Understanding the Scope for a Power Increase for Wireless Broadband Access at 2.4GHz & 5.xGHz Final Report: V1.0

Transcription

1 Understanding the Scope for a Power Increase for Wireless Broadband Access at 2.4GHz & 5.xGHz Final Report: V1.0 Ofcom 10/5/2006 Ofcom Reference C31400/013/1510 Scientific Generics Limited Harston Mill Harston Cambridge CB2 5GG A Generics Group company This report has been prepared solely for Ofcom and may not without permission be disclosed to any third party C Final Report v1-0.doc

2 GLOSSARY ACA Australian Communications Association Access Connection based on a point to multipoint topology ADSL Asymmetric Digital Subscriber Line AFH Adaptive Frequency Hopping ARPU Average Revenue per User AVI Automatic Vehicle Identification Backhaul Base Station connections to an internet peering point BS Base Station CDMA Code Division Multiple Access CEPT European Conference of Postal and Telecommunications Administrations CPE Consumer Premises Equipment DFS Dynamic Frequency Selection EESS Earth Exploration Satellite Service EIRP Effective Isotropic Radiated Power ERC European Radio Communications Committee ETSI European Telecommunications Standards Institute FCC Federal Communications Commission FDD Frequency Division Duplex FDMA Frequency Division Multiple Access FSS Fixed Satellite Service FWA Fixed Wireless Access ISM Industrial, Scientific, Medical ITU International Telecommunication Union JFMG Joint Frequency Management Group LE Licence Exempt LOS Line of Sight MAC Medium Access Control MIMO Multiple-Input Multiple-Output MoD Ministry of Defence NLOS Near Line of Sight OFDMA Orthogonal Frequency Division Multiple Access PAN Personal Area Network PHY Physical Layer PmP Point to multi-point PMSE Programme Making and Special Events POS Point of Sale PSD Power Spectral Density PtP Point to Point QoS Quality of Service RFID Radio Frequency Identification RLAN Radio-based Local Area Network RTTT Road Transport Traffic Telematics SAR Synthetic Aperture Radar SDSL Symmetric Digital Subscriber Line SFN Single Frequency Network SIG Special Interest Group SIP Session Initiation Protocol SME Small or Medium sized Enterprise SOHO Small Office / Home Office SRD Short Range Device TDD Time Division Duplex TDMA Time Division Multiple Access C Final Report v1-0.doc

3 TPC UAV UNII VoIP WBA WiFi WLAN WiMAX WISP Transmit Power Control Unmanned Airborne Vehicle Unlicenced National Information Infrastructure (US) Voice over Internet Protocol Wireless Broadband Access Wireless Fidelity (IEEE802.11b Wireless Networking) Wireless Local Area Network Worldwide Interoperability for Microwave Access Wireless Internet Service Provider C Final Report v1-0.doc

4 CONTENTS EXECUTIVE SUMMARY INTRODUCTION Study objectives Scope of work Industry consultation Structure of Report SPECTRUM REGULATIONS AND USES OF 2.4 AND 5.XGHZ BANDS IN THE UK Current UK regulations regarding RLAN/WBA operation in Licence- Exempt frequency bands above 1GHz Suitability of 2.4GHz and 5.xGHz frequency bands for the provision of WBA Feedback from industry stakeholders on 2.4 and 5.xGHz suitability for WBA Current/Previous usage of spectrum for Licence-Exempt WBA WBA system technical characteristics and deployment scenarios Allocations and other uses of 2.4 and 5.xGHz bands INDUSTRY CONSULTATION Objective and scope of consultation Interviewee and stakeholder views on the suitability of 5GHz bands for WBA Interviewee and stakeholder views on the suitability of 2.4GHz bands for WBA Conclusions from the industry consultation DEVELOPMENTS AND TRENDS Regulatory developments relevant to higher power Technology developments relevant to higher power Self install Relevant standards Equipment developments relevant to higher power CALCULATION OF SOCIAL VALUE OF CHANGES TO REGULATIONS CALCULATION OF BENEFITS Overview of approach to calculating benefits Geographic modelling Network Modelling Demand modelling and consumer surplus Benefits, results and interpretation Other benefits arising from a power increase Feedback from the stakeholders on WBA market development issues Benefits modelling conclusion C Final Report v1-0.doc

5 8 CALCULATION OF COSTS Overview of approach to calculation of costs Affected parties Feedback from the stakeholders on interference cases Calculation of Cost of Interference at 2.4GHz Interference costs at 5GHz Band A ( MHz) Interference costs at 5GHz band A ( MHz) Interference costs at 5GHz band B Interference costs at 5GHz Band C (5725 to 5875MHz) Interference with users of adjacent bands SUMMARY OF AGGREGATE NET BENEFIT/LOSS TO THE UK Benefits Arising from Higher Power Costs Arising from Higher Power Restricting higher power geographically at 2.4GHz OPTIONS AND RECOMMENDATIONS FOR REGULATORY CHANGE Regulating by conducted power vs. regulating by radiated power GHz Regulation Recommendations xGHz Regulations Recommendations Distribution of costs CONCLUSIONS ANNEX A: BIBLIOGRAPHY ANNEX B: INDUSTRY CONSULTATION List of companies consulted Feedback from interviewees on the benefits of increasing power Feedback from interviewees on the preferred bands(s) for a power increase Feedback from interviewees on interference cases Feedback from interviewees on technical developments Feedback from the interviewees on costs and the power increase required Feedback from the interviewees on WBA market development issues ANNEX C: EXPERIENCE OF OTHER COUNTRIES USING HIGHER POWER US Australia New Zealand Ireland Sweden Conclusion ANNEX D: NETWORK MODEL OUTPUTS AND SENSITIVITY ANALYSES Additional network model outputs C Final Report v1-0.doc

6 15.2 Sensitivity analyses ANNEX E: METHODOLOGY FOR COST BENEFIT ANALYSIS Introduction Net private benefits Calculating consumer surplus ANNEX F: COGNITIVE RADIO (CR) TECHNIQUES FOR AVOIDING EESS Introduction Sensing Database Conclusion ANNEX G: FINANCIALLY VIABLE CELL PLOTS Financially viable cells Cells which become financially viable at higher power Assessing the Geographical Impact of Hexagons Potentially Benefiting from Higher Power ANNEX H: SMART ANTENNAS AND REGULATIONS Smart antenna technologies UK regulations FCC regulations Example of compliance with regulations in maximum EIRP and its impact in performance of smart antennas for range extension Conclusions ANNEX I: REVIEW OF PREVIOUS STUDIES EXAMINING COEXISTENCE GHz xGHz ANNEX J: POSTCODES OF AREAS WITH HIGH BUSINESS DENSITY 245 C Final Report v1-0.doc

7 Document history Version Date Author(s) Changes V May 2006 TJO,PJB, RPW,MRR, AMR Initial Release C Final Report v1-0.doc Page 1

8 EXECUTIVE SUMMARY This report describes the results of a study by a consortium led by Scientific Generics Ltd. The other consortium members were Lucent Technologies, DotEcon Ltd and South East England Regional Research Laboratory (SERRL) at Birkbeck University of London. The Office of Communications (Ofcom) commissioned this study to investigate the options for permitting higher power levels and/or making other regulatory changes in the 2.4GHz and 5.xGHz bands for licence-exempt applications. Background Both Ofcom and the Government see the adoption of broadband as a national priority. Existing technologies for providing broadband are either costly or inappropriate for remote or rural communities. Many rural communities do not have access to DSL or cable and Expensive satellite broadband is often the only option. A solution could be Wireless Broadband Access (WBA) although to date the business case for rural WBA has proved to be marginal. Currently operators may use 2.4GHz for providing services but maximum EIRP is limited to 100mW. WBA services may also be provided at 5.8GHz where the maximum EIRP is 2W. A proposed solution is to increase the transmitted power level for licence exempt applications to improve the business case by reducing infrastructure costs. The feasibility of raising power levels depends on the increase necessary and on the consequences of higher power in terms of interference and steps to mitigate that interference. The term WBA is used predominantly throughout this report. Where references are made to previous industry studies or frequency allocations the term Fixed Wireless Access (FWA) is used to be consistent with the existing terminology. Study Objectives The aim of the project is to establish a basis on which Ofcom may consult on whether to permit higher powers in the licence-exempt bands taking account of co-existence issues. The primary objective of the programme of work was to assess the options for revising the regulation of licence-exempt applications, including power levels, to promote use particularly where spectrum is currently under-used. The frequency bands examined were 2.4GHz, and 5GHz band A ( MHz), band B ( MHz and band C ( MHz). Band A is currently allowed for WLAN but restricted to 200mW and indoor use only. Nomadic Radio LANs (RLANs) may be used in band B outdoor and the maximum EIRP is 1W. C Final Report v1-0.doc Page 2

9 The secondary objective of the work programme was to explore the use of higher powers for licence exempt applications in order to provide wireless broadband in rural areas where other broadband technologies might be unavailable due to prohibitively high cost. Scope of work The study was divided into two phases: Phase 1 During the first phase we undertook a series of interviews with key industry players to understand their views on the need for regulatory change and the likely benefits and losses that may arise. Also during this phase we examined in detail previous studies into co-existence that were relevant to this project and researched the current services and users of the 2.4GHz and 5.xGHz bands. Finally, we derived a set of higher power scenarios for further investigation in phase 2. Phase 2 During the second phase the scenarios defined in phase 1 were examined in detail. A model was developed to examine the relationship between power limits and private net benefits. These are the benefits that consumers obtain from WBA net of the price they have to pay for the service. The model is a combined geographic, network and economic model. For each of the scenarios we estimated the costs of interference that might arise as a result of allowing higher powers. Since there are many different types of user and services that might be affected by increasing powers we assessed qualitatively the service/user that would be most likely to incur the highest cost in each of the 2.4GHz and 5.8GHz bands. For the most affected type of service in each band we then assessed quantitatively the likely cost of interference. Based on the results of the model and other research and analysis performed during the project we have made recommendations for consideration for changing the regulations in the 2.4GHz and 5.xGHz bands. These should be considered alongside other options Ofcom has for regulatory change in licence exempt spectrum. Conclusion from Industry Interviews A broad conclusion from the interviews is that the rural Wireless Internet Service Providers (WISP) business case is marginal under the current regulations, at both 5.8GHz and 2.4GHz but particularly at 2.4GHz. Increased availability of ADSL has eroded the ability of WBA to compete for residential subscribers, and in rural areas the cost of extending existing networks to reach the more valuable business customers can be prohibitive. Higher power is therefore useful, in order to: 1 Improve the economics of serving residential customers 2 Improve the ability of rural WISPs to serve business customers C Final Report v1-0.doc Page 3

10 According to industry views, there are two main scenarios for higher power: 1 A modest power increase which is unlikely to cause significant additional interference, but will provide some benefits for rural WISPs 2 A significant power increase which will greatly improve the business case for WBA, but could cause interference for other spectrum users. Many interviewees expressed resistance to increasing power at 2.4GHz because of concerns about interference with WLANs. However, they accepted that the recent rapid increase in ADSL availability means that significant increases in power are required at this frequency to create a viable commercial proposition and that 2.4GHz offers powerful advantages compared to 5.8GHz through lower equipment costs and longer range for the same power. Interviewees agreed that to deploy a higher power 2.4GHz solution the critical issue is whether effective regulatory mechanisms can be developed to restrict use of such systems to rural areas. If such a constraint was feasible then costs of interference could be restricted to these areas only. At 5.8GHz there were mixed views regarding the impact that an increase in power to 4W would have on the business case. Smaller operators value higher power for backhaul and village-village links. Larger operators look for reduced infrastructure costs for covering an area. Some interviewees also expressed a need for more bandwidth. Extension of FWA to band B was considered by some to be desirable. Coexistence Issues 2.4GHz Based on the literature studied and modeling undertaken during this project several conclusions can be drawn on potential interference between high power WBA and other devices at 2.4GHz Much of the interference occurring would be between high power WBA and WLAN due to the proliferation of WLANs in the UK. This is mitigated to some degree since it is expected that WBA systems would use a similar medium access control (MAC) to WLANs and hence they would share spectrum politely. It is expected that residential users of WLAN would be relatively unaffected because the typical user only takes advantage of a small proportion of the capacity available. Business users are much more likely to be operating in an environment where there is a high density of traffic and all WLAN channels in use. In this scenario there is more likely to be a conflict over radio spectrum resources. From the modeling work undertaken it is expected that outdoor WBA would suffer interference from an indoor WLAN over a longer range than vice versa. Commercial WBA operators are unlikely to want to use 2.4GHz for WBA-type services in non-rural areas or near business parks because of the risk of interference from other WLAN users of the band. Anecdotal evidence from operators supports this hypothesis. C Final Report v1-0.doc Page 4

11 If, through appropriate regulation, it is possible to restrict high power use to areas where there is low WLAN usage then there would be little interference. During our research we found little evidence that there had been significant problems between 2.4GHz WISPs using high power and WLAN users in the U.S. where significantly higher power levels are permitted. This could be because U.S. operators have avoided setting up networks in areas where there are high densities of WLAN users because this would impact the quality of service they could provide. Coexistence between high power WBA and Bluetooth is significantly improved by the introduction of AFH in Bluetooth V1.2 Other systems and users that may be affected are RFID, Automatic Vehicle Identification (AVI), analogue wireless video senders and the Ministry of Defence (MoD). The MoD use the band MHz for Unmanned Airborne Vehicles (UAVs) and communications systems, predominantly in rural areas. 5.xGHz A significant number of studies have been undertaken which examine sharing between RLANs or FWA and other users/services. In the current WBA band ( MHz and 5815 to 5850MHz) the main concerns with interference are MoD and Met Office radars and Fixed Satellite Service (FSS). WBA systems must employ Dynamic Frequency Selection (DFS) to avoid radars. The WBA system, under current rules, must change frequency when a radar is detected. The detection threshold must be according to rules set by Ofcom in IR2007. Further research is taking place into coexistence with frequency hopping radars which are not currently protected by DFS. Interference with FSS is determined by the aggregate interference from WBA across Europe due to the broad antenna patterns of the satellites. WBA systems based on an omnidirectional mesh network topology are expected to cause significantly more interference than point-to-multipoint point systems due to omnidirectional antennas and higher activity ratios. In other parts of band C (5725MHz to 5875MHz) where FWA is currently not permitted there are other users/services to consider such as RTTT, PMSE and a higher number of more sensitive fixed satellites. 5GHz bands A and B could be considered for FWA but other services/users are Earth Exploration Satellite Service (EESS), radars, Mobile Satellite Service (MSS) uplinks and Programme Making and Special Events (PMSE). Benefits arising from higher power Several scenarios have been modeled, as set out in Table 1. All scenarios assume that residential users are provided with a 1Mbit/s SDSL service and business users a 4Mbit/s SDSL service. The WBA architecture is assumed to be point-to-multipoint. We have modeled how the benefits to the UK vary with the C Final Report v1-0.doc Page 5

12 maximum permissible EIRP at 2.4GHz and 5.8GHz. The 3.5GHz band, which is licensed with higher power limits for WBA services, has also been modeled for comparison. Frequency (GHz) Higher power scenario 2.4 1W EIRP W EIRP W EIRP 5.8 4W EIRP W EIRP W EIRP W EIRP Table 1 Higher power scenarios For each scenario the consumer surpluses that are expected to arise are shown in Table 2. The consumer surplus is the additional value in GBP for UK consumers and businesses that would be created under each scenario. Frequency Scenario Net consumer surplus 2 Cell radius GBP, million W EIRP W EIRP W EIRP W EIRP W EIRP W EIRP W EIRP 288 Table 2 Net consumer surplus The consumer surplus increases with EIRP because the costs of providing a wireless broadband service fall. These savings/benefits are passed on to consumers in the form of additional consumer surplus. Thus the increase in consumer surplus is greatest for the 2.4GHz 10W and 80W higher power scenarios. This is because consumer surplus arising at 2.4GHz under the current regulations is much lower than for the other cases, which in turn results from the poor financially viability of serving small cells. In contrast increasing power to 1W at 2.4GHz or to 4W at 5.8GHz has a more modest impact. There is also a significant increase for 25W and 200W at 5.8GHz. It should be noted that a high proportion of the total benefit is derived from business subscribers for the services modelled of which a high proportion arises from areas of medium-sized towns such as Chichester and Chesterfield. These 1 3.5GHz is a licenced band for WBA. We have used the same methodology for estimating benefits at 3.5GHz for comparison with the licence-exempt bands. 2 Net of consumer surplus under current regulations C Final Report v1-0.doc Page 6

13 businesses are more than 2kms from a BT exchange and are outside the range for 4Mbit/s services and therefore may benefit from WBA. Costs Arising from Higher Power Interference costs have been quantitatively assessed for the most significant cases: WLANs at 2.4GHz and Fixed Satellite Services (FSS) at 5.xGHz. WLANs present a significant potential cost because of the popularity of these devices over the last few years. Proliferation of WLAN devices is likely to continue in the future with technology developments that will enable them to be used for audio and video streaming applications. Using the same scenarios as the benefits model we have estimated costs of interference from high power WBA into WLAN, ranging between 600,000 at current 100mW power levels to 18m at 80W EIRP. This is the cost of mitigation that business users would incur in changing equipment to 5.2GHz WLAN to avoid interference with WBA at 2.4GHz. Interference costs for residential users are expected to be significantly less than this. Interference with FSS depends on the specific characteristics of each satellite. The location of the geostationary satellite, antenna characteristics and receiver parameters impact the level of aggregate interference from WBA users in Europe that is likely to affect services. Below 5850MHz there are only a few satellites that could be impacted. For the most sensitive satellites the number of WBA devices that can be deployed without satellite operators incurring losses is quite high. It has been estimated that satellite operators could incur losses of 29m if aggregate interference impacts services. These losses arise because the few satellites operating below 5850MHz would not be able to use a part of the spectrum allocated to them. The interference would prevent operators from providing services with the quality of service required by their customers. The value of the loss has been estimated by calculating the ratio of the bandwidth lost to total bandwidth allocated and then multiplying the satellite launch cost by this ratio. The MoD also uses the 2.4GHz below 2.45GHz and the 5.8GHz band. Between 2.4GHz and 2.45GHz MoD applications may suffer interference because there is no mitigation. Due to the nature of the applications the costs of interference are very difficult to quantify and therefore these costs have not been factored into the analysis. At 5.8GHz WBA equipment must employ Dynamic Frequency Selection (DFS) to avoid interference with radars. This is already mandated for equipment using this band. C Final Report v1-0.doc Page 7

14 Recommendations Based on the results of the work, the following options for the regulatory change are recommended for consideration: Licensing by conducted power and antenna gain The UK currently regulates power by defining the maximum Effective Isotropic Radiated Power (EIRP) at both 2.4GHz and 5.xGHz for licence-exempt applications. An alternative approach is to regulate by conducted power and antenna gain. This has the following advantages: Longer range systems are possible for similar interference footprint as EIRP method (assuming all gain is achieved by reducing the azimuth beamwidth). Simplicity of regulation (no need to certify complete systems) Encourages use of directional and smart antennas which are more spectrally efficient At 5.8GHz encourages behaviour which is less likely to cause interference with FSS by reducing the aggregate power that is transmitted into the satellite antenna main beam, i.e. omnidirectional mesh systems will have to transmit lower power than directional point to multipoint systems. Assuming the transmitter and receiver share an antenna no changes are required to the DFS mechanism. 2.4GHz An increase in the power under certain conditions in the 2.4GHz band is recommended since: The benefits of a power increase considerably outweigh the costs of interference. Most licenced users, such as PMSE, have vacated this band because of interference from WLAN. Interference between WLANs does exist but at present only in areas of high usage. Rural areas could significantly benefit from an increase with no interference cost. The following changes to the regulations are recommended: Constrain higher power equipment to 2.45 to GHz only to avoid interference with the MoD Recast the current EIRP regulations into conducted power: 50mW total conducted power (this is typically the maximum conducted power of Wi-Fi equipment) Antenna gain to 3dB C Final Report v1-0.doc Page 8

15 Allow directional antennas to be used: Allow antenna gain up to 30dBi Where antenna gain exceeds 3dB the conducted power must be reduced by 1dB for every 2dB that the antenna gain exceeds 3dBi (assuming gain is achieved equally from azimuthal and elevation directionality this will balance the increase in interference footprint from the elevation gain). Note that this is more conservative than the U.S. approach. Broadly speaking this measure would maintain the current interference environment whilst giving benefit for users of directional antennas. If the backoff is not applied the interference generated will increase with directionality although still at a lower rate than the operating range increases. 5.xGHz In the current 5 GHz WBA band there exists significant scope for increasing the transmitted power levels. The two main coexistence considerations in this band are between WBA and radars and WBA and FSS. Higher power systems employing DFS can operate without interference to radar if lower detection thresholds at the input to the antenna can be accommodated. Assuming that service take-up does not exceed the maximum estimated in this report (158,000 subscribers) then conducted power could be 250mW with a maximum antenna gain of 20dBi without causing service affecting interference, and hence losses with FSS. Therefore, the maximum EIRP in the antenna main beam is 25W. In deriving this result consideration has been given to the fact that WBA users in other European countries also contribute to interference with FSS. There is also scope for extending WBA to other 5GHz bands. Between 5570MHz and 5725MHz in band B the only other significant allocations are radar and PMSE. Due to the relatively low numbers of PMSE users co-ordination between WBA and PMSE should enable this spectrum to be shared. There is also some scope for high power WBA and nomadic use between 5470 and 5570MHz in band B. There is an allocation for Earth Exploration Satellite Service (EESS) but it is not used. Sharing mechanisms should also enable WBA to share this spectrum with future satellites. One such mechanism is described in Annex F and should be further investigated. PMSE is also allocated to this band. Sharing with radars in this can band also be achieved using DFS. The same coexistence situation applies to the upper half of band A ( MHz) as band B between 5470 and 5570MHz. However, there are active earth exploration satellites in this band. There is less scope for sharing between WBA and FSS in the band MHz because satellites in this band are more sensitive to interference and there are more of them than below 5850MHz. Furthermore, PMSE users have only recently been migrated to this band. C Final Report v1-0.doc Page 9

16 Coexistence between high power WBA and RTTT ( MHz) requires further investigation taking into consideration plans for future RTTT deployments. WBA should not be extended to the lower half of band A ( MHz) where there are active Mobile Satellite Service (MSS) uplinks without due consideration of the effects on these services. C Final Report v1-0.doc Page 10

17 2 INTRODUCTION This report describes the results of a study by Scientific Generics Ltd, Lucent Technologies, DotEcon Ltd and South East England Regional Research Laboratory (SERRL) at Birkbeck University of London. The Office of Communications (Ofcom) commissioned this study to investigate the options for permitting higher power levels and/or making other regulatory changes in the 2.4GHz and 5.xGHz bands for licence-exempt applications. A change to the regulations to allow higher powers may benefit rural communities in particular. Many rural communities have limited access to the internet via broadband because of the lack of DSL and cable in these areas. At present wireless broadband access (WBA) systems may be deployed at 2.4GHz but the maximum EIRP that may be transmitted is 100mW. This is considered to be a serious constraint to operating a financially viable broadband access service at this frequency. Most services operating in this band are noncommercial community broadband projects. WBA systems may also be deployed at 5.8GHz ( MHz and to 5.850MHz). The maximum EIRP at these frequencies is 2W EIRP. There are few operators providing commercial services at this frequency. The project is one of a number of projects funded by the UK Treasury under a scheme known as the Spectrum Efficiency Scheme. 2.1 Study objectives The aim of the project is to establish a basis on which Ofcom may consult on whether to permit higher powers in the licence-exempt bands taking account of co-existence issues. The primary objective of the programme of work was to assess the options for revising the regulation of licence-exempt applications, including power levels, to promote use particularly where spectrum is currently under-used. The frequency bands examined were 2.4GHz, and 5GHz band A ( MHz), band B ( MHz and band C ( MHz). Band A is currently allowed for WLAN but restricted to 200mW and indoor use only. Nomadic Radio LANs (RLANs) may be used in band B outdoor and the maximum EIRP is 1W. The secondary objective of the work programme was to explore the use of higher powers for licence exempt applications in order to provide wireless broadband in rural areas where other broadband technologies might be unavailable due to prohibitively high cost. C Final Report v1-0.doc Page 11

18 2.2 Scope of work The study was divided into two phases: Phase 1 During the first phase we undertook a series of interviews with key industry players to understand their view on the need for regulatory change and the likely benefits and losses that may arise. Also during this phase we examined in detail previous studies into co-existence that were relevant to this project and researched the current services and users of the 2.4GHz and 5.xGHz bands. A further objective of phase 1 was to derive some scenarios for further investigation. Phase 2 During the second phase the scenarios defined in phase 1 were examined in detail. A model was developed to examine the relationship between power limits and private net benefits. The model is a combined geographic, network and economic model. For each of the scenarios we estimated the costs of interference that might arise as a result of allowing higher powers. Since there are many different types of user and services that might be affected by increasing powers we assessed qualitatively the service/user that would be most likely to incur the highest cost in each of the 2.4GHz and 5.8GHz bands. For the most affected type of service in each band we then assessed quantitatively the likely cost of interference. Based on the results of the model and other research and analysis performed during the project we have made recommendations for options for changing the regulations in the 2.4GHz and 5.xGHz bands. 2.3 Industry consultation During the first phase of the project we held a number of interviews and discussions with key players who are active in wireless broadband for rural regions, in order to gain insights into the status of WBA in the UK. These interviews concentrated on gathering views from two key stakeholder groups representing supply and demand in this area: Operators Equipment vendors In addition, we have also sought input from other stakeholder organisations including the Low Power Radio Association (LPRA), British National Space Centre (BNSC), Met Office, European Space Agency (ESA), MoD and Joint Frequency Management Group (JFMG). C Final Report v1-0.doc Page 12

19 2.4 Structure of Report The report is divided into 11 sections: Section 2 summarises the current regulations relating to Radio LANs (RLANs) and WBA in the UK at 2.4GHz and 5.xGHz, current/previous usage of this spectrum for WBA and allocations for other uses in the same bands. Section3 describes the objective and scope of the industry consultation and results and conclusions arising from it Section 4 describes the results of the review into previous studies examining coexistence Section 5 describes the results of research into regulatory and technological developments that are relevant to this project Section 6 introduces the approach taken to the calculation of the change in social value (welfare) that could be brought about by increasing powers for WBA at 2.4GHz and 5.xGHz Section 7 describes the approach to calculating benefits. The detail of the geographic, network and economic models that have been developed are provided in this section. The results and conclusions arising from the modelling are described Section 8 describes the approach to calculating costs of interference that may arise from higher powers. For most services these are calculated qualitatively but at 2.4GHz the effect on WLANs was calculated quantitatively. At 5.8GHz (band C) the impact on fixed satellite services was also calculated quantitatively Section 9 briefly compares the benefits and losses arising Section 10 provides options and recommendations for regulatory change and other policy implications Section 11 provides overall conclusions and other recommendations not related to regulations The report also contains several Annexes: Annex A provides a bibliography of previous studies Annex B provides details of the feedback from interviewees Annex C describes the results of research into experience of other countries where higher power is allowed Annex D provides details of the network model and the sensitivity analyses conducted on the benefits model Annex E provides details of the methodology used for the cost benefit analysis Annex F describes an investigation into possible approaches for improving coexistence between Earth Exploration Satellite Service (EESS) and RLANs in 5GHz band A and B Annex G provides plots of an assessment that was conducted into the areas of the country that would benefit from higher powers Annex H describes the results of a qualitative investigation into changes to regulations that would benefit intelligent antenna techniques C Final Report v1-0.doc Page 13

20 3 SPECTRUM REGULATIONS AND USES OF 2.4 AND 5.XGHZ BANDS IN THE UK The bands considered during this project as being potentially suitable for licenceexempt WBA are: 2.4GHz (2.4GHz GHz) 5 GHz band A ( GHz) 5 GHz band B ( GHz) 5GHz band C ( GHz) which includes RTTT (5.795GHz to 5.815GHz) and 5.85GHz to 5.875GHz which is not currently allocated for FWA. Section 3.1 outlines the current regulations regarding WBA operation in these bands. Section 3.2 provides an initial assessment of the suitability of these bands for WBA. The term WBA is used predominantly throughout this report. Where references are made to previous industry studies or frequency allocations the term Fixed Wireless Access (FWA) is used to be consistent with the existing terminology. 3.1 Current UK regulations regarding RLAN/WBA operation in Licence-Exempt frequency bands above 1GHz GHz (2.4GHz GHz) The minimum equipment parameters for RLANs operating at 2.4GHz is defined in IR2005, UK Interface Requirement 2005 Wideband Transmissions Systems Operating in the 2.4GHz ISM Band and Using Spread Spectrum Modulation Techniques. The EIRP (Effective Isotropic Radiated Power) is limited to 100mW. For frequency hopping systems the maximum power spectrum density is 100mW/100kHz. For other forms of spread spectrum modulation the power density must not exceed 10mW/MHz. Indoor and outdoor use is permitted. Typical equipment is Wi-Fi wireless LAN and operates at 17dBm conducted power with a 3dB gain omidirectional antenna. Equipment must comply with the harmonised European Norm EN RLANs operating in this band are licence-exempt. C Final Report v1-0.doc Page 14

21 GHz Band A ( GHz) The minimum technical parameters regarding RLANs operating in band A are defined in IR2006, UK Interface Requirement 2006 Short Range, Broadband Data Services (High Performance RLAN) operating in the Frequency Range MHz. The EIRP is effectively limited to 23dBm in a (minimum) occupied bandwidth of 20MHz. Indoor use only is permitted. Operation in the upper half of the band ( GHz) requires the use of DFS (Dynamic Frequency Selection) and TPC (Transmitter Power Control) to minimise interference with licenced users of the band. Annex 1 of Recommendation ITU-R M provides details of the DFS and TPC requirements. Equipment must comply with the harmonised European Norm EN Typically, equipment is Wi-Fi wireless LAN and operates at 15dBm conducted power with up to 8dB gain omidirectional antennas. RLANs operating in this band are licence-exempt. Band B ( GHz) The minimum technical parameters regarding RLANs operating in band B are defined in IR2006, UK Interface Requirement 2006 Short Range, Broadband Data Services (High Performance RLAN) operating in the Frequency Range MHz. The EIRP is effectively limited to 30dBm in a (minimum) occupied bandwidth of 20MHz. Indoor and outdoor use is permitted. Operation in this band requires the use of DFS and TPC to minimise interference with licenced users of the band. Annex 1 of Recommendation ITU-R M provides further details of this provides details of the DFS and TPC requirements. Equipment must comply with the harmonised European Norm EN Typically, equipment is either Wi-Fi wireless LAN operating at 15dBm conducted power with up to 15dB gain antennas or WiMAX wireless access equipment operating at up to 23dBm with up to 7dB gain antennas. RLANs operating in this band are licence exempt. C Final Report v1-0.doc Page 15

22 Band C ( GHz) The authorisation regime for FWA in band C is defined in Ofcom document entitled The Authorisation Regime for Fixed Broadband Services operating in the frequency band MHz (Band C). The minimum equipment technical parameters are defined in IR2007, UK Radio Interface Requirement 2007 Fixed Broadband Services Operating in the Frequency Range 5725MHz-5850MHz. The band is subject to a light licensing scheme whereby all terminals must be registered with Ofcom. Applications for authorisation are made on line. The current FWA allocation is GHz and MHz. The 20MHz between 5.795GHz and 5.815GHz is allocated to RTTT. The spectrum from 5.85 to 5.875MHz is not currently allocated to FWA. However, there have been several studies into coexistence between FWA and other systems in this frequency range as summarised in sec. 4. Operation in this band requires the use of DFS and TPC to minimise interference with licenced users of the band. The details of DFS and TPC requirements are specified in Annex 1 of Recommendation ITU-R M There is currently no harmonised European Norm for equipment operating in this band. Therefore, there is a Voluntary National Standard (VNS2107). VNS references EN with regard to technical parameters such as frequency tolerance and testing including DFS. The limits included in IR2007 are: Maximum EIRP 33dBm PSD < 100mW/MHz Restrictions on EIRP spectral density at elevation angles above the horizontal The TPC dynamic range must exceed 19dB relative to 33dBm DFS detection thresholds are: -62dBm for devices with max. EIRP <200mW -64dBm for devices with max. EIRP 200mW to 1W -67dBm for devices with max. EIRP >1W Typically, equipment is (pre 3 ) WiMax Wireless access equipment operating at up to 23dBm with up to 10dB gain antennas. The 2W limit maximum power level in 5.8GHz band for FWA systems arose through consultation with stakeholders who wished to have the same rules in the 5.8GHz band C as FCC UNII rules i.e. 4W. Following discussions with MoD, 3 WiMAX is not allowed to be used for equipment unless it has passed interoperability testing. Testing only started in Late pre WiMAX is used to describe equipment that is basically WiMAX (ie d compliant) but not passed interoperability testing. C Final Report v1-0.doc Page 16

23 JFMG and satellite operators the 2W limit in conjunction with a light licensing regime was agreed as a compromise. 4 The table below summarises the power limits, relevant UK IR documents and ETSI specs. OFCOM Band 2.4GHz Frequency MHz Power limits 5 GHz Band A MHz 100mW EIRP 200mW max mean (over burst) EIRP 5 GHz Band B MHz 1W max mean (over burst) EIRP 5 GHz Band C , , 2W EIRP Relevant UK IR2005 IR2006 IR2006 IR2007 interface doc ETSI specs EN EN EN EN GHz GHz 100mW EIRP Table 3 Summary of licence-exempt frequency band above 1GHz International comparison Table 37 Annex C provides a comparison between regulations in the UK, US and Europe at 2.4 and 5.xGHz. Any changes to the UK regulations should consider the wider availability of equipment so that the UK can benefit from international economies of scale. 3.2 Suitability of 2.4GHz and 5.xGHz frequency bands for the provision of WBA This section provides an initial assessment of the suitability of the frequency bands that could be used for licence-exempt rural WBA systems (primarily 2.4GHz and 5.xGHz in the UK). This assessment takes account of both technical and commercial factors. Basic technical and commercial requirements for WBA systems In order to select the most appropriate bands it is necessary to draw up a list of the key requirements for the future rural WBA system. These requirements should be strongly similar to the requirements for any broadband delivery system, and are set out below: 4 Source: Ofcom. attachment The reason for 2W in the 5.doc from Ahmad Atefi to Andy Rhodes sent 26/4/05 C Final Report v1-0.doc Page 17

24 Criteria Additional Information / Issues High capacity suitable for broadband services Potential for licence exempt use Compatible with widespread adoption ADSL: 1.5Mbit/s downstream, 256kbit/s upstream SDSL: 2 or 4 Mbit/s Other users of the band can tolerate LE applications LE users can coexist Cost of CPEs Cost of infrastructure The QoS provision meets requirements for commercial viability Ease of deployment It should be possible to operate Near Line of Sight (NLoS) propagation Line of Sight (LoS) increases deployment difficulties and restricts technology options LoS likely to need redundant techniques e.g. mesh, multi-hop, SFN to overcome obstacles Equipment supply International harmonisation and standardisation of regulations creates economies of scale and improves availability or cost of equipment Scalable Networks must be easily scalable There must be no need to re-deploy as network grows e.g. repointing of CPE antennas High Base Station utilisation Related to coverage and penetration; Future Proofing and New Services In addition to scalability of the realisable network architectures this requires the capacity to cope with demands for higher bandwidth from the market, and also the capability to support new services such as VoIP or quality of service Customer support issues Additional costs through customer support issues (e.g. through interference) that arise in wireless networks that are not issues for cable/dsl operators Backhaul Extra capacity and different technical features may be required for backhaul or connection to core networks Antenna size and antenna height Roof height antennas will require two installers, increasing cost. Self install, either indoor or outdoor is preferred. C Final Report v1-0.doc Page 18

25 Competition From growing deployment of DSL, satellite, other licenced wireless operators Planning consent Infrastructure; antenna visibility and prime site congestion, possibility for installing CPEs in conservation areas and listed buildings Co-ordination problems with other operators In a licence-exempt environment operators generally need to manage their own co-ordination. QoS QoS is more difficult to maintain in a licence-exempt environment compared with a licenced one Lifeline service and other high reliability services Usually not available, hence no revenue from primary telephone service Overview of suitability of 2.4GHz and 5.xGHz bands for WBA This section sets out the main issues relating to the use of 2.4GHz and 5.xGHz for WBA services. 2.4GHz Spectrum for Licence Exempt Applications Low cost equipment is available in this band due to the proliferation of WLAN technologies such as b, g. There is enough bandwidth for data services, and there are technologies designed for this purpose (up to a theoretical 54Mb/s) which could form part of a future solution for rural WBA, with or without band regulation or equipment modifications. However, there are several issues: Previous experiences of WBA in this band have been characterised by excessive interference leading to poor QoS and high customer support costs Allowing higher power transmission for WBA operators only may improve resistance to interference. With higher power operators have opportunity to effectively trade area coverage with resistance to interference However, higher power for all users may result in many users in the band increasing power, which could cancel out the benefit of being able to trade coverage and interference resistance It is likely that there is insufficient spectrum for operation of a wide area network including backhaul elements using Wi-Fi-based equipment in the 2.4GHz band where there are many other Wi-Fi users. However, in areas where there is little spectrum usage Wi-Fi-based WBA may be able to offer an adequate broadband service. Lack of spectrum will also cause co-ordination problems between co-located services or other operators of equipment in this band. C Final Report v1-0.doc Page 19

26 The band is shared with other users such as the MoD, PMSE, SRDs, AVI and other RLANs. Sharing and coexistence issues will be considered in Annex I GHz Spectrum for Licence Exempt and Light Licenced Applications There are 3 band allocations at 5 GHz compatible with high rate data services in the UK. As at 2.4GHz, there is Wi-Fi equipment available at relatively low cost that can operate in UK 5 GHz bands A and B 5. Also, several vendors currently provide proprietary equipment in band C at higher prices. IEEE standardisation should reduce costs of equipment for this band in the future by creating economies of scale. There are many other types of systems that share spectrum with the three 5 GHz bands, and some of these are primary in status. These include Earth Exploration Satellite Services, Fixed Satellite Services, radar installations and Road Traffic Telematics Systems (RTTT). Hence the coexistence picture will require detailed analysis if any regulatory changes are suggested, or any other steps taken to increase the prevalence of wideband devices in this spectrum. Sharing and coexistence issues will be considered in Annex I Other Bands Spectrum Below 1GHz for Licence-Exempt Applications There are spectrum allocations for licence-exempt applications at 868MHz, 433MHz and below, but these do not have enough bandwidth for broadband services. 17 GHz specific comments There is spectrum allocated at 17 GHz for broadband data systems, however the directional nature of the transmission and its cost limits its applicability to backhaul for the purposes of this analysis. Frequencies above 10 GHz begin to propagate in a strongly directional manner which makes them useful for line of sight applications only. Further, it becomes expensive to build equipment at these high frequencies, and the volumes have tended to be lower than for more consumer-oriented technologies that are available in the lower frequency LE bands. If this frequency band was used for backhaul as an integral element of the rural WBA scheme, or of other similar systems, there would effectively be more spectrum available for the system deployment as a whole. However, 17 GHz is also a MoD band. Therefore any interference issues with MoD use would need to be assessed. 5 Note that a can function in US (U-NII) bands: GHz, GHz, and GHz, and so could potentially operate in UK band C for Fixed Wireless Access C Final Report v1-0.doc Page 20

27 The likelihood of any equipment being made available for WBA in this band in the near future and the likely high cost compared with 2.4GHz and 5.xGHz equipment precludes it from further consideration Summary of frequency bands, regulation and use for FWA Band Current regulatory position for FWA Current Commercial position for FWA <1GHz Allowed Not enough bandwidth for economic use 2.4GHz Allowed at 100mW Used for community self help schemes 5GHz A Not allowed 5GHz B Allowed for Nomadic Hardly used. Some Endpoints at 1W confusion regarding 5GHz C 17GHz Allowed at 2W Light Licence RLAN devices allowed but limited to 100mW EIRP regulations. Strong growth Notes May be many illegal users who have not applied for licence for equipment Table 4 Summary of frequency bands regulation and use for FWA 3.3 Feedback from industry stakeholders on 2.4 and 5.xGHz suitability for WBA This section summarises the main views of interviewees on the suitability of 5.xGHz and 2.4GHz bands for WBA. Fuller details are given in section Interviewee views on the suitability of 5 GHz bands for WBA Interviewees felt that the WISP business case at 5 GHz is marginal. Demonstrating that 5 GHz WBA technology is viable requires: At least a modest power increase (e.g. to 4W EIRP) in band C Changing the band structure in favour of WBA (e.g. releasing the RTTT allocation and aligning the use of WBA in band B and band C), thus making further channels available for interference avoidance Increased international harmonisation (particularly with US regulation) to support overall market success for WBA. C Final Report v1-0.doc Page 21

28 3.3.2 Interviewee views on the suitability of 2.4GHz bands for WBA Interviewed WLAN operators (i.e. hotspot operators and corporate RLAN providers) have serious concerns about allowing higher power in this band: The main concern they raised is that they thought that the use of higher power equipment could not be constrained to rural areas. Their view is that allowing higher power equipment in rural areas leads to more high power equipment availability in the UK which in turn causes leakage of equipment into urban areas. They expected that some operators & DIY home users may use high power equipment anyway, if it is available, regardless of the regulations. Interference in urban areas would therefore increase, which would be difficult to mitigate Interviewees report that in their US operations they are already experiencing significant interference in this band. However, equipment in this band is significantly cheaper than at 5 GHz, and can give better performance (longer range at the same power). For many small community operators, 2.4GHz is the only viable option unless a significant degree of subsidy is available. Some interviewees suggested that technologies such as directional antennas, TPC or smart antennas might allow a power increase in this band without adding to interference. 3.4 Current/Previous usage of spectrum for Licence-Exempt WBA This section presents a brief overview of the deployment of Licence Exempt WBA systems in the UK, covering both 2.4GHz and 5.8GHz systems GHz current/previous usage In the telecoms boom days of the mid 1990 s, new entrant operators saw wireless access technology as a way for them to compete against BT s wireline local loop monopoly. Business plans of the time envisaged rolling out FWA networks in target regions, building a subscriber base by undercutting BT s pricing for basic voice services and selling a range of (usually unspecified) value-added services to these subscribers to increase ARPU. Business models quickly evolved from residentially-focussed plans (e.g. Ionica, albeit using licenced spectrum) to targeting the more lucrative business customers, particularly SMEs. For example in 1995 Atlantic Telecom moved into FWA in Scotland, expanded by making several acquisitions (acquiring customers and access to fibre) and launched itself as a major ISP. The network roll out involved using 2.4GHz FWA technology, at that time licenced to Atlantic but shared with other non-telecoms spectrum users. By 2000 Atlantic had deployed 133 base stations providing ca. 100,000 connections to Scottish cities. The company developed ambitious plans to expand this network, envisaging moving into England and then into Europe. C Final Report v1-0.doc Page 22

29 In the late 1990s many other companies evaluated FWA technology and developed similar plans, including cable operators seeking to infill gaps in their networks and a variety of new entrants. DSL technology was in its infancy, US wireless operators such as Teligent and Winstar had attracted substantial funding, and the future seemed promising. However, despite the business plans looking good on paper, the promise of FWA deployment (and subsequently WBA deployment) failed to be realised in practice. The economics of higher frequency systems proved unattractive expensive base stations required relatively high numbers of customers per sector to break even, and marketing delivered customers scattered across the network. As competition increased the time to payback lengthened, incumbents dropped their wireline prices, investors lost confidence and one by one these operators collapsed. For Atlantic and other operators using 2.4GHz systems things seemed different. There were no expensive spectrum licence fees to pay, and the lower frequency equipment was significantly cheaper. It seemed possible that these systems could be economic at much lower levels of penetration, particularly by targeting SMEs and higher-usage residential customers. However, again there were problems the unlicenced spectrum was subject to interference from other spectrum users. In the backhaul, interference affected all of the customers aggregated into that backhaul link. For CPE, outdoor antennas were subject to interference from other spectrum users and (more importantly perhaps) were impacted by trees and vegetation. This meant additional customer complaints and site visits, driving up operating costs. This proved fatal for Ionica, and ultimately resulted in Atlantic exiting the FWA market. Interest in FWA in the UK only resurfaced with the rise of ADSL. In the early stages of ADSL roll-out considerable pent-up demand was created in areas served by BT exchanges which were not scheduled for ADSL upgrade. Many community-based groups formed, often with RDA support, and turned to 2.4GHz WBA systems as a cost-effective means of getting broadband service. At least 260 such organisations were created, with over 90% using licence-exempt spectrum. Whilst these organisations succeeded in providing service, often on a best-efforts basis, they encountered problems using 2.4GHz equipment. For example Invisible Networks, a Cambridge WISP, encountered external interference to 2.4GHz backhaul links, limited range, signal blocking by vegetation and high cost of connection to an internet backbone network. In addition, early equipment proved unreliable and actual speeds were much less than headline speeds. These problems, together with the perception that ADSL would come soon and give a more reliable, more secure service, slowed down customer growth once those that were desperate for broadband had been signed up. Today the potential for new 2.4GHz WISPs in the UK residential market is limited to areas which are unserved by ADSL. Even if improvement in 2.4GHz network economic performance can be achieved (for example, by using mesh network architectures) it is likely that WISP economics will not allow effective competition with ADSL. In the business market, the potential for interference from other spectrum users means that pure 2.4GHz networks are unlikely to be effective in C Final Report v1-0.doc Page 23

30 providing the required QoS. Thus under current regulations it is likely that use of 2.4GHz equipment will be confined to existing community operators already using the technology to serve residential customers, and commercial WISPs targeting business customers by using 2.4GHz equipment as a cost-effective last drop in a mixed frequency radio network. This latter situation is discussed in more detail below GHz current/previous usage Following the availability of spectrum at 5.8GHz for WBA in the UK in 2003 many commercial and community WISPs have deployed 5.8GHz systems. For example these include commercial operators like Telabria 6, Link Suffolk 7 and Airworks 8, and community-based groups such as CLEO 9 and the Connected Communities Project in Western Isles. These groups have typically deployed 5.8GHz equipment for local distribution via a point to multipoint architecture, and also for point to point links. Many of these operators have relied on public funding to set up their networks, a good indication that the business case is not compelling. Economic issues are the high cost of CPE, range limitations, susceptibility to interference from foliage, low subscriber density in target areas and competition from ADSL. The commercial operators have largely targeted business users, only serving households where subsidies are available to offset high CPE costs or where a 2.4GHz last drop can be used. In view of the economic limitations on a 5.8GHz network under current power regulations described above, the business case is marginal and relies on finding areas which are poorly served by DSL and have a reasonably high proportion of businesses. Thus, new opportunities for 5.8GHz WISPs in the UK are limited. However, this situation could change. Large scale deployment of WiMAX could lower CPE costs, and increased demand for symmetric services could grow the market for WBA significantly (provided current range limitations for SDSL cannot be substantially increased). Under current regulation significant movement in favour of wireless in both of these areas will be needed in order for an attractive business case for 5.8GHz WISPs to develop. 3.5 WBA system technical characteristics and deployment scenarios Two common WBA architectures at 2.4 and 5.8GHz are point-to-multipoint (PmP) and omnidirectional mesh (mesh). Although directional mesh systems exist they are more complex and expensive than omnidirectional systems. The latter have C Final Report v1-0.doc Page 24

31 the advantage that no electronic or mechanical beamforming techniques are required and network configuration is simpler. Directional mesh is sometimes used for high capacity microwave systems operating at higher frequencies. Beamforming at 2.4GHz and 5.8GHz is not practical because of the size of the antenna systems required. Point-to-Multipoint This is the traditional approach and the one that most existing equipment is designed for. There are similarities with Broadcast TV and Cellular telephony. The network operator sets up base sites which are positioned to be able to see a large number of customers premises. There are many different technologies used (FDMA, TDMA, CDMA, OFDMA) to ensure that data from different customers does not interfere. At the customers site a similar type of radio device is used to receive data from the central site and pass that to the computer or other device needing the broadband data and to transmit data from the broadband device to the network operators site(s). A key requirement of this type of system is that since the base site is expensive to install it must be shared amongst many customers to keep the overall system costs down. This in turn requires that the site must be able to cover a large area in turn requiring the use of high power radio transmitters at the customer and operator sites. Current WBA equipment normally uses directional antennas. At the base station these increase the receiver sensitivity (and hence operating range) and they also improve the spectral efficiency by reducing the frequency reuse distances required. At the CPE directional antennas are used to increase operating range and reduce interference susceptibility and emissions. However, they have the significant disadvantage that in most cases they must be professionally installed to ensure they are pointing directly at the correct base site. In addition if the base site has to be moved all the CPE antennas must be realigned at considerable cost. Because of the cost of professional installation, omni directional antennas have often been suggested as a cost reducing approach for WBA. However, the consequence is that the operating distance reduces considerably and so more base sites are required. This significantly increases the cost of providing coverage although it is more neutral where the system is radio capacity constrained where the base sites are required anyway. Using omnidirectional antennas at the CPE effectively makes the CPE s nomadic thus blurring any distinction between nomadic and fixed services. At the base station the use of omnidirectional antennas is theoretically lower cost than directional antennas because less equipment is installed at the site. This should make this approach attractive to serve low density rural areas. However, this equipment is different to that used in the equivalent sectored site as higher C Final Report v1-0.doc Page 25

32 transmit power is normally required. In practise the economies of scale normally make the simpler equipment cost more attractive than the directional antenna equivalent. Mesh The Mesh approach attempts to remove the large expensive basestation. Instead the approach uses enhanced CPEs which can relay data across many radio hops to connect the customers together. The CPE has to have more capacity than the equivalent point to multipoint CPE so it can relay data to several other users as well as its own. In addition, for any radio power that the CPE can achieve, there is a corresponding critical density of CPEs required for the system to operate. Below that density CPEs can see less than two others and so cannot help to relay data. In practise at (or above) this critical density point-to-multipoint currently has the economic advantage. Where omni directional antennas are used the CPE must operate at higher power than the point-to-multipoint CPE to compensate for the loss of antenna gain or alternatively the nodes must be closer together, which is typically the case for a mesh network. Also for easy deployment the mesh system must have very high capacity links, which in turn reduces operating range at a given transmitter power, so further damaging their cost effectiveness in real deployments. Where directional antennas are used the mesh becomes difficult to manage. Every node requires at least two other nodes with directional antennas pointing at it. This creates a significant operational cost when new nodes are added. Different WBA technologies have different regulatory consequences. For example point-to-multipoint has high power base sites. However, there are few of them operated by a single body. Mesh systems have very large numbers of nodes that can generate higher levels of interference because each data packet is transmitted several times as it passes over successive radio links. C Final Report v1-0.doc Page 26

33 3.6 Allocations and other uses of 2.4 and 5.xGHz bands GHz management and allocation The management of the 2.4GHz band and allocations within it are shown in Figure 1. Figure 1 2.4GHz band management and structure GHz usage MoD The MoD is one of the primary users of the 2.4GHz band with an allocation between 2400 and 2450MHz. Military use of this band includes fixed, telemetry and mobile services and long-range airborne telemetry links which are particularly carefully protected. Some of this use is by both the RAF and the USAF for training purposes 10. It is believed that there may be increasing use of this band for Unmanned Aerial Vehicles (UAVs). However, WRC-07 is considering additional spectrum for UAVs and airborne telemetry. The Cave report also states 10 Cave Audit, December 2005 C Final Report v1-0.doc Page 27

34 that future generations of Radio Relay equipment is expected to be designed so that it can operate across this band. The MoD s BOWMAN Personal Role Radio uses this band. PMSE PMSE is analog or digital video transmission equipment for outside television broadcast purposes, such as sporting events or news coverage. The equipment is mobile and often transmits high power. PMSE use of this band is declining. JFMG has provided SG with information showing that the number of applications for licences is beginning to decline despite the growth in PMSE in general (see Figure 2). The decline is believed to be because of the increased potential for interference broadcasters are likely to suffer due to rising numbers of other civil systems such as ISM devices, SRDs and WLANs. 350 Decline in usage of PMSE (Total number of short term licenses per year) GHz 2.4GHz Year Figure 2 Short-term licences issued by JFMG source: JFMG C Final Report v1-0.doc Page 28

35 Wideband Data Systems Existing wideband data transmission systems in the 2.4GHz band fall mostly under the Bluetooth and Wi-Fi banners. Equipment categorised as wideband data must comply with EN and is limited to 100mW EIRP. SRDs Short Range Devices operate in the same frequency band as RLAN equipment, and therefore are in the same band that any higher powered FWA variant at 2.4GHz would be likely to use. They operate at a lower power limit and should be designed with their secondary user status in mind, which means they include interference tolerant features such as frequency hopping and data retries. Recommendation is a reference document provided by the CEPT which provides a guide to spectrum availability within CEPT countries. The CEPT and national administrations aim to minimize the regulations so that there are as few application specific constraints as possible. SRDs are limited to 10mW EIRP Due to the licence-exempt nature of SRDs there is a great diversity of applications. Technical characteristics of systems vary greatly too. RFID RFID (Radio Frequency Identification) is a widely used technique for tagging and tracking goods. There is also expected to be considerable growth of RFID devices in the future. IDTechEx estimate the RFID market to be worth approximately $12billion by Most RFID systems are passive and operate at 13.56MHz. However, active systems are also growing. 2.45GHz is a relatively popular band for active systems because of its worldwide availability. Due to the need for tags to be very low cost active 2.45GHz systems represent a very small minority of all tags compared with passive tags and low frequency active systems at 315 and 433MHz. AVI AVI (Automatic Vehicle Identification) is a specific form of RFID used for monitoring rail rolling stock. If information about the stock is required off-board then the reader will be at the track side. If there is a need for on-board system to know its precise location then the rolling stock may carry the reader. AVI systems in Europe must comply with EN Electromagnetic Compatibility and Radio Spectrum Matters Short Range Devices; Automatic Vehicle Identification for Railways Operating in the 2.45GHz Frequency Range. C Final Report v1-0.doc Page 29

36 Video senders Video senders are commonly used for connecting set-top boxes, VCRs and DVD players to multiple television sets around the home. They usually use analogue modulation and must meet the EN standard for short range devices. Problems with interference from Wireless LANs have been reported. It is anticipated that video senders in the future are likely to be based on IEEE wireless LAN equipment. Task Group e is aiming to enhance the current MAC (Medium Access Control) to enable applications such as voice and streaming audio and video. No reliable market information regarding the installed base or sales of consumer video senders could be found GHz management and allocations The management of the 5GHz bands A and B and allocations within them are shown in Figure 3. The management of 5GHz Band C and allocations within it are shown in Figure 4. C Final Report v1-0.doc Page 30

37 GHz usage Figure 3 5GHz Bands A and B Management and Structure Band A ( MHz) Aeronavigation it is believed that there are no plans to use this spectrum for this application. 12 Fixed satellite service the lower part of band A is used for Mobile Satellite Service (MSS) feeder links. These are mostly LEO satellites. Globalstar claims to be the most widely used handheld satellite phone provider. It operates 13 satellites in this band. ICO filed for bankruptcy in 1999 and emerged from bankruptcy in There is a single MEO ICO satellite recorded as being active in the Handbook of Satellite Services in Europe 13. However, it appears that ICO is not currently offering services. Mobile (incl HIPERLANs or equivalent) - this band is used by IEEE802.11a WLAN equipment. Use is limited to indoors to minimise interference to MSS feeder links. Equipment must conform to EN which includes TPC and DFS requirements. Furthermore, equipment must conform with ECC Decision 12 ECC DEC(04)/08 13 Microcom Systems Ltd, October 2005 C Final Report v1-0.doc Page 31

38 (04)/08. EIRP is limited to 200mW and maximum mean EIRP spectral density is limited to 0.25mW/25kHz in any 25kHz band. Band A ( MHz) EESS there are EESS radars operating in this band such as the Canadian Radarsat and European Envisat. It is considered by the EESS community that the U.S. decision to allow outdoors WLANs in this band will render it unusable for EESS over the U.S. and Canada. 14 Note that Radarsat-2 operates between this band and band B at 5.405GHz and Radiolocation the MoD has radars operating in this band. There are eight different types two of which can be operated in either a fixed frequency or frequency hopping mode. Mobile (incl HIPERLANs or equivalent) this band is used by IEEE802.11a WLAN equipment. Use is limited to indoors to minimise interference to EESS. Equipment must conform to EN which includes TPC and DFS requirements. Furthermore, equipment must conform with ECC Decision (04)/08. EIRP is limited to 200mW and maximum mean EIRP spectral density is limited to 10mW/1MHz in any 1MHz band. Active space Research This band is used for both sensing on planetary missions. It is believed that this band has not been used for sensor applications on any planetary missions. ESA have confirmed that it has not used this band for this application but could in the future for Mars missions. 15. Band B ( MHz) Radiolocation the MoD has radars operating in this band. There are eight different types two of which can be operated in either a fixed frequency or frequency hopping mode. EESS ( MHz) There is believed to be only one device on board an EESS which uses this band. This is the altimeter on the Topex-Poseiden mission. This is a joint CNES and NASA mission to map the ocean topography (sea level height changes) 16. ESA does not have EESSs operating in this band. It is possible that there will be future use of this band by ESA if the outdoor WLAN use in the U.S. prevents the MHz band being used. Maritime Radionavigation ( MHz) The maritime radar application includes shipborne weapons systems radars and Vessel Trafficking System (VTS) radars. ITU-R Rec. M. 629 Use of the radionavigation service of the frequency bands MHz, MHz, MHz, from Edoardo Marelli of ESA to Paul Bearpark, 31 October from Edoardo Marelli of ESA to Paul Bearpark, 23 March from Edoardo Marelli of ESA to Paul Bearpark, 31 October 2005 C Final Report v1-0.doc Page 32

39 9 500 MHz and MHz indicates that "Only a small number of these types of radars operate in the band MHz". Mobile (incl HIPERLANs or equivalent) this band is used by RLANs including IEEE802.11a WLAN equipment. Indoor and outdoor use is allowed. Equipment must conform to EN which includes TPC and DFS requirements. Furthermore, equipment must conform with ECC Decision (04)/08. EIRP is limited to 1W and maximum mean EIRP spectral density is limited to 50mW/1MHz in any 1MHz band. Active Space Research ( MHz) This band is used for both sensing on planetary missions. It is believed that this band has not been used for sensor applications on any planetary missions. ESA have confirmed that it has not used this band for this application but could in the future for Mars missions. 17 Deep Space Research ( MHz) ESA ground stations for communications in this band in Spain, Australia and the U.S.A Amateur and Amateur satellite ( MHz) Amateur users of this band use the spectrum on a secondary basis and therefore should avoid interference with primary users of the band PMSE - PMSE is analog or digital video transmission equipment for outside television broadcast purposes, such as sporting events or news coverage. The main applications is ENG/OB. The equipment is mobile and often transmits high power. There are few wireless cameras operating in this band. It is understood that programme makers using this band do not expect to suffer interference from RLANs in the band despite there being an allocation for outdoor systems operating up to 1W 18. It is thought that sharing has not been a problem because of the low number of outdoor RLANs. Note that this is in contrast to band C where programme makers expect that they may suffer interference from FWA. As an indication of usage Table 5 shows the number of licences that were issued for PMSE use in this band in Frequency MHz MHz MHz MHz MHz MHz MHz Licence type Short term Annual Table 5: PMSE licences issued during 2005 in band B 17 from Edoardo Marelli of ESA to Paul Bearpark, 23 March Conversation with Paul Gill of JFMG, 27 January 2006 C Final Report v1-0.doc Page 33

40 Band C ( MHz) Fixed Satellite Service Figure 4 5GHz Band C Management and Structure All satellites operating in this band are geostationary. Satellites located over Europe and western Asia between 45 West and 90 East are considered during this project as being potentially susceptible to interference from FWA in the UK (see ). The majority of satellites only operate above 5850MHz. The following list shows satellites believed to operate below 5850MHz. Satellite Frequency Range Position Express AM MHz GEO 40.0E Express AM MHz GEO 80.0E Yamal MHz GEO 90.0E Yamal MHz GEO 49.0E Table 6 Satellites operating below 5850MHz Source: Handbook of Satellites in Europe, Microcom Systems, October 2005 C Final Report v1-0.doc Page 34