Arcep considers also that the quality of UMTS services cannot be guaranteed by using only the 2GHz band, with currently has a maximum spectrum availability of 2x60 MHz. Consequently, the 2x110MHz GSM allocations will be progressively migrated to UMTS spectrum. The introduction of 900MHz and 1800MHz spectrum areas will partially cover the capacity demands. Market players From the industry point of view, market players consider that UMTS900 has been standardised during 2005, which would lead to the necessity of the provision of new equipment and terminals. European framework Arcep is also considering the European and International developments in regards to the utilisation of 900MHz and 1800MHz bands. In European level there are processes in progress regarding the formulation of the regulatory framework for the deployment of 3G networks within the 900MHz and 1800MHz. This is highlighted with the technical studies which are also in progress. Arcep will use the results of these studies for the determination of the frequency distribution among the different operators. Taking into consideration the issues above, Arcep proposes: The reutilisation of 900MHz and 1800MHz to be implemented in phases and not in a big-bang approach. There will be a transition phase where GSM and UMTS will coexist in the same band. This way, GSM will ensure the provision of global services even in the case where international users wont be able to access the UMTS network or compatible terminals are not available. The equitable treatment of all the operators of 2G and 3G networks via the redistribution of the frequency bands. The authority will modify the authorisations of the utilisation of the frequency bands of all the parties involved. Hence, the consultation aims to determine whether the distribution of the 900 and 1800MHz bands need to be redefined in order to guarantee the frequencies are distributed equitably among all 2G and 3G mobile network operators. Arcep has stated its intent to survey the market again to assess interest in a fourth UMTS licence. If there is no interest for the fourth 3G licence, the scheme will be defined based on the existing conditions. Any entitled operator can demand the total or partial reutilisation of the bands which have been provided to this operator in order to develop 3G networks. This is in accordance to Arceps decisions early this year in regards to spectrum authorisations for utilising the frequency bands of 900 and 1800MHz. As stated above, SFR and Orange have already expressed their desire to put this demand into practice.
The authority, based on the results of the discussion will put in place a reutilisation scheme for the bands 900MHz and 1800MHz for 3G networks in 2007-2008.

A report to GSMA
59
The situation in Finland There are three main mobile (GSM/UMTS) operators in Finland. All operators have been allocated spectrum from both the GSM-900 and the E-GSM bands. The Ficora (Finnish NRA) frequency plan indicates that there are frequency allocations within the GSM-900 band which expire by the end of 2006. In general, spectrum licences are not provided in a technology-neutral approach, although Ficora is moving towards this approach for mobile services. Ficora initially planned to allow the GSM-900 band to be allocated for the deployment of UMTS network by early 2009. However, the Finnish operators have requested to be able to deploy UMTS900 networks in an earlier time frame and in May 2006, the Finnish Ministry of Transportation and Communications officially decided to allow the deployment of UMTS networks in the 900MHz band. Further to this decision, Ficora is currently reviewing the frequency allocations and it will publish a revised plan by the end of 2006, allowing the three operators to utilise the 900MHz band for UMTS networks. One of the key issues in the Finnish market is the fact that all three operators have already met the coverage requirements of their 3G/UMTS operating licences. The operators therefore did not object to the change of marketing conditions under which the UMTS licences were granted. One of the issues that needs to be addressed is the re-calculation of the new spectrum fee, although there are no plans by Ficora to apply a levy for allowing operators to utilise the GSM-900 band for UMTS networks. A second issue for the deployment of UMTS900 is the handset and equipment availability. Finnish operators expressed their desire to utilise the 900MHz spectrum as soon as the equipment becomes available. Finnish manufacturers indicated that the network equipment will become available by the last quarter of 2007. The situation in Germany There are four mobile network operators in Germany. T-mobile (D1 network) and Vodafone (D2 network) were assigned frequencies from the 900MHz band. E-Plus (E1 network) was licensed in 1994 and was assigned frequencies from the 1800MHz band. In 1997, O2 was granted a licence (E2 network) as the fourth GSM mobile operator and was assigned frequencies from the 1800MHz band. Different expiration dates apply for the operators licences. This is due to the different market entry dates, the variation of the licence period and the different legal frameworks applicable at the respective time. The D-network licences (Tmobile and Vodafone) expire on 31 December 2009, the licence of E-Plus Mobilfunk GmbH expires on 31 December 2012 and the O2 licence will end on 31 December 2016. According to FNA, the German NRA, there is a demand for GSM spectrum in the 900MHz band in the German GSM mobile market, especially for E network operators (E-Plus and O2), which, if it can be satisfied, will be a precedent for the

A report to GSMA
60
future development of the GSM and the UMTS market towards GSM and UMTS markets convergence. GSM and UMTS services are (still) deemed to be as separate markets today. However, due to the demand for 900MHz spectrum, preparations have started in order to deploy frequency bands currently allocated to GSM, as extension bands for mobile services of the third and later generations services. These processes will eventually result in creating one single market for IMT-2000 technologies such as UMTS in future. The regulator, in view of the foreseeable future convergence of GSM and UMTS services and the principle of technological neutrality, has recently provided a total of 2x10 MHz so as to allow reasonable use from the technical, regulatory and competitive perspective. GSM frequency bands will in future also be used on the basis of wideband technologies such as UMTS/IMT-2000. Future wideband technology usages will be subject to 5MHz blocks. For this reason the regulator divided the recently allocated frequency range (880-890/925-935MHz) into 2x5 MHz blocks each (upper and lower band) and to provide spectrum in 5 MHz blocks accordingly10. The use of frequency assignments from the 880890/925935 MHz bands will initially be limited to 31 December 2016 in connection with the adjustment of the regulatory framework for GSM usages. According to the German UMTS/IMT-2000 strategy11, frequencies for UMTS/IMT2000 will be allocated basically only in those bands identified by a WRC (global harmonisation)12. The situation in Sweden Although some countries e.g. France and Finland, have publicly expressed plans in favour of UMTS900 deployments, Sweden has no plans to address the issue in the short-term. Sweden has three main GSM operators and one UMTS greenfield operator and 900MHz licences have been granted with limitations for GSM usage. Although the Swedish regulatory authority (PTS) is in favour of a technology-neutral policy, the mobile network licences are technology specific. The operators are charged for the
10
FNA - Consultation for the award of further spectrum for public digital cellular mobile
communications below 1.9GHz (GSM concept)

11 12
FNA - Strategic Aspects of the Spectrum Regulation WARC-92 identified bands of 1885-2025MHz and 2110-2200MHz (core bands) as
the initial bands for IMT-2000 development WRC-2000 identified three additional bands for terrestrial IMT-2000 i.e. 2500-2690MHz, 806-960MHz and 1710 1885MHz

A report to GSMA
61
use of spectrum, however, the regulator is considering the adaptation of an auction policy for future spectrum licences. The situation in Italy In Italy, the current GSM-900 licences do not allow operators to deploy UMTS networks utilising the 900MHz band. Mobile operators like TIM and Vodafone have not expressed any request yet to AGCOM, the Italian regulator, to use their 900MHz licences for UMTS technologies. At the end of December 2005 TIM freed some spectrum in 900MHz band that was previously used to supply analogue telecommunications. This freed spectrum, which equates to 2x11.8MHz (880.0891.8/925.0-936.8MHz) is due to be auctioned during the course of 2006.
5.3.2 Asia-Pacific
In the Asia-Pacific region, some regulatory and spectrum management agencies have already adopted a technology-neutral approach to spectrum allocations. This is the case for Australia, New Zealand and Hong-Kong, although there are differences in how the spectrum is acquired. For example, in New Zealand spectrum can be traded by operators in a manner that best serves their needs providing the usage does not cause interference in adjacent bands, while in Australia, a licensee holding an apparatus licence may authorise, by written instrument, other persons to operate radio-communications devices under the licence. The situation in New Zealand The spectrum policy in New Zealand is, in principle, technologically neutral. The use of spectrum is mainly decided by the market, which is generally considered to be in a better position than government and regulatory authorities to make decisions on technical innovation. The operator which has been authorised the management rights over a spectrum band, decides the technology to be used in that spectrum band, under the condition that it does not cause interference in adjacent bands. The Spectrum Management Unit (under the Ministry of Economic Development (MED)) seeks to apply technological neutrality as far as is practicable. The General User Licence (GUL) allows for flexible arrangements in regards to spectrum allocations. The operators, under the current spectrum management regime, are able to migrate from legacy technologies. New technologies are partly or wholly self-managing, so that operators can access frequency bands without the need of an individual licence. It is the operators responsibility to ensure that their transmissions do not cause harmful interference for other users. New Zealand has an advantage which supports technology-neutral spectrum policies. Due to its physical isolation from neighbouring countries, much of the spectrum can be used without having to coordinate with neighbours over potential

A report to GSMA
62
interference. For this reason, New Zealand is considered to be an ideal country in which to test and develop new wireless technology products, like UMTS900. In a document which outlines future trends in spectrum usage in New Zealand13, the MED anticipates significant technology changes within the existing 800MHz/ 900MHz cellular bands which will allow future developments with 3G services in the 800MHz/ 900MHz bands, the main spectrum bands currently used for cellular services in New Zealand. The current holders of management rights to these bands, which will expire in 2011 and 2012 are Telecom New Zealand (TNZ) and Vodafone. TNZs management rights represent all of their spectrum holdings for cellular services in the 800MHz band. Vodafones rights for possible renewal account for approximately 66 percent of their holdings in the 900MHz band (its core network for 2G services). It also uses the 1.8GHz and 2.1GHz bands for 2G and 3G services, respectively14. A discussion paper on the renewal of management rights in the 800MHz and 900MHz bands, based on recent ITU work, released by The Ministry of Development (MED) in July 2006, states that the 800MHz and 900MHz bands will continue to be used for cellular services until at least 2025. These bands are currently used for 2G and 3G services and are candidate bands for 4G services, anticipated to be implemented from 2015. The document invites stakeholders views on such issues as the future of these bands, existing use of the spectrum, the appropriate renewal period, and the most appropriate pricing approach. In response to the paper, Vodafone expressed the view that the 800MHz and 900MHz spectrum in the next 10 to 20 years will be utilised not only by UMTS, but also by systems beyond UMTS and IMT-2000. The 800 and 900MHz bands have been identified for IMT-2000 (3G) services and are candidate bands for Systems beyond IMT-2000 (3G LTE). TelstraClears opinion is that both 800MHz and 900MHz bands in New Zealand will be used for 2G, 3G and, depending on decisions by the ITU, 4G services also. In Australia, TelstraClears parent company Telstra Corporation decided to deploy a UMTS-850 3G network, using the 3GPP standard for the 850MHz band. TelstraClear also uses this standard in the 2.1GHz band. TelstraClears preference is to deploy services in 800MHz/ 900MHz bands, if spectrum becomes available, because of the better propagation characteristics and therefore lower deployment costs. TelstraClear does not foresee that constraints, such as the availability of suitable handsets, could act as show-stoppers for the deployment of 3GSM GPP services in these bands. TNZs intension is to invest in upgrading its existing 850MHz CDMA network to meet new business and customer demands. TNZ is concerned that limiting the
13
MED Radio Spectrum: Future Spectrum Usage, A Forecast Of Technical Issues For The
Periods 2001-2006 & 2006-2011

14
MED Radio Spectrum Management: Renewal of Management Rights for Cellular Rights

A report to GSMA
63
850MHz spectrum availability would jeopardise their ability to offer higher data rates, since all of its existing 850MHz spectrum allocations are required to meet existing data forecasts in the short to near term. Therefore TNZ does not consider any of the 850MHz is or will be unused spectrum and they seek to renew the full 20MHz of 850MHz that they currently hold. In order to facilitate the development of competition in New Zealand there is currently 15MHz cap, set for an initial period of three years, on the provision of spectrum for 3G mobile services. The Ministry of Communications reviewed the spectrum cap policy in 2004 and the Government announced its decision to continue this policy until May 2007. Other issues that MED takes into consideration are: Technology changes that could affect the use of the 800MHz and 900MHz bands Industrys perspectives in terms of future investments Cellular coverage in remote/rural areas. Access to spectrum in the 850MHz and the 900MHz range is important to minimise site costs. The entire spectrum in this range is held either by TNZ or Vodafone; hence the unavailability of spectrum in the 850MHz/ 900MHz range for new entrants is a barrier to entry as it raises the cost of entry.
The situation in Hong-Kong In Hong-Kong cellular mobile services are provided in a variety of systems. There are eleven 2G licences, five of them operate in the 825-960MHz bands, based on the GSM-900 standard (three systems), IS-95 CDMA standard (one system) and IS-136 TDMA standard (one system). The remaining six licences operate in the 1710-1880 MHz bands providing Personal Communications Services (PCS) using the GSM-1800 standard. Before the expiration of these licences (July 2005 to September 2006), OFTA issued a consultation paper on the licensing of 2G mobile services. In this paper, OFTA states that the approach to be adopted regarding the licence renewal will be consistent with its established technology-neutral policy. According to the Hong Kong Third Generation Mobile Services Licensing Information Memorandum issued in July 2001, the existing 2G Operators are free to use any technology, regardless of whether it is 2G or 3G, in the spectrum under their 2G Licences. In line with this approach, existing 2G operators will be allowed to refarm their spectrum for 3G, if they so wish, under the current terms and conditions of their existing Licences for the remaining period of validity15.
15
OFTA (2004): Licensing of Mobile Services on Expiry of Existing Licences for Second
Generation Mobile Services

A report to GSMA
64
According to OFTA the licensees should make the best use of assigned spectrum with the most efficient technology available for the provision of more advanced mobile services. Nevertheless, the regulator is aware that 3G equipment is not yet available to operate in 2G spectrum. Furthermore, the TAs view is that the convergence of 2G and 3G services should be subject to a transition period, which would allow sufficient time for mobile operators to upgrade their networks for the provision of advanced mobile services. TA expects that by the end of this transition period, the gap between the capabilities of 2G and 3G networks will narrow down and the two networks will ultimately converge. Even if operators launch services of more advanced technology, such as 3G services, within the first 5 years of receiving a Mobile Carrier Licence (valid for 15 years), the same framework will continue to apply. The situation in Australia The spectrum licences in Australia are provided under a tradeable, technologyneutral policy for a fixed non-renewable term of fifteen (15) years. This means that the licence is not limited to any particular technology, system or service. The licence holder has the freedom to deploy devices anywhere within its licence area, provided that the devices are compatible with the core conditions of the licence and the technical framework for the bands. Within the licence area, licensees are free to operate whatever type of communications service they choose, and are able to change that service in response to technical improvements or changes in consumer demand. Some types of device must be registered with the regulatory authority before they can be operated. As spectrum licences are tradeable, licensees can negotiate in the open market with others to buy and sell spectrum space as the need arises, or authorise third parties to use their spectrum space. Spectrum licences can be combined or subdivided to form new licences. The frequency bands 825845 MHz paired with 870890 MHz are under the spectrum-licensing regime. Telstra recently announced an intention to re-farm this spectrum to provide UMTS services. The Australian Communications and Media Authority (ACMA) has allocated a large amount of spectrum for mobile telecommunications in bands compatible with both US and European technologies. The allocated bands are16: 825845 MHz paired with 870890 MHz (as used in the Americas) 890915 MHz paired with 935960 MHz (as used in Europe) 17101785 MHz paired with 18051880 MHz (as used in Europe)
16
ACMA (2004) - From DC to Daylight: Accounting for Use of the Spectrum in Australia A Spectrum
Management Strategy

A report to GSMA
65
19201980 MHz paired with 21102170 MHz (as used in Europe) and 19001920 MHz in major city areas
The spectrum bands 890915 MHz & 935960 MHz is under the apparatus licensing regime. These bands were assigned to three major telecommunication operators in 1992. All three operators have rolled out extensive mobile telephone networks using GSM technology. While initially providing only voice and low-rate data, upgrades via the GPRS and EDGE platforms have enabled average mobile data capabilities of 100 kbps. However, any equipment operation, like UMTS900, under an apparatus licence is subject to device-compliance conditions and the conditions specified in the licence by ACMA. The situation India The GSM operators in India currently have their operations in 900MHz and 1800MHz bands. The Telecom Regulatory Authority of India (TRAI) considers that since W-CDMA equipment is not currently available for 900MHz and 1800MHz bands (present GSM spectrum) and that there is no additional spectrum available in the 900MHz band, the GSM operators are not able to provide 3G services until they are allocated spectrum from the 2GHz area. Therefore, TRAI does not intend to allocate the GSM-900 band for 3G services. The Authority, upon receiving inputs from market players and operators, has recommended that the Government should not treat the allocation of 3G spectrum in the same way as 2G spectrum17. Hence, the allocations for 2G and 3G networks should be allocated separate bands. Under this approach, the Authority has identified the 450MHz, 800MHz, and 2.1GHz bands for immediate allocation for 3G services.
5.3.3 Arab/Middle East

The telecom monopoly of the incumbent operator in Qatar (Qtel) ended in November 2006 and the law was issued to enable the introduction of competition into Qatars telecommunications sector. The Qtel monopoly included mobile services and the legal and regulatory processes have started towards the liberalisation of the market. Qtel operates GSM services across the state and has launched 3G services in the 2.1GHz band. The regulator has reserved spectrum in which a second operator could deploy GSM services and intends to adopt a technology neutral approach in order to encourage innovation, allow better utilisation of the spectrum and speedup the deployment of convergence services. The policy will be based upon certain technical criteria in order to minimise interference not only with adjacent countries, but also among services within the same band.
17
TRAI (2006) Recommendations on Allocation and pricing of spectrum for 3G and broadband wireless
access service

A report to GSMA
66
In Egypt, the three GSM operators have been allocated spectrum from all GSM bands. The GSM-900 and E-GSM bands are fully occupied and only the 1800MHz band has unallocated/free spectrum. All GSM licences are valid for 15 years. Two of the licences expire in 2013, while the third has been renewed last year. Only one of the GSM operators has acquired a 3G/UMTS licence, under the condition of covering 75% of the land area within the next 4 years. The 2.1GHz band has been allocated for the provision of UMTS services. Although the regulator is in favour of a technology neutral approach (allowing the deployment of same generation networks in the same frequency band), the GSM/UMTS licences are technology specific. Although there have been no enquiries from the operators to deploy UMTS network in the 900MHz, the regulators main concern in allowing such development is spectrum availability in this band. 2G re-farming processes will be considered in the future to increase spectrum efficiency, under ITU decisions, but the regulator is concerned to protect GSM services in the 900MHz band. In Bahrain, there are two GSM operators occupying the full bandwidth of the EGSM and the GSM-900 bands. Both operators are in the process of launching 3G services using UMTS technology. The regulator has provided GSM licences under a technology specific policy, although technology neutral policies have been adopted, in general. The UMTS networks will be deployed using the core band of UMTS. Allocating spectrum from the GSM-900 band to UMTS services has not been considered yet by the regulator. The likelihood of such development will be addressed in the future under a general spectrum management policy according to ITU decisions, public consultations, internal coordination and international developments18. In Saudi Arabia, there are two GSM operators allocated spectrum from the 900MHz band. By the end of first quarter of 2007 a third licence will be issued. The remaining free spectrum of the 900MHz band will be assigned to the third operator. CITC, the national regulator, has adopted a technology neutral/service specific approach in regards to spectrum licensing. The licences are provided for specific type of services (e.g. mobile, fixed, broadcasting) and the licensee is allowed to deploy any technology that meets best his traffic demands. In other words, the mobile operators are able to migrate users from GSM technology to any other 3G technology according to their timeframe and business needs. CITC has coordinated its decisions with adjacent countries, although none of the operators has yet deployed UMTS services in the 900MHz. In UAE, there are two GSM operators occupying the full spectrum in the area of 900MHz and E-GSM. Both licences have been provided under a Public Sector Telecommunication Services regime which actually allows the operators to deploy the mobile technology of their preference within their allocated spectrum areas. Due to the slow uptake of UMTS the operators have only recently started to deploy
18
Telecommunication Regulatory Authority (TRA) (2006): Spectrum Policy and Planning - A
Joint Policy Issued by the TRA and the Ministry of Transportation (MoT) of the Kingdom of Bahrain

A report to GSMA
67
UMTS networks using the core band across the country. The regulator does not consider competition or licence pricing issues to act as barriers for UMTS900. However, the operators have to coordinate and inform the regulator before proceeding to UMTS900 deployment.
5.3.4 Sub-Saharan Africa

In Africa, 3G deployment is still immature and in most countries, operators have not yet applied for a 3G licence. The majority of countries have technology-specific policies for spectrum allocations for mobile networks. Although re-farming processes are being considered from some African NRAs, most of the 900MHz band spectrum has been allocated to GSM operators or has been reserved for future GSM applicants. In Kenya, two GSM mobile networks have been deployed. Both operators have been allocated frequency allocations from the 900MHz and 1800MHz bands. The EGSM band, although currently allocated for another use, could in the future be designated for GSM networks. The GSM licences have been granted under a technology specific policy. Kenyas frequency allocations allow portions of the 880960MHz band may in future be used to implement terrestrial component of IMT 200019. The operators have stated that they are planning to deploy 3G networks, although none of them has formalised this request to the regulator. Re-farming processes, as part of a migration plan from GSM to 3G/UMTS networks, are considered by the regulator as future spectrum management strategies. In South Africa, three GSM operators occupy the available spectrum in the E-GSM, GSM-900 and GSM-1800 bands. The licences are valid for 10 years and they have been renewed recently. Two of the operators have been granted licences to launch 3G networks, which operate in the 1900MHz and 2.1GHz bands. The regulator has adapted a technology neutral approach for spectrum allocations; however the second and third generation mobile networks do not operate under this policy. Potential applicants for 3G licences have submitted inquiries in regards to the availability of spectrum in the 800MHz band for the deployment of CDMA networks. This band is currently occupied by broadcasting services and the applicants questioned whether the CDMA services can coexist with the broadcasting services. The regulator will consider re-farming processes in the future as a means of spectrum relief, although issues like user migration and the cost implications also need to be considered. In Zambia, currently there are three GSM operators. All operators have allocated an equal amount of spectrum (6MHz) from the 900MHz and the 1800MHz bands. The E-GSM frequency band is unused. The regulator has reserved frequency bands in the 900MHz and 1800MHz area to allow a fourth operator to enter the market. Two of the operators have expressed their desire to deploy UMTS networks. Despite the fact that the spectrum management policy is under a technology
19
Communications Commission of Kenya (2006) Table of Radio Frequency Allocations

A report to GSMA
68
neutrality regime and the frequencies are allocated in a first-come-first-served basis, the operators choice is to perform initial testing for their UMTS networks in 1800MHz and 2.1GHz frequency bands. The first phase of the deployment is estimated for early 2007. In order to implement a technology neutral approach, the regulator has bilateral agreements with neighbouring countries. Border coordination is continuously revisited to allow an interference-free environment on each side of the border. In Zimbabwe, three GSM networks operate in the 900MHz and 1800MHz band. The E-GSM band is currently not is use, but this is due to handset unavailability. The GSM licences, which are technology specific, have been granted following a beauty contest process. The incumbent operator has expressed an interest to deploy a 3G network. The regulator is currently reviewing the request and the process will be formalised by the end of 2006. Frequencies will be allocated from the free area of 1700MHz band. Re-farming strategies will be considered in case better spectrum efficiency is required. In Ivory Coast, there are currently three GSM operators and the regulator has provided four additional GSM licences to start operating within 2007. Four licences have been provided for 20-years period, while three licences are valid for 10-years. The first GSM licence expires on 2016. The GSM licences occupy the whole spectrum of the GSM-900, E-GSM and GSM-1800 bands. Although the regulator is currently reviewing the introduction of a 3G/UMTS network in the 2.1GHz band, none of the current GSM operators has applied for a 3G licence. All GSM licences are technology specific. In Mauritius, there are two GSM operators, which have been allocated spectrum in all GSM bands. The regulator has adopted a technology neutral policy, although the frequency bands for the mobile networks are technology specific. Hence, the regulator has granted licences for the deployment of Public Land Mobile Networks (PLMN). These licences expire on 2015 and 2016 and are valid for 15 years. Since 2004, one of the operators has begun to deploy 3G/UMTS network in the core band of UMTS, with very low penetration rate mainly due to the expensive pricing of handsets. Although none of the operators have expressed any interest in the deployment of UMTS900 network, the regulator is in the process of assessing the implications of such development. Spectrum re-farming processes have been adopted in the past for the coexistence of UMTS together with other services. According to the regulator, competition issues wont be a barrier for UMTS900 deployment and no additional payment is being considered for granting such a licence. The regulators main concerns are technical and compatibility issues in order to minimise any interference or degradation may be caused to adjacent services. These barriers could be lifted under international harmonization processes, ITU and WRC decisions.

A report to GSMA
69
5.4 Time frame

There are many parties which need to be involved in re-farming of 900MHz spectrum: regulators, spectrum policy bodies, operators, service providers, chipset and handset vendors. Cross-border co-ordination is needed to assess and address potential interference problems. As noted in this report, it will be highly beneficial to the industry to have the widest possible commonality of approach to ease the introduction of UMTS900. Because of the long timescales which will be involved, it would be beneficial to the industry for the issue to be acted on as soon as possible, with public policy for the long-term management of GSM spectrum needing worldwide co-ordination. Currently, Europe accounts for 48.3%20 of WCDMA connections around the world. This fact together with the activities regarding UMTS900 networks development, not only from the regulatory/EU side but also from the mobile operators point of view, make Europe the most potential region for widely adopting UMTS900 technology sooner. No later than the end of 2007, three key players - UK, Finland and France -, will allow UMTS900 deployment. Considering these factors, an estimated timeframe could be as follows: by beginning 2008, 50% of the European NRAs allowing UMTS900 networks by beginning 2009, the rest will follow
In APAC region, the active WCDMA connections reach 47.8%22 of the number worldwide. Ten countries (Australia, Hong-Kong, Japan, S. Korea, Malaysia, New Zealand, Philippines, Seychelles, Singapore and Taiwan) in this region have deployed WCDMA networks. A technology neutral approach for mobile service seems a preferable option for most of the NRAs. However, in some of the countries the 900MHz. band is assigned to other services, since other systems like CDMA2000 provide an alternative 3G solution. Taking into consideration the above, an estimated time schedule could be as follows: by end of 2007, 20% (key players) of the APAC NRAs grading licences for UMTS900 by end of 2008, this percentage may reach 50% by end 2009, the rest 50% will follow
In Middle East, there are six countries with 3G/UMTS licences (Bahrain, Kuwait, Oman, Qatar, Saudi Arabia and UAE). In total, their 3G connections account for 0.8%22 of the worldwide WCDMA connections. NRAs tend to adapt a technologyspecific approach for mobile services and showing preference for the core-band of UMTS for 3G services. Due to neighbouring coordination that needs to take place, processes for UTMS licences may slow down. The next two years will be crucial for the UMTS uptake in this region; hence NRAs could aim for licencing UMTS900 networks by end 2008.
20
Wireless Intelligence

A report to GSMA
70
In sub-Saharan Africa, 3G/UMTS deployment is in an embryonic stage. Accounting for 0.5%22 of total WCDMA connections, UMTS networks are commercially launched only in Mauritius and South Africa. Second generation networks, mainly GSM, are the main revenue drivers in this region and NRAs aim to safeguard GSM operations. Coordination activities with neighbouring countries are extensive in this region; hence processes for UTMS licences may slow down. Taking also into account the very slow uptake of UMTS services, 10-20% of the NRAs wont be considering UMTS900 deployment sooner than the end of 2009.
5.5 Compatibility study

The re-farming of the 900MHz band implies that a part of the band gets used to run UMTS, while the other part gets still used to service users of GSM-only handsets (or dual-mode GSM-UMTS2100 handsets). Then there are two main implementation issues: One must take sure that adjacent GSM and UMTS channels dont interfere Operators must allocate appropriate spectrum capacity to GSM and UMTS, and must in particular make sure that enough capacity is left for GSM operations
In relation to the UMTS900 deployment in GSM frequency bands, ECC has prepared the following study reports: i) ii) UMTS and GSM compatibility operating in 900MHz and 1800 MHz bands UMTS compatibility operating in 900/1800 MHz with adjacent operating systems
A preliminary study by comparing the GSM and UMTS link budgets in 900MHz band has indicated that the cell range of GSM speech service is similar to that of UMTS CS64 (circuit switched 64 kbps). This means for a GSM900 operator, by re-using the GSM900 sites without adding any new sites, UMTS CS64 video-telephony service can be offered by co-locating GSM and UMTS sites. For the practical deployment of both GSM and UMTS in the 900MHz band a minimum of 2x10MHz per operator is needed21. According to the first report, the ECC report concludes that UMTS900/1800 can be deployed in different types of environments (urban, suburban and rural areas) in co-existence with UMTS and/or GSM under the following conditions: 1. UMTS900/1800 networks can co-exist with other UMTS900/1800 networks in the same geographical area with a carrier separation of 5 MHz. The recommended carrier separation between two uncoordinated UMTS networks is 5 MHz or more, while in a coordinated scenario is 5 MHz or even less. The UMTS capacity degradation is less than 1% for both uplink and downlink
21
ECC/CEPT (2006): Compatibility study for UMTS operating within the GSM-900 and GSM-
1800 frequency bands

A report to GSMA
71
2. UMTS900/1800 can be deployed in urban, sub-urban and rural areas in coexistence with GSM900/1800 macrocells in coordinated operation and/or in uncoordinated operation. According to the simulations, the UMTS downlink capacity loss varies between less than 1.5% (urban) and less than 1.2% (rural), while the uplink capacity loss is expected to be smaller than 5% (urban) or smaller than 3% (rural). The GSM system outage degradation is negligible. When UMTS900/1800 networks and GSM900/1800 networks are in uncoordinated operation, the recommended carrier separation between UMTS carrier frequency and the nearest GSM carrier frequency is 2.8MHz or more. When UMTS900/1800 networks and GSM-900/1800 networks are in coordinated operation (co-located sites), the recommended carrier separation between UMTS carrier frequency and the nearest GSM carrier frequency is 2.6 MHz or more. 3. UMTS900/1800 can be deployed in urban, sub-urban areas in co-existence with GSM-900/1800 microcell and/or picocell in uncoordinated (non-located sites between different networks) operation. In this case, the GSM system outage degradation is less than 1% for both downlink and uplink. The recommended carrier separation between the UMTS carrier frequency and the nearest GSM microcell and/or picocell carrier frequency is 2.8MHz or more. Further separation of the carriers will help to reduce the GSM micro/pico cell system outage degradation. 4. In order to avoid or minimise the interference between two operators, it is suggested for the operator who plans to deploy UMTS and GSM in the same band that it is better to use the so called Sandwich frequency arrangement as shown below.
Fig. 5.1 UMTS carrier arrangement between GSM carriers Sandwich arrangement
Source: ECC

A report to GSMA
72
The conclusions of the second report for the deployment of UMTS900 systems can be summarised as follow22: UMTS900 can be deployed in the same geographical area in co-existence with GSM-R as follows: 1. There is no need of an additional guard band between UMTS900 and GSM-R, a carrier separation of 2.8MHz or more between the UMTS900 carrier and the nearest GSM-R carrier is sufficient. 2. Activation of the GSM-R uplink power control is recommended, especially for the train mounted MS, otherwise the impact on UMTS UL capacity could be significant when the UMTS carrier is paced at 5MHz adjacent to the GSM-R band. However, it has to be recognized that this is only applicable in low speed areas as elsewhere the use of uplink control in GSM-R will cause significantly increased call drop out rates. For high speed trains the power control is not fast enough in order to allow for some interference gains. 3. UMTS operators should take care when deploying UMTS in the 900MHz band, where site engineering measures and/or better filtering capabilities may be needed in order to install UMTS sites on the railway track or close to the railway track when the UMTS network is using the 5 MHz channel adjacent to the GSM-R band. When UMTS900 is deployed in the same geographical area in co-existence with PMR/PAMR (CDMA PAMR, TETRA, TAPS) operating at frequencies above 915MHz, some potential interference from PAMR/PAMR BS to UMTS900 BS could be a problem. Some measures than could be taken into consideration are: reducing PMR/PAMR BS Tx power, space separation, installation of external filters, increasing Guard band. The potential interference from UMTS900 to aeronautical DME operating at frequencies above 970 MHz is not a problem. Some additional margins may be required for the protection of aeronautical DME operating at frequencies between 960 and 970MHz, where the required additional margins are dependent on DME carriers and aircraft positions.
22
ECC/CEPT (2006): Compatibility between UMTS 900/1800 and systems operating in
adjacent bands

A report to GSMA
73
6 Handset and chipset analysis

6.1 Introduction
Handset availability is crucial to the successful rollout of UMTS900 as the market will need to be populated with handsets which will work in the UMTS900 band and in other GSM and UMTS bands also. A reasonable threshold of UMTS900 handsets will needed in the market for it to make commercial sense for operators and end users to switch from one network to another. Research with chipset and handset vendors established the progress in chipset and handset development to date and outlined the roadmap and timeline for device development. Key findings 1. All major chipset (and consequently handset) manufacturers agreed that UMTS900 will happen. They are all either currently developing or planning to develop UMTS900 chipsets and handsets. 2. They all expect to develop chipsets supporting UMTS2100/900 (dual band) before moving towards quad band. The usual strategy is to include at least 2 upper bands and 1 lower band. A dual/tri/quad-band chipset refers to a chipset capable of operating in the corresponding number of radio frequency bands. The frequency bands can be configured by the use of onboard software to meet the different requirements of different regions. 3. In the case of UMTS900: dual band usually refers to UMTS2100/900; tri band usually refers to 2100/1900/900; quad band can refer to any combination with 2100/1900/900/xxx. All UMTS chipset manufacturers interviewed support GSM 1800/900 in their UMTS900 chipsets. 4. The volume of production is unpredictable as it depends on demand from operators. However, operators success in rolling out UMTS900 services lies in the availability of devices to support the transition to UMTS900. This was also the case in the transition from 2G to 3G. Vendors stressed that operators will need to create the pull to ensure that the UMTS900 market is seeded with handsets / devices. 5. Manufacturers expect that between a few hundred thousands and up to a few million handsets will be produced by 2009. With these low quantities, volume related costs will not come down significantly in the early stages of deployment. Advances in chipset design, e.g. higher levels of integration, will be the main driver for cost reduction rather than economy of scale. Furthermore, chipset manufacturers do not in general expect cost reductions to be necessarily reflected in prices. However, the opportunity for economies of scale reductions in handset prices will be severely limited if there is no common international policy to make spectrum available for UMTS900.

A report to GSMA
74
6. In low income countries, a minimum cost of ownership of a mobile phone is critical for increasing penetration or phones and retention of service. Key components of the cost are phone acquisition costs23 (which includes handset cost) and taxation24. Price elasticity is discussed in Section 7 of the report. Handset availability research was conducted through both desktop research and interviews with major handsets and chipsets manufacturers.
6.2 Chipset availability

Suitable chipsets have to be available before handsets can be developed on top of them. The roadmap of chipsets is generally consistent across all major chipset manufacturers which took part in interviews. They all expect to start the product line-up with chipsets supporting UMTS2100/900 (dual band) before moving towards quad band for easy roaming. Some are considering adding additional frequency bands such as 2.5 GHz which is yet to be approved by regulatory bodies. However the timelines for chipset development will vary and manufacturers are considered in two groups in the following market development scenario: Market Leaders and Followers. The edge that Market Leaders enjoy in the first two years will become less apparent in 2008 and the distinction between the two groups will blur.
6.2.1 Market development scenario

2006: Market Leaders: At least one major chipset manufacturer currently has dual band (UMTS2100/900) platforms developed, one with HSDPA and one without. One has development kits for their tri band platforms (UMTS2100/1900/900), thousands of which have been distributed to handset manufacturers for testing purposes. Followers: Development of UMTS900 chipsets is still underway because they are currently focusing on manufacturing chipsets supporting UMTS2100 or 1900, which are experiencing higher growth.
23
Economic Impact of Wireless and Policy implications in Asian developing countries. McKinsey, 3GSM
World Congress Asia 2006, Singapore, 16th October 2006

24
http://www.gsmworld.com/tax/

A report to GSMA
75
2007: As the market develops and more customer segments are established, so there will be different requirements for multi-band baseband chipsets. The total market for such chipsets is expected to be too small for economies of scale to develop. Market Leaders: By 2007, market leaders are expecting to have quad-band platforms available. HSDPA support will become as standard functionality and some will vendors expect to include HSUPA. The cost difference between dual band and other multi band chipsets is expected to be USD 5-10 per unit. This is mainly due to technological advances in chipset manufacturing processes and reduction in the number of components in each chipset made possibly by using new Systemon-Chip (SoC) technologies. Followers: Prototype development gradually transforms into commercial shipment and adoption by handset manufacturers. 2008: All manufacturers expect that demand for UMTS900 handsets will start to take off as operators start to develop the markets. However, volumes will remain low, and only a few million chipsets are expected to be shipped. Ovums research with chipset vendors indicated that did not expect to get significant economies of scale. However, some manufacturers are confident that by 2008 they will be able to integrate several different components into one and power management is expected to improve significantly. Costs are expected to decrease by 20-30% annually in the following years, driven by integration and manufacturing processes rather than volumes. A study by inCode estimates that chipset costs could be reduced by 58% in 2010 from 2007/8 levels25.
6.3 Handset availability

Handset manufacturers generally take about 3 months to develop handset prototypes from chipsets, and a further few months to ramp up production for commercial rollout. The time required for product and production development depends heavily on the handset complexity and size. Handset manufacturers are similarly divided into two groups, market leaders and followers. Most market leaders will have had the advantage of having developed UMTS 850 handsets. Once again, the distinction between leaders and followers will blur by 2008. Manufacturers stressed that the demand for handsets will have to come from operators to seed the market. The handset market is now much more mature than when the initial 3G networks were deployed and take-up was slow due to
25
UMTS900 Evolution Handset Roadmap and Penetration Analysis

A report to GSMA
76
disappointing and expensive handsets and poor network coverage. This maturity and the availability of UMTS 850 devices should allow a much cleaner market entry for UMTS900, but market pull is still needed.
6.3.1 Market development scenario

2006: Market leaders and followers: Given the availability of chipsets from chipset market leaders, prototypes are still under development. All vendors are still focusing on devices supporting other spectrum bands, with UMTS2100 and 1900 experiencing the strongest growth and hence placing significant demand on the product development teams. 2007: Market leaders: market leading manufacturers, with UMTS 850 experience consider the technology development for UMTS900 not to be a major issue it is simply a change in radio frequency. These manufacturers expect to be able to roll out dual band handsets by the second half of 2007. Although customer migration to UMTS900 is expected to take place, demand from operators is expected to be negligibly small compared to devices supporting other spectrum given a small number of operators actually having the infrastructure in operations. Initial demand of at least hundreds of thousands will be needed to make mass production commercially viable. The volumes will be insufficient to allow economies of scale to be achieved. Since UMTS900 has better coverage than UMTS in higher bands, some handset manufacturers expect that the market for such handsets will be mainly for voice traffic in the early stage, and hence mid or low end handsets are more likely to gain popularity. Therefore, higher volumes of handset shipments will be needed to compensate for the lower profit margins on low value devices. compared to high end devices. Followers: UMTS2100 and 1900 are expected to continue to dominate growth areas. Manufacturers are still cautious about demand from mobile operators because of different regulatory environments in different markets. 2008: Market leaders and other followers: The handset range will start to diversify and address to both high-end and middle range users, given the availability of a wider range of chipsets. Orders from operators are expected to start to build, but economies of scale will only start to develop once more widespread adoption takes place. However, manufacturers generally do not expect their reduction in costs to be reflected in prices.

A report to GSMA
77
7 Economic analysis
This section of the report provides a high level analysis of some of the economic factors which may influence the deployment of UMTS900 networks and services.
7.1 Impacts of the introduction of UMTS900 on prices and demand

It has been demonstrated through our modelling simulations that UMTS900 can effectively generate cost reductions of up to 40% in CAPEX and 30% in overall costs when compared to a baseline case scenario of deployment using UMTS at 2100MHz. These lower costs are primarily due to the radio propagation characteristics in the lower band which provide greater reach of UMTS900 and improved in-building coverage. Our modelling simulations did not include a comparison between UMTS900 and GSM technologies. Intuitively, however, given similar costs for coverage build up at 900MHz, overall costs are not likely to be different between these two technologies. If this is the case, then in a greenfield implementation case (areas without existing mobile coverage), provided adequate spectrum is available, operators are likely to choose the UMTS900 technology over GSM because of the potential upside in revenues due to high value add data services and a desire to align their networks to the 3GPP roadmap. In this section of the report we discuss the effects that the introduction of UMTS900 might have on the price of services and on the demand for services. There is however uncertainty about the relationship between these factors in new markets.
7.1.1 Impact of UMTS900 on the price of services

It is unlikely that operators will pass on savings with CAPEX and OPEX to customers immediately, especially in markets where overhauling of existing infrastructure is needed. The major reasons for operators not to reduce their prices will be: increased overall indebtedness - as capital expenditure for new infrastructure is usually financed by debt shareholder pressure to realise higher returns to compensate for recent capital outlays demanded by the overhaul
In a market with high mobile penetration and where operators deploy more than one technology (e.g. GSM and UMTS in multiple bands), operators do not have

A report to GSMA
78
incentive to immediately reduce the price of services for one specific technology (UMTS900 in this case) because this may impact their ability to charge higher prices for similar services deployed over technologies with a higher cost base. Some players, however, may decide to reduce investments in GSM or UMTS2100 and expand coverage and or capacity mainly with UMTS900. Their pricing strategy might include more aggressive price cuts to stimulate subscriber acquisition. In markets where mobile penetration is still low, it is more likely that the cost reductions delivered by UMTS900 are more quickly passed on to the end-user to accelerate penetration. There are other cases where the cost advantages of UMTS900 may be more readily translated into reduction of retail prices of mobile services. These include markets in which the regulator imposes price control based on regular cost reviews and cases where competition in the market is strongly price-focused. In the medium term, as investments in the GSM network have already been amortised and the declining contributions of the GSM platform represent a residual net present value which is low when compared with the potential NPV with an UMTS only network, we might expect a more accelerated replacement of GSM by UMTS900. When this happens, prices will tend to change reflecting the lower long run costs of UMTS. When replacing GSM by UMTS900, there is also a case for cross-subsidisation of voice and data services by an operator. As data services enabled by 3G enjoy higher margins, operators might opt to offer bundles with very low effective price per minute for voice to attract GSM subscribers from other networks. When operators are introducing UMTS900, they may have to provide additional handset subsidies to cover the higher costs of the lower volume UMTS900 handsets. These costs have not been considered in the analysis, as any such handsets will be dual or triple mode and be able to generate revenues on the operators GSM or UMTS2100 network, which offset the subsidy.
7.1.2 Impact of UMTS900 on the demand for services

We look at the impact of UMTS900 on the demand for services from three different perspectives: Elasticity of demand to changes in price Expansion of the addressable market Effect of tariffs on the service usage
Elasticity of demand to changes in price Assuming that cost differentials of UMTS900 are eventually translated into price reductions, these price reductions may drive an increase in usage. Although the general idea is very intuitive, its quantification and forecasting are complex. Demand elasticity to price is a concept widely accepted but rarely demonstrated.

A report to GSMA
79
The analysis of demand elasticity to price at a micro level requires marking a base of users (to monitor changes of usage after price alterations) and isolating external effects (e.g. seasonality, public holidays or major events, competition-driven bundling of services) that also interfere in the usage of mobile services. The results obtained are relevant to the segment of users analysed and remain valid for a short period of time. These results can hardly ever be generalised. On a macro level, some results can be more easily obtained but their relevance is more questionable as many other factors might also contribute to the change in usage (notably the effects of fixed to mobile substitution in the increase in usage). Figure 7.1 plots minutes of use of mobile voice services for different levels of price per minute using historic data from the UK market. Figure 7.1 Average revenue per minute vs. average minutes of use per week (per user) for voice
0.29 Price per Minute (USD) 0.27 0.25 0.23 0.21 0.19 0.17 0.15 110 120 130 140 Minutes of Use 150 160 170 R2 = 0.9237
Source: Ovum, based on Wireless Intelligence data The graph shows that a 10% change in the average price per minute for voice calls results in approximately 12% change in usage (expressed as minutes of use per week). However, one should note that not all of the increase in usage is due to price reductions and that part of it is due to increased usage of mobile services due to fixed-mobile substitution and also a general increase in traffic levels. It should also be noted that there are major differences in elasticity of demand in relation to price in different regions. Using data available from Wireless Intelligence, we have compiled the evolution of minutes of use and average price per minute for mobile voice services in Western Europe and in Asia Pacific.

A report to GSMA
80
Figure 7.2 Demand elasticity to price for mobile voice services in Asia Pacific and in Western Europe ASIA PACIFIC
0.2 Price per Minute (USD) Price per Minute (USD) 0.3
WESTERN EUROPE
0.15
0.25
0.1 R2 = 0.8519 0.05 190 210 230 250 270 290
0.2 R2 = 0.7422 0.15 110
130 Minutes of Use
150
170
Minutes of Use
Source: Ovum, based on Wireless Intelligence Data These graphs show that a reduction of price of USD 0.05 in Asia Pacific will in average produce an increase of 60 minutes of use in a week while in Western Europe, for the same decrease in price, there is approximately 35 minutes increase in use of mobile voice services. These differences are partially explained by the average level of development and mobile penetration in these regions. Figure exemplifies the expected differences between developed and developing countries. Figure 7.3 Demand elasticity to price in developed and developing countries
P
Average Price per Minute
P1 P2
Demand for Mobile C ommunication Services in Developed countrie
P3
Demand in Developing countries
Minutes of Use
Source: Ovum
Since the potential customer base in developing countries is much larger and more price sensitive, it is safe to assume higher price elasticity to demand, which implies a bigger incentive for operators in developing countries to lower their prices compared to counterparts in developed countries.

A report to GSMA
81
In the graphical example, Figure , prices are reduced from P1 to P2 in developed countries while a bigger reduction from P1 to p3 happens in developing countries. The situation is such that operators in developed countries clearly have less incentive to lower prices with its relatively inelastic and saturated mobile market, compared to an extremely elastic demand in developing countries. The price reduction from P1 to P3 in developing markets may also reflect a land-grabbing behaviour of operators trying to secure a higher market share (possibly even with temporary offers priced below cost). From a cost perspective, operators in developing countries usually benefit from a much lower customer acquisition cost, and also lower costs for electricity, rent and labour. In the GSMA report Tax and the digital divide it has been estimated that the price elasticity is between -0.54 (post-paid) and -0.76 (pre-paid) in developing countries. This means that a 10% increase in prices causes between 5.4% and 7.6% decline in usage. From Figure 7.2 we observe that the price elasticity for Asia Pacific using blended post-paid and pre-paid data is -0.57 which is compatible with these estimations. It is difficult, at this stage, to predict by how much mobile voice prices will be affected with the introduction of UMTS900 in regions such as the sub-Saharan Africa. However, one could reasonably assume that a 10% decline in prices will boost mobile usage (or the average minutes of use) by at least 6%. Expansion of the addressable market One valuable characteristic of UMTS900 is that it makes it possible to extend mobile coverage in developing countries. In Africa, for instance, the lower overall costs of UMTS900 (in relation to UMTS2100) mean that it may be economically viable to deploy 3G services in suburban or rural areas where this was not possible before. The practical consequences are that the addressable market is enlarged and the penetration of mobile services accelerates. However, it is important to note that in some developing countries, the longer handset replacement times (up to 4 years in some cases) and the prevalence of second hand handsets mean the uptake curve is shifted one or two years in relation to more developed markets. Initiatives such as GSMAs Emerging Market Handsets are already addressing this issue at the 2G level and will need to be extended to 3G in order to help accelerate the penetration of UMTS900 capable handsets in markets such as the sub-Saharan Africa.

A report to GSMA
82
Figure 7.4 Acceleration of penetration of mobile services
A ddressable M arket after UM T S 9 0 0
Penetration
A ddressable M a rk e t befo re UM T S 9 0 0
Time
Source: Ovum theoretical model Effect of tariffs on the service usage In developing countries, the prevalence of pre-paid plans can distort the effect of demand elasticity to price. This happens because in low-income countries, the barrier to usage is commonly the value of the recharge card, independent of the number of minutes bundled in it. Therefore, if savings obtained from lower operational costs are passed on to the price per minute and operators are able to include more minutes in the bundles or reduce the price of existing bundles, this will expand the usage of services. In developed countries, the broader availability of UMTS is speeding up the launch of infotainment services combined with flat-fee style tariffs. As these new data tariffs and compelling services are made available to end users, the average profile of usage of data services will change. While voice pricing will remain the main flex-factor affecting 3G migration and growth for the next few years, flat-rate data will become an increasingly effective tool by which pure-play UMTS and second or third ranked mobile operators in developed countries can improve their position. There are several effects that take place in a mobile network in few months after the successful introduction of flat-rate tariffs: Heavy mobile data users are attracted to the network Existing users start to gradually increase usage of data services, initially mostly based on web browsing

A report to GSMA
83
Users are encouraged to experiment new data services such as VoIP and video streaming using mobile data cards or applications downloadable to smartphone handsets
The introduction of UMTS900 and consequent broader availability of 3G services is likely to speed up the transition to flat-rate and bundle tariffs which will in turn cause a substantial increase in usage of data services.
7.1.3 Impact of UMTS900 on the demand for services in various regions

A detailed analysis of the impacts of UMTS900 on the demand, as considered in our simulations, is provided in Section 4 of this report.
7.2 Impacts to country economics

UMTS900 is an ICT technology which provides a very good opportunity to produce economic benefits. To the extent that it can reduce prices and extend the supply of mobile services to new areas, UMTS900 can help to unlock considerable value that can subsequently revert to new investments. There are many other ways in which UMTS900 can bring economic benefits. These are illustrated by examining how it might contribute to economic growth. The theory of economic growth identifies several channels through which ICT can influence growth.
7.2.1 Changes in ICT related capital investments

There is scope for UMTS900 investment in all markets: In countries where mobile coverage is already high, capital investment is likely to be made for replacement of 2G systems by UMTS900 In countries where coverage is not complete in suburban and rural areas, capital investments may be directed to extensions of coverage using UMTS900.

A report to GSMA
84
Figure 7.5 Economic benefit of lower prices and expanded supply of mobile services
Price
Demand
Price reduction passed on to users P1 P2 Economic Benefit Supply
Lower cost base of UMTS 900 allows expansion of supply
Quantity
Figure 7.6 Components of the Economic Growth Accounting
7.2.2 Changes in the size of the labour force

The wider availability of mobile services, enabled by the UMTS900 technology, in developing countries can improve the chances of the unemployed to find employment. There are several ways by which this may occur: enabling unemployed people to call for opportunities rather than rely on word of mouth communications and avoiding unnecessary journeys

A report to GSMA
85
taking advantages of UMTS900 data capabilities to access the Internet to seek for job opportunities or to receive updates on job vacancies having access to a mobile phone may improve employability. An employee who has the means to receive and make calls will be more productive in the eyes of the employer as they will be easier to contact when away from their location of employment improved mobile coverage in the region will stimulate SME business development and growth.
The increased expenditure for the build-up of the UMTS900 infrastructure will also help to create other jobs. For instance, employees of the mobile industry spend money on restaurants, holidays, food etc and generate additional jobs as a result. Previous GSMA reports26,27 estimate that a multiplier between 1.5 and 1.7 for developed countries and 2.0 for developing countries exists, i.e. for every ten direct jobs created in the mobile industry by the introduction of UMTS900, there will be other five to ten induced jobs created in other industries.
7.2.3 Changes in the productivity of labour

The positive effects that ICT investments have in productivity are referred to as capital deepening. In developing countries, the effect will be felt particularly for workers with unpredictable schedules. Examples would be any workers who are involved in repair and maintenance activity, plumbers, electricians and workers involved in collection and delivery of services. For such workers, access to UMTS900 services allows them greater accessibility and therefore superior knowledge of demand for their services. In developed countries, the data capabilities enabled by UMTS900 can be used to keep workers connected to companies back-office systems (e.g. e-mail, CRM, ERP) independently of their location or time of the day. Indeed, the productivity of all workers who, at any time or for any reason, have a need to be away from their main location of work, or who need to travel frequently can be enhanced by mobile communications.
26
Lewin, D: The economic contribution of mobile services in the Europe Union before its 2004
expansion
27
Lewin, D and Sweet, S: The economic benefits of mobile service in India

A report to GSMA
86
7.2.4 Solow28 residual

Economic theory developed over the last 20 years has identified other mechanisms to recognise benefits. Figure .7 decomposes the element of GDP growth for countries that are member of the G7 for the period 1995 to 2001. This analysis by Jorgensen shows that in addition to changes in Capital Input and changes in Labour Input, there is a residual, commonly named total factor productivity (TFP), which captures any growth that is not associated with input usage. Enabled by technological spillovers, phenomena such as creative deconstruction and network effects play an important role in recycling business practices promoting growth both in the ICT and in non-ICT sectors. To the extent that UMTS900 facilitates new working practices (e.g. by improving the mobility of the workforce) it will indirectly contribute to a creative deconstruction where firms may reengineer business processes rather than simply automating them so as to take advantage of the productivity gains that UMTS900 enables. Also, new firms, with new business models enabled by mobility, displace existing firms which operate legacy business models.
Figure 7.7 Growth components in G7 countries

Sources of Economic Grow th (G7 Countries 1995-2001) 5
Annual Contribution (%)
4 3 2 1 0 -1 -2 US Canada UK France Germany Italy Japan
Labour
Non-ICT Capital
ICT Capital
Non-ICT Total Factor Productivity
ICT Total Factor Productivity
Source: Dale Jorgenson (Harvard University - 2004) As 3G services become more widely available (with the help of UMTS900), network effects quick in29. With more companies connected to the mobile internet,
28
Named after Robert Solow, winner of the 1987 Nobel Prize for his work on the Exogenous
Growth Model

A report to GSMA
87
UMTS900 services can help reduce transaction costs by enabling companies and workers to trade and negotiate electronically. Mobile internet also helps to improve the dissemination of information to allow markets to operate more efficiently. This is particularly important in the financial market where banks benefit from having prompt access to financial information. Previous studies point to a strong positive relationship between investment in telecommunications and economic development30,31: A developing country which had an average of 10 more mobile phones per 100 population between 1996 and 2003 would have enjoyed [annual] per capita GDP growth that was 0.59 percent higher than an otherwise identical country. These studies however do not account for the combined voice and data capabilities of 3G communications. The extension of Internet to a much broader audience and its usage in completely different contexts of use will require an upward review of the existent estimates.
7.2.5 Regional differences

In Western Europe it is expected that economic benefits from increased investment in UMTS900 infrastructure or improvement in labour productivity from the direct use of UMTS900 technology will not be major factors when compared to the long term benefits that the diffusion of mobile broadband wireless access and mobile internet will have in reshaping business practices. In the Middle East there is still room for increase in the penetration of services and capital investment is one key economic benefit. However, as oil-rich countries look into diversifying to tourism and financing, UMTS900 can be also instrumental in improving labour productivity. In Asia Pacific there is a mixed picture. In countries with high population density, the economic benefit that deployment of UMTS900 can yield is to speed-up the substitution of 2G to 3G thus stimulating the reshape of business models. In
29
One consequence of a network effect is that the purchase of a good by one individual
indirectly benefits others who own the good - for example by purchasing a telephone a person makes other telephones more useful. This type of side-effect in a transaction is known as an externality in economics, and externalities arising from network effects are known as network externalities.
30
Rller, Lars-Hendrik and Leonard Waverman, Telecommunications Infrastructure and
Economic Development: A simultaneous Approach, American Economic Review, September 2001, pp 909-923 using OECD data.
31
Ndiri, M Ishaq and Banani Nandi, Telecommunications Infrastructure for Economic
Development, January 2003, NYU working paper on developing countries.

A report to GSMA
88
countries such as Australia and New Zealand, UMTS900 can also play an important role in bringing 3G to more remote areas improving economic activity of indigenous population and reducing transaction costs for farmers. In sub Saharan countries the case is clearly to bring prices down and to make mobile services accessible to a larger percentage of the population while also expanding the supply by extending the coverage to rural areas. Capital investment will influence growth in a higher degree that it does in other areas but UMTS900 has a much bigger role in helping people find employment or as an enabler of informal economic activities. Farmers could also have their bargaining power improved as they would have better access to price information of both inputs and outputs.
7.3 Adoption of complementary technologies or services

The benefits of UMTS900 to the economy are not restricted to the economic growth that increased voice-centred mobile service penetration and usage prompts. Because of its native broadband data capabilities the introduction of UMTS900 has a much broader impact in the ICT sector as a whole.
7.3.1 Third Generation Mobile Services

One direct consequence of the introduction of UMTS900 will be to speed up the penetration of 3G services in the different regions. As shown in Figure 7.3Figure 7.8, apart from Western Europe where 3G subscribers will account for mare than 60% of the mobile subscriber base by end of 2010, the forecasted penetration of 3G as a total of mobile subscribers is forecasted to stay relatively low with the second highest penetration below 15% in Asia Pacific.

A report to GSMA
89
Figure 7.38 Total Mobile connections and 3G connections to 2010

Africa 400 Subscribers (in 1,000,000 s) 350 300 250 200 150 100 50 0 Q3 2004 Q3 2005 Q3 2006 Q3 2007 Q3 2008 Q3 2009 Q3 2010 Total 3G Subscribers (in 1,000,000s) 1,800 1,600 1,400 1,200 1,000 800 600 400 200 0 Q3 2004 Q3 2005 Q3 2006 Q3 2007 Q3 2008 Q3 2009 Q3 2010 Total 3G Asia Pacific
Middle East 250 Subscribers (in 1,000,000s)
Western Europe 600 Subscribers (in 1,000,000s) 500 400 300 200 100 0 Total 3G
200 150 100 50 0 Q3 2004 Q3 2005 Q3 2006 Q3 2007 Q3 2008 Q3 2009 Q3 2010
Total 3G
Q3 2004
Q3 2005
Q3 2006
Q3 2007
Q3 2008
Q3 2009
Q3 2010
Source: Ovum, based on Wireless Intelligence data In the portrayed situation there are huge opportunities for UMTS900. Not only this technology can help to accelerate the penetration of mobile services as a whole but it also has the potential to increase considerably the participation of 3G as a percentage of the total mobile connections. Supposing that UMTS900 captures half of the mobile connections growth from 2010 and that the overall growth remains unchanged, penetration of 3G connections is impacted as follows:

A report to GSMA
90
Figure 7.9 Possible impact of UMTS900 in the penetration of 3G 3G penetration (% of total connections) without UMTS900 Africa Asia Pacific Middle East Western Europe 2.8% 12.9% 7.1% 61.3% 3G penetration (% of total connections) with UMTS900 8.2% 16.8% 11.5% 62.7%
7.3.2 Internet
UMTS900 will have two main effects in relation to Internet access. Firstly, it will allow more individual users to access content in the Internet using web-browse enabled handsets or data cards. The positive effects will be felt particularly in rural areas where, in many occasions in developing countries, Internet access is not available because there are no fixed lines or broadband service available. Secondly, it will provide one more channel for enterprises to keep its workforce connected to the Internet even when on the move. In developing countries, the broader availability of Internet access for enterprises may also stimulate that more companies decide to launch their own websites. All in all, UMTS900 can be instrumental in reducing the digital divide between developed and developing countries and urban and rural environments.
7.3.3 Broadband spread

In most cases the rollout of UMTS900 infrastructure will be based on release 5, release 6 or even release 7 of 3GPP. In practice, this means that the UMTS900 infrastructure will most likely be enabled for HSDPA (rel.5), HSUPA (rel.6) and HSPA+ (rel.7). The initial deployments of HSDPA demonstrate that this technology typically deliver average speeds of 1.8Mbps for content download. Comparatively: Average speed is twice or 3 times that of 3G in the real world An email with a 4Mb attachment can be downloaded in around 30s (1min30s with UMTS) Access to Intranet is possible in 4.5s (10s with UMTS) A web page of 150Kb appears in 3.5s on an HSDPA users screen instead of 8s with UMTS

A report to GSMA
91
Mobile Streaming of a TV channel encoded in MPEG-4 and streamed at 300kbps is enabled
Later in release 6 with HSUPA, high speeds will be enabled also for upload and with release 7, smart antennas and improvements such as Multiple Input and Multiple Output (MIMO) are likely to double or treble the average speeds available to users of UMTS. The increased coverage and enhanced in-house penetration achieved with the deployment of UMTS900, coupled with the broader availability of laptops and desktops with UMTS connectivity, can positively affect the rate of adoption of broadband services in a country.
7.3.4 Other complementary services

Linked to the improvements that UMTS900 will help to foster in relation to intensification of Internet usage and spread of broadband access, there are two other areas of complementary services that are worth mentioning: e-Commerce - Used as a broadband access service, UMTS offers one more channel for internet based e-Commerce. In addition, service enablers embedded in the mobile infrastructure make it possible to use mobile handsets to perform commercial transactions and payments. Public e-Services Services such as e-Government, e-Learning and e-Health can be made available as mobile applications in the UMTS network.
Provided the enhanced coverage and increased penetration of 3G services that UMTS900 can deliver, these services can be accessed in different contexts of use, for instance while commuting, and will add to elimination of dead time and to the improvement in individual productivity.
7.4 New services for customers

Less tangible but not least important is the influence that UMTS900 can have on the introduction of innovative communication services. Linked to the wider adoption of flat-rate and bundle tariffs, data based services are set to blossom. One such service is voice over IP over mobile. It allows users to make use of VoIP software clients to make calls, for instance, to international destinations without paying extra (e.g. mobile to computer) or at reduced prices. Besides the savings to end-users, indirect economic benefits can be achieved through increased international communications. As outlined before, investments in UTMS 900 are likely to be linked with mobile infrastructure compatible with releases 5 and 6 of the 3GPP standards. These releases introduce the IP Multimedia Subsystem which is a powerful enabler of rich services. These are services that combine presence capabilities (e.g. as in the case

A report to GSMA
92
of Instant Messaging), location capabilities and combinational capabilities (e.g. the ability to combine audio, text messages and/or video streaming during a communication session). The indirect economic benefit of these services is that they will facilitate to a great extent the conveyance of information and the consumption of content. IMS also plays a pivotal role in the convergence of networks. Besides the well known and extensively discussed economic benefits to mobile operators related to the lower operational costs and improved revenue potential delivered by converged networks, users also benefit from services such as FMC where the user is always best connected either via mobile or via fixed line communications and have a seamless experience through the use of a single handset. These are all services that are gradually developing with upgrades in the existent infrastructure. The deployment of UMTS900 will represent a chance to speed up considerably the diffusion of such services.

A report to GSMA
93
Annex A Model description

A.1 Introduction
The main benefit of UMTS deployment in the 900MHz band is the lower signal attenuation compare to 2.1 GHz band. This results in higher cell coverage in rural areas where there is less demand for services, and better quality of coverage in urban areas especially for indoor users due to better signal penetration. Apart from the cheaper deployment (smaller number of Node Bs for coverage driven network planning), it will also decrease the number of handovers for users with medium and high mobility. The 2.1 GHz band can be used to complement the 900MHz band in areas with high capacity demands to serve hotspots such as train stations, stadiums and tourist areas.
A.2 Network Dimensioning

We perform simple coverage-based dimensioning, simple capacity-based dimensioning and finally a simple combined coverage and capacity dimensioning. A simple approach that prevents an under-estimation of the required number of cells is to curry out the dimensioning exercise using both methods and selecting the higher prediction for the number of cells required. This is obviously a somewhat cruel method and a more accurate dimensioning exercise will consider the interaction between capacity and coverage in a UMTS system. For the coverage driven dimensioning we initially have a link budget calculation to estimate the maximum path loss. The path loss is then used to find the maximum range by using the well known Cost 231-Hata-Model propagation model with correction factors for suburban and rural environments. The cell range estimation is following based on the number of sectors used and finally the site number is calculated as a ratio of the total area over the site area for different geotypes. For the capacity driven dimensioning we initially calculate the maximum capacity per carrier based on the pole capacity and the assumed noise rise. The overall (voice, SMS and data) offered traffic in busy hour per geotype is then estimated. Finally the maximum number of sites per geotype is calculated as a ratio of the offered traffic over the site capacity (i.e. the carrier capacity times the carriers per site). The required number of node Bs that fulfill both network coverage and capacity requirements, for different geotypes, given their sector configuration is taken as the maximum number of the two independent calculations of node Bs. The process of evaluating the required number of node Bs is summarised in Error! Reference source not found.Figure A.1. A more detailed description of the propagation models, the link budget estimation, the sectorisation and the pole capacity estimation follows in the next sections.

A report to GSMA
94
Figure A.1 Process of calculating maximum number of Node-Bs required

Bandwidth Demand UMTS Connections Voice, SMS, Data Demand
Coverage Demand
Geotype Table ()
UMTS Coverage
System Parameters
Pole Capacity
Link Budget
Busy Hour Traffic Data
Cost 231 - Hata Model Transmission Range & Node B Coverage
Nodes B from coverage driven calculation
Nodes B from capacity driven calculation
Take Maximum
Source: Ovum Coverage Demand: The geotypes represent the morphological structure of the communication environment. Multiple geotypes can be defined, reaching very detailed description of the environment. Due to practical reasons, however, the number of geotypes should be limited for dimensioning purpose to obtain fast results. Our geographical area has been partitioned into 4 classes. An example of type of data used is provided in the following Figure A.2.
Figure A.2 Example of type of data used

A report to GSMA
95
Geotype
Area Coverage
Population Coverage
UMTS Coverage
UMTS Capable Handsets 80% 80% 60% 50% Voice 160% 160% 85% 70%
Traffic Share SMS 160% 160% 85% 70% Data 160% 160% 85% 70%
Dense Urban Urban Suburban Rural
2% 26% 8% 20% 70% 47% 27%
70% 70% 60% 50%
Source: Ovum
A.3 Propagation Models

In our calculations we use the Cost 231-Hata-Model Error! Reference source not found. for 900MHz and 2.1GHz frequency bands for urban environments and the corresponding correction factors for suburban and rural areas. 900MHz - Urban area
L( R) = 69.55 + 26.16 log( f ) 13.82 log(hB ) + [44.9 6.55 log(hB )]log( R) a (hM )
(1) where
a(hM ) = [1.1log( f ) 0.7]hM [1.56 log( f ) 0.8]

and hB is the base station (BS) antenna high in meters, hM is the user equipment (UE) antenna high in meters, f is the used frequency in MHz and R is the distance between BS and UE in km. For some typical values, i.e. hB = 30m , hM = 1.5m and f = 920 MHz , and so above formula is simplified to:
L( R ) = 35.2 log( R) + 126.6

2.1GHz - Urban area
L( R ) = 46.3 + 33.9 log( f ) 13.82 log(hB ) + [44.9 6.55 log(hB )]log( R) a(hM ) a( hM )
is given again by
Where
a(hM ) = [1.1log( f ) 0.7]hM [1.56 log( f ) 0.8]
For the same typical values as above and for f = 1950 MHz we have:
L ( R ) = 35 . 2 log( R ) + 137 . 4

A report to GSMA
96
Suburban environments
L(R )
suburban
= L ( R ) urban
2 log
f 5 .4 28
Therefore for
f = 920 MHz a typical value will be
L( R) suburban = 35.2 log( R) + 116.6 , and for f = 1950 MHz we have: L( R) suburban = 35.2 log( R) + 125.2 .
Rural environments
L( R ) rural = L( R ) urban 4.78 log 2 ( f ) + 18.33 log( f ) 40.94

Again, for and for
f = 920 MHz a typical value will be L( R) suburban = 35.2 log( R) + 98 ,
f = 1950 MHz we will have L( R) suburban = 35.2 log( R) + 105 .
A.4 Link Budget

An example of a link budget used based on 12.2 kbps voice service is shown at Figure A.3. It must be noted that input data provided is just illustrative and link budget has been built for all different scenarios and regions examined on a case specific basis. Figure A.3 Link Budget Calculation table for 12.2 kbps voice service
Values Transmitter (mobile) Max. mobile transmission power (W) As above in dBm Mobile antenna gain (dBi) Body loss (dB) Equivalent Isotropic Radiated Power (EIRP) (dBm) 2.1GHz 0.25 24.0 2.0 0.0 900MHz 0.25 24.0 2.0 0.0 a b c Formula
26.0
26.0
d=a+b-c
Receiver (base station)

A report to GSMA
97
Thermal noise density (dBm/Hz) Base station receiver noise figure (dB) Receiver noise density (dBm/Hz) Receiver noise power (dBm) Interference margin (dB) Total effective noise + interference (dBm) Processing gain (dB) Required Eb/No (dB) Receiver Sensitivity (dBm) Base station antenna gain (dBi) Cable loss in the base station (dB) Fast Fading margin (dB) Max. path loss (dB) Log-normal fading margin (dB) Soft handover gain (dB), multicell In-car loss (dB) Interference due to co-location with GSM 900MHz (dB) Allowed propagation loss for cell range (dB)
-174.0 5.0 -169.0 -103.2 3.0 -100.2 25.0 3.1 -122.0 18.0 2.0 4.0 160.0 7.0 3.0 10.0
-174.0 5.0 -169.0 -103.2 3.0 -100.2 25.0 3.1 -122.0 18.0 2.0 5.0 159.0 10.0 3.0 10.0
e f g=e+f h=g+10*log(3840000) i j=h+i k=10*log(3840/12.2) l m=l-k+j n o p q=d-m+n-o-p r s t
146.0
139.0
u=q-r+s-t-v
Source: Ovum
A.4.1 Link budget explanation

Brief description of all the values and formulas used in the link budget (Uplink case). Maximum Transmission Power: 3GPP defines four classes for the mobile terminal with the following nominal power and tolerance values [5]: 1. Power Class 1: 33dBm (2W), +1/-3dB 2. Power Class 2: 27dBm (0.5W), +1/-3dB

A report to GSMA
98
3. Power Class 3: 24dBm (0.25W), +1/-3dB 4. Power Class 4: 21dBm (0.125W), +2/-2dB We use Power Class 4 for voice (Power Class 3 for data). For the base station, 3GPP has defined upper limits for micro and pico base stations. This is 38 dBm for micro and 24 dBm for pico BS [6]. For the macro BS, there is no definition of an upper limit of output power; typically, it varies between 10 and 45 W ( 2.5 dB tolerance), with the step of 5W, depending on manufacturer and product line. Antenna Gain: The gain on a mobile antenna depends mostly on its size and the application of the UE. In a typical handset application, such as voice, it is assumed that the antenna gain is equal to 0 dBi. UE used as data cards (PCMCIA) can contain an antenna with 2 dBi gain. For the BS antenna the gain depends on the directionality of the antenna. e.g. for 65o horizontal beam width the gain is 18 dBi. Body Loss: A body loss parameter is introduced due to the usage of the UE near the users body. It defines the additional loss in the transmitting and receiving path. Typical values are 2-3 dB on average for voice services and 0dB for data. Receiver: The receiver sensitivity level is mainly limited by thermal noise. The thermal noise power spectral density N 0 , normalized to 1 Hz bandwidth is 174 dBm/Hz. This noise floor is further limited by the quality of internal components (Low Noise Amplifiers (LNA), filters, synthesizers, etc) which generates additive noise. This noise contribution is expressed as the receiver noise figure , F ,and has typical values of 5 dB. The total noise power at the receiver is limited by its filter bandwidth (for UMTS, the chip-matched filter bandwidth B equals to the frequency band occupied by the scrambling code, i.e. approximately 3.84MHz, which owing to spectral side lobes, results in 5MHz carrier raster); thus the total receiver noise power in dBm is given by:
N r (dBm) = N 0 (dBm / Hz ) + F (dB ) + 10 log(3.84 10 6 ) .

Interference margin: Interference margin is a function of the cell loading. The more loading is allowed in the system, the larger interference margin is needed and the coverage area shrinks. For coverage-limited cases a smaller interference margin is suggested, while in capacity-limited cases a larger interference margin should be used Error! Reference source not found.. In the coverage limited cases the cell size is limited by the maximum allowed path loss in the link budget, and the maximum air interface capacity of the cell is not used. Typical values for the interference margin in the coverage limited cases are 1.0-3.0 dB, corresponding to 20-50% loading.

A report to GSMA
99
Cable loss in the base station: Typically, the overall cable system attenuation should be less than 3 dB for macro base station. Processing gain: The ratio between the chip rate
Wc and
the user data rate
Wd
is called processing gain
G p . This is
3.84 10 6 G p (dB ) = 10 log W d
Required Eb/N0 : The typical Eb/N0 values for different services with 10-20 ms interleaving and BLER=10% are presented in Figure A.4 Summary of uplink Eb/NoFigure A.4 and Figure A.5 for the uplink and downlink respectively, as they have been derived in Error! Reference source not found.. Figure A.4 Summary of uplink Eb/No Voice 12.2. kbps Static Multipath 3km/h Multipath 120km/h Source: Ovum 2.9 dB 4.2 dB 5.5 dB Data 64 kbps 1.0 dB 2.2 dB 3.4 dB Data 144 kbps 0.4 dB 1.7 dB 2.9 dB Data 384 kbps 0.6 dB 2.0 dB 3.4 dB
Figure A.5 Summary of downlink Eb/No Voice 12.2. kbps Static Multipath 3km/h Multipath 120km/h Source: Ovum Fast Fading Margin: Fast fading margin or power control headroom is needed in the UE transmission power for maintaining adequate closed-loop fast power control in unfavorable propagation conditions such as near the cell edge. At the cell edge, the mobiles can transmit almost at the maximum available power; thus there is no more headroom to follow and reduce the negative influence of the small-scale or fast fading. In order to include this phenomenon in the dimensioning process, the fast fading margin has to be applied. The fast fading margin depends on the mobile speed and its typical value [1]: Varies between 4 and 5 dB for slowly moving mobiles (3 km/h) 4.4 dB 7.0 dB 7.0 dB Data 64 kbps 2.5 dB 5.3 dB 5.3 dB Data 144 kbps 2.3 dB 5.0 dB 5.0 dB Data 384 kbps 2.4 dB 5.1 dB 5.1 dB

A report to GSMA
100
Ranges between 1 and 2 dB for mobiles moving with a velocity of 50 km/h Is marginal and may be assumed at 0.1 dB for high speed mobiles (120 km/h)
Soft Handover Gain: Soft handover occurs if a mobile is connected to at least two cells at a time. If the cells belong to different Node Bs, the uplink signal combining is performed at RNC level. If the cells belong to the same Node B, the uplink signal combining is performed at the same Node B. In the downlink, the combining is done by the mobiles RAKE receiver. The total soft handover gain is assumed to be between 2dB and 3 dB Error! Reference source not found..3dB. Other Losses: Indoor losses: Typical value 15 dB In-car losses: Typical value 8 dB Log-normal fading margin: 10 to 12 dB for indoor, and 6 to 8 for outdoor. Interference due to co-location with GSM 900MHz (dB): An interference loss (in dB) has been considered in the case of coexistence of WCDMA and GSM carries in the same frequency band of 900MHz. If proper network and frequency planning is followed along with carrier frequency separation, as specified in ECC Compatibility Study for UMTS in 900MHz and 1800MHz frequency bands [7], then these losses can be minimal. For the link budgets above, the cell range R can be calculated for the given propagation models as it is briefly explained in Figure A.6.

A report to GSMA
101
Figure A.6 Calculation of cell range R
System parameters
Max path loss between mobile and base station antennas [dB]
Required coverage probability
Margins to guarantee service also within shadowing and indoors
Area type
Propagation model to convert dB to kilometres (Okumura-Hata)
Source: Ovum
A.5 Sectorisation and calculation of the area of coverage

After the cell range estimation, the site coverage area and site-to-site distance can be calculated. The cell area is used to calculate the number of sites required in a specific environment to fulfill the coverage requirements. Some example figures for coverage area and site-to-site distance calculation is provided in Figure A.7 Figure A.7 Coverage area and site-to-site distance for different number of sectors as a function of cell radius R Omni Two Sectors Three Sectors Six Sectors
Site-to-site distance
3 R 2 3 3 2 R 2
2R
1. 5 R
3 R 2 3 3 2 R 2
Site area
3 3 2 R 4
9 3 2 R 8
Source: Ovum

A report to GSMA
102
A.6 Frequency bands

Frequency band available assumed in FDD is 2x5MHz. Bearing in mind that every UMTS carrier is a 5MHz carrier this leads to 1 carrier per sector/cell for either the uplink or the downlink. Also, we have considered three sectors Node-Bs and thus we have considered three carriers per Node-B. Of course, model allows for flexibility regarding of available FDD spectrum and considered number of carriers per Node-B consequently. Proper frequency planning is further required in order to calculate optimal number of carriers per Node-B. For simplicity reasons we follow initial approach (1 carrier per sector/cell) described above.
A.7 Pole Capacity for Capacity-Based Dimensioning

A.7.1 Capacity based dimensioning using different services
The load factor for the uplink is defined as
nUL =
(E b
N0 ) NR (1 + i) , W
and the
noise rise as
Noise _ Rise =
1 1 nUL
For
nUL = 1 ,
the corresponding noise rice approaches to infinity and the system
has reached its pole capacity and becomes unstable. We want to operate our system for capacity based dimensioning at the 75%
(nUL = 0.75) of its pole capacity; this will give 6dB noise rise.
We calculate the joint service throughput that will give us the desired noise rise for our capacity driven dimensioning. Let us denote by to noise ratio for service-k,
( Eb / N 0 ) k
the desirable power
is the activity factor for the service (Typical values
is 1 for data, 0.65 for 50% voice activity in uplink and 0.58 for voice activity in downlink) and W=3.84 Mcps.

A report to GSMA
103
The other cell to own cell interference ratio transmission and 0.65 for 3 sectors. By
is 0.55 for omni directional
ck
we denote the ratio of the overall data
from service k, over the overall data from all the services. Then the available throughput of a cell given the noise rise and the ratio between the services will be calculated as:
N 0 )v (E N ) (E N ) Rcv v (1 + i ) + b 0 1 Rc11 (1 + i) + b 0 2 Rc 2 2 (1 + i ) = W W W (1 + i) nUL = [(Eb N 0 )v cv v + (Eb N 0 )1 c11 + (Eb N 0 )2 c 2 2 ] R W nUL W R= [(Eb N 0 )v cvv + (Eb N 0 )1 c11 + (Eb N 0 )2 c2 2 ] (1 + i) nUL =
In this way, cell throughput (R) is thus calculated. Example: Assume A Mbits of voice, B Mbits of data1 service, C Mbits for data 2 service (overall for the busy hour), then the data ratios for all the services will be:
(E b
cv =
A B C , c1 = , c2 = . A+B+C A+ B+C A+ B +C
Also, v
= 0.58,1 = 2 = 1,
and
( Eb / N 0 ) k
is given from the 3GPP table
describing desirable signal-to-noise ratio for different services and types of environments. Finally, we divide the overall data (A+B+C) over the calculated R for the required number of sites. Important: It must be noted that since we use BLER=10% for the desirable signal-to-noise ratio, we have to include the packet retransmission for the data services. Thus, the overall data are (A+(B+C)/(1-BLER)).

A report to GSMA
104
A.8 Traffic Assumptions for Capacity-based dimensioning

Usually telecommunication systems are dimensioned for the busy hour, where the peak load (network traffic) is served with a required quality of service and given resource availability and it is qualified by its blocking probability Error! Reference source not found.. In our calculations we assume that the busy hour carries 10% of the daily traffic.
A.9 References
[1]. M. J. Nawrocki, M. Dohler, A. H. Aghvami, Understanding UMTS Radio Network Modeling Planning and Automated Optimization, Wiley 2006. [2]. H. Holma, A. Toskala, WCDMA for UMTS, Radio Access For Third Generation Mobile Communications, Wiley 2002. [3]. 3GPP TR 25.816, UMTS900MHz Work Item Technical Report. [4]. Cost 231 Final Report, Digital Mobile Radio Towards Future Generation Systems. [5]. 3GPP TR 25.101, User Equipment (UE) radio transmission and reception (FDD). [6]. 3GPP TR 25.104, Base Station (BS) radio transmission and reception (FDD). [7]. ECC within CEPT, ECC Report 82, Compatibility study for UMTS operating within the GSM 900 and GSM 1800 frequency bands, May 2006.

A report to GSMA
105
Annex B Operator benchmarks

B.1 Cingular
B.1.1 Background
Under a previous agreement with NTT DoCoMo, US provider AT&T Wireless (now Cingular) was required to build and market UMTS networks in four major United States cities by the end of 2004. At CTIA 2004, Cingular announced that their 3G network would be a 1900MHz only implementation of UMTS and would launch by the end of that year as planned. Eventually, Cingular chose to deploy on both 850 MHz and 1900MHz.
B.1.2 Currently available networks

Cingular is the largest wireless service provider in the US. It currently runs:
An AMPS and TDMA Service on 850MHz till 18 February 2008 A GSM/GPRS/EDGE 850/1900MHz. A UMTS/HSDPA 850/1900MHz which has been in operation since December 2005. It also supports HSDPA in selected cities. It is currently the leader in 3G in US market.
B.1.3 Reason for Cingular switching from CDMA to UMTS 850

Why UMTS 850? Since the acquisition of Bellsouth (GSM at 850MHz) and AT&T (GMS 1900MHz), Cingular has more spectrum in 850 MHz band than 1900MHz. Given the spare capacity, Cingular had been planning for the past two years on utilizing more of its 850 MHz holdings for 3G, since the demand for TDMA is decreasing as Cingular migrates customers to its GSM network. The carrier's remaining 3.6 million TDMA customer account for only about 1 percent of its network traffic. Cingular holds an average of 58 megahertz of spectrum in the nation's top 100 markets which represents over two thirds of total population, and it would like to use its 850 MHz spectrum licenses as much as possible, leaving its 1.9 GHz spectrum for international roaming opportunities. Most international handsets include support for the 1.9 GHz spectrum bands in addition to the 900MHz and 1.8 GHz spectrum bands used in Europe. Cingular's UMTS/HSDPA technology can use the same spectrum for both voice and data, while CDMA2000 1x EV-DO technology requires clean'' channels for voice and data separately. EV-DO Revision A, which is the next network step for both

A report to GSMA
106
Sprint Nextel Corp. and Verizon Wireless, will offer CDMA carriers the capability to run both data and Voice over Internet Protocol services over the same channel. Why Cingular? Given the fact that Cingular has more subscribers than its competitors, Cingular is far more powerful to demand for handset support for rolling out UMTS in an unconventional spectrum than any other operators in the US. Among other U.S. carriers, a number of smaller carriers are expected to rollout UMTS 850 under their roaming agreements with Cingular. Apart from 850 MHz, Cingular and T-Mobile USA Inc. plans to launch their own UMTS networks using the 1.7/2.1 GHz spectrum that it recently purchased in auctions.
B.1.4 Milestones and progress so far

In December 2005 Cingular commercially launched these 3G networks in 16 major cities. Development and deployment of UMTS with HSDPA continued since then. This included replacement of the six UMTS systems which had been previously deployed by AT&T Wireless. By 2Q 2006, Cingular has UMTS coverage of more than 27 cities. Device support Despite the fact that Cingular looks more powerful than anyone else in the US market, handset support was very limited in the early stage. Although the range has now grown to 5 handsets, with 2 from LG and 3 from Samsung, only LGs CU 500 has proved to be popular so far. Cingular announced that it also has agreements with its strategic handset providers Nokia, Motorola Inc. and to deliver UMTS devices by the end of 2006. The carrier said the devices would include ''attractive form factors and feature functionality that takes advantage of the increased data speeds.'' e.g. Samsungs Blackjack, which also supports push emails. A few PC data cards are also currently available. Laptop access is another popular access technology to UMTS 850, thanks to the flat rate data pricing. It typically costs USD 60 per month for unlimited access (effectively about 10GB per month) to UMTS based data service, with download speed about 400kbps.

A report to GSMA
107
B.2 Telstra
B.2.1 Background
In recent years Telstra has lost overall retail mobile connection share, although it gained subscribers on its CDMA and 3G networks. It has lost ground in the GSM market where the bulk of the subscribers are, dropping 6.2% of connections on a year-on-year comparison. Telstra has responded by launching its 850MHz WCDMA network in October 2006 which will ultimately replace its existing CDMA network, simplifying Telstra's mobile network technology and cutting costs. The success of Telstra's national network with UMTS 850 is integral to maintaining its market share.
B.2.2 Currently available networks

Telstra currently has : a 3G network (UMTS2100) running at selected locations in Sydney, Melbourne, Geelong, Brisbane and Gold Coast, Perth, Adelaide and Canberra, a CDMA/EV-DO 1X network running at 850 MHz and a GSM network with GPRS technology a recently launched UMTS 850 network.
All other operators UMTS services in Australia are running at 2100MHz. Telstra announced that it would replace its existing CDMA/EV-DO network with Next G (UMTS 850) in November 2005 and the network became operational in October 2006. The network will be upgraded and developed by using Ericssons software enhancements to base stations in 2007. Telstra will also continue to extend the national coverage until it has the same or better coverage as its existing CDMA coverage (98% of the population) before it is expected to decommission the network in early 2008. The current CDMA network and the new network will run concurrently for some time before Telstra retires the CDMA network. Arrangements have been made to sell redundant CDMA equipment in a secondary market.
B.2.3 Reason for Telstra switching from CDMA to UMTS 850

Telstra considered 15 different solutions and chose this option as the one that gave it the best economic and customer-focused proposition: 1. Why choose UMTS 850: A national WCDMA network at 2.1 GHz would require three to four times more base stations than at 850MHz.

A report to GSMA
108
Roaming made easy: International roaming possibilities are greatly increased in WCDMA compared to CDMA, especially in Europe where few CDMA networks exist. Telstra anticipates a set-up similar to many GSM deployments, where the lower band (850MHz in this case) is deployed for coverage, and the upper band (2100MHz) is used as a capacity overlay.
2. One network only: To remove duplication and unnecessary costs
The cost of managing two networks with separate core switching systems was too great for Telstra. The networks also had a massive coverage overlap. At the time of the announcement, Telstra had over 4,916 GSM towers, over 3,488 CDMA towers and over 2,100 WCDMA towers at 2.1GHz, with three core switching systems. Geographically, the GSM footprint is 70% duplicated by CDMA, and the CDMA footprint is 80% duplicated by GSM. Although a large amount of capex will be required initially, Telstras migration from CDMA to UMTS 850 will save opex and capex in long run. Annual opex is important to operators and is usually four times infrastructure capex.
3. One major upgrade is desired : The national footprint of one 3G network will provide a greater advantage than subsequent incremental speed upgrades.
B.2.4 Milestones and progress so far

Telstra is constructing Australias only national 3GSM 850 wireless network which will reach more than 98% of the Australian population when it is completed. Telstra will operate services over both the GSM and CDMA networks until 2008 when its national 3GSM 850 network will provide the same or better coverage as the other networks. Progress: Nationally, by 30 June 2006 over 60% of the 5,112 sites had equipment installed, with over 80% of transmission lines completed First video call on the 3GSM 850 network has already been placed between country Victoria and Sydney Currently transmitters are able to throw a signal 80 kilometres on the new network. Work is still in progress to upgrade the range to 200km, which is on track for completion in the latter part of 2007. Telstra has provisioned the network for capacity. Additional backhaul capacity has been rolled out (using both fibre and microwave solutions) including some entirely new deployments in rural Australia. In the radio access network, Telstra anticipates a set-up not unlike many GSM deployments, where the lower band (850MHz in this case) is deployed for coverage, and the upper band (2100MHz) is used as a capacity overlay.

A report to GSMA
109
The Option datacard offered at launch supports HSDPA at 850/1900/2100MHz and EDGE/GPRS at 850/900/1800/1900MHz. It will be able to support throughput on the downlink to a peak of 7.2Mbit/s. This will be enabled by software upgrade, which Telstra is planning for early in 2007. Subsequent technology iterations to the network peak of 14.4Mbit/s and beyond will require users to upgrade their CPE to a PC card to be supplied by ZTE from November 2006, which will also support circuit-switched voice calls (using the PC as a softphone). This amount of activity in the field offered synergies in expanding and upgrading its existing footprint. Consequently, an additional 100 3G towers were deployed in Sydney, and 2,000 GSM towers were upgraded to 3G. Concurrently, Telstra has upgraded most of its GSM footprint to support EDGE.
Telstra believes that the Next G network is a source of competitive advantage. As such it has supplanted the fibre to the node (FTTN) plans as a focus for the business. The FTTN plans were indefinitely shelved in September 2006 as a result of unreasonable regulatory outcomes. However, contrary to Australian media reports, this does not mean Telstra aims to provide broadband to rural Australia primarily with the new network: it is pitched as a premium offering, rather than any kind of FTTX replacement. The source of Telstras competitive advantage is two-fold: coverage and performance. Both are enhanced by the better signal propagation achieved at 850MHz. The coverage of the network is unparalleled in Australia. When completed, it will cover 1.9 million square kilometres of country, up from 1.65 million square kilometres at launch. By contrast, competitor 2G/3G networks will cover capital cities and some suburbs and transport corridors. Telstra claims its network currently covers 98% of the population compared to Optuss 37%. Telstra has an initial throughput advantage, which it hopes to maintain through incremental upgrades over the next couple of years. Roaming is covered by tri-band HSDPA and quad-band GSM support: Telstra has negotiated 3G roaming agreements with 31 countries worldwide so far.
B.2.5 Customer migration will be a challenging process

Telstra plans to slowly decrease CDMA handset sales and then switch users over once UMTS850 coverage is as good or better than the CDMA network. Since CDMA and GSM are designed for different purposes given that CDMA covers larger distances and was deployed first and foremost for rural areas and does what GSM could not, rural users handset replacement cycles are generally much longer than those of city. Telstra will have to migrate 1.3 million CDMA users. At launch, Telstra be offered its customers attractive and competitive plans to encourage them to take up the new service, including offers like $0 upfront for

A report to GSMA
110
handsets on 24 month contracts. Collectively Telstra claims there are in excess of 30 devices expected to become available in the next 12 to 18 months.
B.3 Issues common to all UMTS900 operators

B.3.1 Interference between UMTS900 and GSM900 networks
Based on simulation results for different scenarios in which UMTS and GSM co-exist in the 900MHz band, the ECC report has concluded that these two technologies can technically co-exist without significant interference that can cause call dropping in either system.
B.3.2 Implications for operators to allow transition from GSM 900 to UMTS900
Cingular and Telstra, the only two operators currently supporting UMTS 850, require devices that support UMTS 850/1900 and UMTS 850/2100 respectively. Handset support is still limited and volumes will only build when tri band (2100/1900/850) devices are available. In case of UMTS900 is slightly different. The availability of such device technology will not be a major issue, given most handset manufacturers have either developed prototypes or have gained experience in supporting UMTS 850 well before the rollout of UMTS900. Handset and chipset manufacturers will converge towards developing quad band devices (2100/1900/900/850) as most mobile markets in the world support UMTS2100 (except North America) and UMTS900 (except North America, Japan, Australia). The success of UMTS900 deployment lies in good incentives for subscribers to migrate or connect to UMTS900, such as price and quality of the devices, and the price of operators subscription services. This will only happen when tri band or quad band devices become popular.
B.3.3 Impact of device availability

Device availability is considered to be a key factor for successful service adoption. It can be shown that limited handset choices have restricted the uptake for all 3G services, including Cingulars 3G and Telstras Next G. From Ovums analysis, the following handset factors noted below affect 3G customer uptake around the world and the key successful factors (KSF). Telstra declines to reveal its subscriber numbers, but it is understood the Next G service had at least 30,000 subscribers at the end of October 2006 (11% of all Telstras WCDMA subscribers), less than a month after its launch.

A report to GSMA
111
Figure B.1 Key factors for successful service introduction
Factors Quality of handsets (physical size, battery life, features, etc.)
KSF 3G handsets match or exceed 2G equivalents
Examples and proof points The launch of Nokia 7600 and clamshell LG 3G handsets (which are slimmer and smaller in size than handsets made by all other mainstream manufacturers) in Q1 2004 gave Carphone Warehouses sales and Hutchison 3s subscriber uptake a significant boost. The chart below shows that WCDMA subs grew rapidly as the choice of handsets increased significantly from Q2 2004. The chart shows that the there is only a small reduction in price during this period (highlighted blue).
Range of handsets
Wide range of handsets at numerous price points and with brands which appeal to a wide range of different customer segments
Figure B.2 Key Relationship between available WCDMA devices and take-up
30 25 20 15 10 5 0
WCDMA Subscribers world-wide (Millions)
35

A report to GSMA
112
Annex C GSM-900 and E-GSM bands

Table 1: GSM-900 and E-GSM frequency band allocations in EU-15 countries in 5MHz blocks (Note includes blocks of 4.8MHz also)
Country Operator T-mobile Austria Mobilkom One Tele.ring Proximus Belgium Mobistar KPN/Orange Sonofon Denmark TDC Telia Orange Radiolinja Finland Sonera Finnet Orange France SFR Bouygues Telecom T-mobile Germany Vodafone E-Plus O2 Vodafone Greece TIM Hellas C osmote O2 Ireland Vodafone Meteor Tim Italy Wind Vodafone LuxGSM Luxembourg Tango VoxMobile TMN Portugal Telecel Optimus Vodafone/Movil Spain Telefonica Moviles Amena TeliaSonera Sweden Tele2 Vodafone Vodafone KPN Mobile The Nethe rlands Orange Telfort T-Mobile T-mobile UK Vodafone O2 Orange GSM- 900 2 x 2x5MHz 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz E- GSM 2x5MHz 2x5MHz 2x5MHz 2 x 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz 2x5MHz JAN 2011 DEC 2010 DEC 2010 DEC 2010 MAR 2021 MAR 2021 DEC 2024 DEC 2009 DEC 2009 DEC 2016 DEC 2016 SEP 2012 SEP 2012 MAR 2023 2011 2011 2015 FEB 2015 JUN 2018 FEB 2015 JUN 2008 MAY 2008 APR 2020 MAR 2007 OCT 2006 NOV 2012 FEB 2025 FEB 2015 JUN 2020 DEC 2010 DEC 2010 DEC 2010 MAR 2010 MAR 2010 NO EXP DATE NO EXP DATE NO EXP DATE NO EXP DATE
The E-GSM band is assigned for res e a rch purposes and for a re g ional GSM opera to r 2 x 2x5MHz (E-GSM) 2 x 2x5MHz (E-GSM) 2x 5MHz (E-GSM) 2 x 2x5MHz (E-GSM) 2x 5MHz (E-GSM) Not-allocated spectrum in GSM-900 & E-GSM bands (5MHz blocks)
Expiration Date DEC 2015 DEC 2015 JUL 2011 NOV 2010 JUL 2013
SOURCE: EUROPEAN RADIOCOMMUNICATIONS OFFICE (ERO) / NRAS

Umts900 Full Report

Uploaded by

Document Information

Original Description:

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Umts900 Full Report

Uploaded by

Copyright:

Available Formats

Market Study for UMTS900

Project Number CLW28 Version V1.1

Ovum Consulting, Cardinal Tower, 12 Farringdon Road, London EC1M 3HS

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

3.1 Demand and network cost analysis

Figure 3.1 Percentage increase in coverage area

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

Ba seline scenario: 2.1 GHz only

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

Baseline scenario: 2.1 GHz only

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

Baseline scenario: 2.1 GHz only

Asia-Pac Low Demand

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

Baseline scenario: 2.1 GHz only

Asia-Pac Low Demand

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS900 and UMTS2100

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

Figure 3.8 Updated NPV percentage improvements over UMTS2100 only

Source: Ovum model

3.2 Case studies examples

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

3.2.1 Sri Lanka

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

3.2.3 Saudi Arabia

3.2.4 South Africa

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

3.3 Key regulatory issues

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

3.4 Handset and chipset analysis

3.5 Economic analysis

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

By improving productivity By stimulating change in business working practices.

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

UMTS 900 Market Study Final Report V1.1

MARKET STUDY UMTS 900

4.1 Implications of UMTS900 technology