3.8 GHz to 4.2 GHz band: Opportunities for Innovation Summary Huawei welcomes the opportunity to comment on this important consultation on opportunities for innovation in the 3800-4200 MHz band. We consider the band 3800-4200 MHz to be a key element in the development of 5G, and for the support of wide coverage enhanced Mobile Broadband, massive machine connectivity and critical and real time communications for high reliability or extremely low latency applications. We consider the availability of 400 MHz of additional spectrum adjacent to the band 3400-3800 MHz to be an important element for the establishment of the highest performing 5G networks globally, both in terms of the development of a vibrant device eco-system, and in terms of facilitating the evolution of mobile networks from existing 4G technologies to new air interfaces, with the possibility of making this available to UK citizens cost effectively. We consider that licensed spectrum is essential for mobile networks, including 5G, in order to allow predictable quality of service and end user experience as well as effective management of available resources in a spectrally efficient manner. As such, we are encouraged by Ofcom s approach regarding licensed use of 3800-4200 MHz in the proposed Tier-2, and consider this an enabler for the deployment of 5G in this band. However, we consider that the proposed geographic licences for Tier-2 would result in suboptimal spectrum sharing, and would not allow the full value of 3800-4200 MHz to be exploited. Our preference is for nationwide licences, which might include restrictions in specific geographic areas depending on the frequency of operation. We present a number of approaches for defining such Tier-2 licences. Our preference is that the licenses specify the maximum permitted interference levels as experienced at the input of Tier-1 receivers. Compliance with these regulatory limits, would then be the responsibility of the Tier-2 licensees, with the possible involvement of the incumbents, and potentially with ex ante qualification and/or ex post regulatory oversight. We do not support the inclusion of a third tier for opportunistic use of 3800-4200 MHz, on the grounds that a) the demand for (likely restrictive) opportunistic use is not evident for this band (in light of existing and future spectrum availability at 5GHz), b) it carries risks of harmful interference especially to Tier-2, and c) it overcomplicates the sharing environment. Page 1/11
We encourage Ofcom to consider a two-tier framework in the first instance, and to consider a third tier only when Tier-2 licensed use of 3800-4200 MHz to deliver 5G communications networks in the UK has been well established, and when the cost/benefits of a third tier can be better evaluated. We consider that technology and service neutrality are important cornerstones of progressive spectrum regulation; the market is best placed to decide on the most appropriate technology and service. As such, we consider that the band 3800-4200 MHz should be available for use by all 5G applications, including large-cell and small-cell network deployments, and subject to technical conditions that are least restrictive, and are informed by early coexistence trials. We consider that access to 3800-4200 MHz for 5G would bring significant benefits to UK citizens and consumers, by allowing a wide range of services and applications as we move forward towards a new era in mobile communication. Regulators and stakeholders in industry should carefully assess the bands where spectrum sharing between mobile communications services (IMT) and satellite services might apply. Such assessment should account for the current and forecasted use of the bands by the two services, the social and economic benefits that could derive from the use of the bands by the two services, as well as technology and ecosystem trends of the two services. Such assessment will allow for appropriate trade-offs to be made across the radio spectrum by identifying those bands where exclusive use by mobile communications services (IMT) is most beneficial and should be promoted, those bands which should be used by satellite services, and those bands where sharing between the two services is viable. Page 2/11
Consultation questions and our responses Question 1: Given the nature of the incumbents and their use of the spectrum, what new types of applications do you foresee could access this spectrum on a shared basis? Please provide details on the potential applications and their characteristics of use as identified in the spectrum sharing framework. We consider the band 3800-4200 MHz to be a key element in the development of 5G, and for the support of wide coverage enhanced Mobile Broadband, massive machine connectivity and critical and real time communications for high reliability or extremely low latency applications. We considers the availability of 400 MHz of additional spectrum adjacent to the band 3400-3800 MHz to be an important element for the establishment of the highest performing 5G networks globally, both in terms of the development of a vibrant device eco-system, and in terms of facilitating the evolution of mobile networks from existing 4G technologies to new air interfaces, with the possibility of making this available to UK citizens cost effectively. We consider that licensed spectrum is essential for mobile networks in order to allow predictable quality of service and end user experience as well as effective management of available resources in a spectrally efficient manner. We are encouraged by Ofcom s approach regarding licensed use of 3800-4200 MHz in the proposed TIer-2. We consider that such Tier-2 use should be based on nationwide licences rather than geographic licences. We elaborate on this further in our response to Question 3. We consider that technology and service neutrality are important cornerstones of progressive spectrum regulation; the market is best placed to decide on the most appropriate technology and service. As such, we consider that the band 3800-4200 MHz should be available for use by all 5G applications, including large-cell and small-cell network deployments, and subject to technical conditions that are least restrictive. Question 2: Based on information provided in this Section, can you identify any barriers to enhanced sharing in the 3.8 GHz to 4.2 GHz band? Please use the Spectrum Sharing Framework, which identifies four types of barriers to spectrum sharing: lack of information; market barriers; technology barriers; and authorisation barriers. As highlighted before, we consider licensed access to the band 3800-4200 MHz to be important for future 5G networks. We see the following barriers in the context of a two-tier sharing framework, where 5G communication networks use 3800-4200 MHz through the licensed Tier-2. We are not supportive of a three-tier sharing framework. See our response to Question 3. Lack of information We acknowledge that lack of information about the use of spectrum by incumbents has been one of the major barriers to spectrum sharing in the past. We acknowledge Ofcom s realisation of this issue, and welcome its initiatives to make such information available for the 3800-4200 MHz band. Page 3/11
Ofcom s online interactive visualisation tool regarding the use of the band by the Fixed and Fixed Satellite services is certainly a step in the right direction. The availability of the raw underlying data on the characteristics of each individual existing assignment would greatly facilitate sharing studies in this band. Importantly, information on the received wanted signal power in individual existing assignments (reflecting their margin above sensitivity) would enable analysis of the possibilities for efficient spectrum sharing mechanisms in this band. Market barriers We consider that regulatory uncertainty about the future conditions for the use of 3800-4200 MHz can be a substantial barrier to investments in 5G infrastructure by Tier-2 licensees in this band. We welcome Ofcom s proposals for managing new Tier-1 (incumbent) use of the band following the award of Tier-2 licences. See our response to Question 4. We consider that uncertainty about the potential for harmful interference from Tier-3 equipment will have a similar negative impact on investments in 5G by Tier-2 licensees. For this and other reasons described in our response to Question 3, we do not support a threetier sharing framework. Technology barriers We believe that advanced sharing technologies, such as database assisted access, can play an important role in enabling efficient spectrum sharing between Tier-2 and Tier-1. We note that efficient spectrum sharing is more viable when the potential victim receivers are few in number, and have fixed and known geographic locations. As such, we do not envisage any technology barriers for such a two-tier sharing framework. For a toolbox of options for implementing spectrum sharing between Tier-2 and Tier-1, see the draft ECC report developed at PT1 on Operational guidelines for spectrum sharing to support the implementation of the current ECC framework in the 3600-3800 MHz range. Similar concepts apply to 3800-4200 MHz, as outlined in our response to Question 3. However, we consider that technologies for effective spectrum sharing with victim receivers that have a high degree of mobility are as yet not proven. For this and other reasons described in our response to Question 3, we do not support the introduction of Tier-3 equipment for opportunistic use of 3800-4200 MHz. We consider that the introduction of Tier- 3 would inevitably lead to a less predictable quality of experience for Tier-2 end users. Furthermore, we consider that the market (and not the regulator) is best placed to devise the most appropriate spectrum sharing technologies. For this reason, our recommendation is that the regulator has to the extent possible minimal involvement in the implementation and on-going operation of spectrum sharing technology solutions in this band. This would also help reduce any regulatory delays in the availability of the 3800-4200 MHz band. Authorisation barriers We acknowledge the difficulties of coexistence studies and the challenges facing the regulator in defining least restrictive licence conditions for new users while achieving a reasonably low likelihood of harmful interference for existing users. Page 4/11
We recommend that Ofcom undertakes coexistence field trials at an early stage in its consultation process, preferably in collaboration with interested stakeholders connected to both tiers 1 and 2. Such trials would identify the viability of efficient spectrum sharing, help calibrate theoretical coexistence models, provide guidance on the most suitable spectrum sharing solutions, and reduce the timelines (otherwise extended by endless paper studies) in deriving effective authorisation rules and regulatory restrictions which might apply to Tier-2 in the 3800-4200 MHz band. Question 3: Do you agree with our initial assessment of a potential application of a tiered authorisation approach in this band? If yes, please provide as much detailed information as possible of how you consider any tiered authorisation approach may enable greater spectrum sharing and how it could be implemented in practice. If no, please describe the spectrum access method that you consider may best meet any requirements you have to access spectrum in the 3.8 GHz to 4.2 GHz band. Please give specific details of how you would envisage this working in practice, where appropriate with reference to the tools and enablers identified in the Spectrum Sharing Framework. We support a two-tier sharing framework We agree with the potential for a two-tier sharing framework between Tier-1 incumbents and Tier-2 new licensees and consider that licensed authorisation of the band in Tier-2 can be an important facilitator for the deployment of 5G networks in the UK. Importantly, we consider that such Tier-2 use should be based on nationwide rather than geographic licences, and that it should be subject to least restrictive technical conditions which avoid over-cautious assumptions on the risk of harmful interference to Tier-1. We consider that spectrum use by Tier-2 equipment can be guided through the use of advanced database technologies to achieve efficient spectrum sharing and to extract greatest value from the 3800-4200 MHz band. We elaborate on these points later in our response to this question. We do not support a three-tier sharing framework We caution against a three-tier approach in the 3800-4200 MHz band. We consider that the inclusion of a third tier for opportunistic use of the band would be inappropriate for the following reasons: 1) Uncertain demand for opportunistic access Tier-3 spectrum access for 3800-4200 MHz is aimed at equipment which might opportunistically exploit white spaces that are unused by tiers 1 and 2. In all likelihood, such equipment will be licence-exempt (they would fall under Tier-2 if they were licensed) and be based on IEEE 802.11 derivatives, and other proprietary RLAN-like technologies. We are not aware of a great demand for opportunistic access to spectrum for use by such equipment in 3800-4200 MHz. We note that assuming that the Tier-1/2 sharing framework for 5G and the incumbents is implemented efficiently, there is unlikely to Page 5/11
be much unused white space spectrum available for opportunistic access by Tier-3. And where such white spaces do exist, their use is likely to be subject to stringent technical conditions to protect tiers 1 and 2. On the other hand, we are aware of initiatives in Europe and US to make large swathes of spectrum at 5GHz (5150-5920 MHz in Europe) available for licenceexempt equipment. Given the large amount of spectrum being considered at 5GHz, the existing licence exempt equipment ecosystem at 5GHz, and the likely very restrictive conditions for Tier-3, we see little merit in opportunistic use of 3800-4200 MHz, and consider it to be contradictory to the aim of efficient use of the radio spectrum due to its potential impact on Tier-2 (see below). 2) Risk of harmful interference We consider that the introduction of Tier-3 devices for opportunistic access would carry a substantive risk of harmful interference, especially to Tier-2. This is because while the Tier-1 incumbents in the band use fixed receivers at known locations, this is not the case for Tier-2 users which will include 5G mobile user equipment, and possibly consumer-deployed 5G small-cell base stations. This means that while Tier-2 equipment can benefit from database assisted spectrum access technologies to avoid harmful interference to fixed Tier-1 receivers, this is considerably more challenging in the context of interference from Tier-3 to Tier-2. The efficient protection of mobile/nomadic Tier-2 equipment from the radiations of Tier-3 equipment is by no means trivial and is unproven. Also, as noted earlier, Tier-3 equipment will most likely be licence exempt (they would belong to Tier-2 otherwise). It is well understood that licence exempt radio equipment pose greater challenges in terms of ex post regulatory intervention to mitigate reported cases of interference caused by faulty or non-compliant use of the equipment. For this reason, the merits of Tier-3 equipment in this band and the impact on 5G networks in Tier-2 should be carefully assessed prior to any consideration of authorisation. 3) Complexity The efficient sharing of spectrum and effective management of harmful interference between two different services is not trivial. The complexity of achieving an efficient introduction of Tier-2 use to the 3800-4200 MHz is exacerbated by the existence of three different incumbents in the band. We consider that the sharing scenario in a two-tier framework will already be complex enough, and will require advanced database assisted technologies to allow the efficient use of the spectrum resource. We consider that the simultaneous introduction of a third tier would over complicate the sharing environment, and would delay the implementation of sharing solutions in 3800-4200 MHz. Such delays would reduce the value of the spectrum, and negatively impact the benefits of 5G communication systems (Tier-2) to the UK citizens and consumers. Page 6/11
The era of efficient spectrum sharing is still in its infancy. For this reason, we believe it is prudent that simpler sharing frameworks should be considered first, with more advanced frameworks considered at a later stage, aided with the availability of mature interference mitigation technologies (e.g., sensing, high-order beam steering, rapid frequency agility, etc) and with increased experience by regulators and operators in the implementation of spectrum sharing frameworks. For the above reasons, we encourage Ofcom to consider a two-tier framework in the first instance, and to consider a third tier only when Tier-2 licensed use of 3800-4200 MHz to deliver 5G communications networks in the UK has been well established, and when the cost/benefits of a third tier can be better evaluated. Implementation of a two-tier framework A two-tier spectrum sharing framework for 3800-4200 MHz can be designed in a variety of ways. However, it is important that the framework results in least restrictive technical licence conditions for Tier-2, based on a pragmatic assessment of the risk of harmful interference to Tier-1. We consider that this is essential to allow the efficient sharing of spectrum, and to realise the full value of the 3800-4200 MHz band. We acknowledge that the design of a two-tier spectrum sharing framework may vary across countries to account for national circumstances. We do not consider this to be problematic, so long as the implementation of the framework for example, the way information relating to any potential restrictions is conveyed to the Tier-2 licensees can follow standardised interfaces and technologies. An example of these is Licensed Shared Access, which is being standardised at ETSI. Here we describe a number of options for the design of a two-tier spectrum sharing framework for the 3800-4200 MHz band. Tier-2 licensing: Nationwide vs. geographic In this call for input, Ofcom proposes the use of geographic licences for Tier-2, implying that the operation of the licensees would be limited to specific geographic areas. While this might be a simple approach, we believe that it would result in sub-optimal spectrum sharing, and would not allow the full value of 3800-4200 MHz to be exploited through national coverage of 5G networks. We consider that a preferred approach would be the use of nationwide licences, which might include restrictions in specific geographic areas depending on the frequency of operation. We consider that such nationwide licences would allow the MNOs to use the available spectrum in a more efficient way by adapting their network deployments in geographic areas which would otherwise not be available at all under geographic licences. Tier-2 license conditions As is common practice, the Tier-2 licences would include regionally/globally harmonised technical conditions such as block edge masks, as defined by the CEPT, and often mandated by the European Commission. Page 7/11
The Tier-2 licences would include directly or indirectly additional location-specific and frequency-specific restrictions on the radiated powers of Tier-2 transmitters in order to manage the risk of harmful interference to the (fixed) Tier-1 receivers. The extent of these power restrictions would be a function of the geographic and frequency separation between the Tier-1 transmitters and Tier-2 receivers, as well as other relevant parameters including Tier-1 and Tier-2 equipment deployment geometries, antenna gains, patterns and pointing angles, as well as propagation conditions. We see three broad approaches for the specification of these additional technical licence conditions: a) In a first approach, the licences specify the maximum permitted interference levels 1 from Tier-2 transmitters as experienced at the input of Tier-1 receivers. Compliance with these regulatory limits, for example by appropriately adjusting the EIRP of equipment and/or their deployment characteristics, would then be the responsibility of the Tier-2 licensees, with the possible involvement of the incumbents, and with ex ante qualification and/or ex post regulatory oversight. The licences may also prescribe other relevant technical parameter values which must be used in assessing compliance with these regulatory limits 2. b) In a second approach, the licences specify a coordination mechanism involving the regulator, through which the maximum permitted EIRP of the Tier-2 equipment could be calculated 3. This would account for the operating frequency, location, and deployment characteristics of individual Tier-2 equipment. 1 The limit on received interference can be specified either as the maximum permitted effective cochannel interference power I(0), or the maximum permitted electric field strength E( f), where f is the interferer-victim frequency separation. I(0) would be specified in units of dbm/(b 1 MHz) at the input of a Tier-1 receiver (post antenna) of bandwidth B 1. One can use adjacent channel interference ratios ACIR( f) to work back from interference I(0) to the maximum permitted received interferer power P Rx ( f) from Tier-2, and then use coupling gain G (combined effect of receiver antenna and propagation) to work back from P Rx ( f) to the maximum permitted Tier-2 equipment EIRP P( f). Note that compliance with I(0) would allow the gain and angular discrimination of the Tier-1 antenna to be accounted for. Alternatively, E( f) would be specified in units of db V/m/(B 2 MHz) over the air at the location of the Tier-1 receiver antenna. Note that B 2 is the bandwidth of the Tier-2 transmitter. E( f) can be derived from I(0) as a function of frequency f, ACIR( f), and receiver antenna gain. One can then use propagation gain G Prop to work back from E( f) to the maximum permitted Tier-2 equipment EIRP P( f). E( f) is typically used in circumstances where the characteristics of the receiver antenna (gain, directionality, and pointing angle) are not known for each individual victim receiver. As such, instead of specifying I(0) at the input to a victim receiver, the regulator specifies E( f) over the air at the input of the victim receiver antenna, thereby avoiding the need to consider the receiver s antenna characteristics. E( f) is also used in circumstances where the geographic location of the victim receivers cannot be disclosed. For this reason, E( f) is often applied over a protection zone, which represents the geographic area over which the victim receivers might be located. Finally, note that in principle, different limits on the received interference can be specified for each individual incumbent receiver to account for the receivers margins of operation above minimum sensitivity. 2 Examples include values of adjacent channel interference ratio, which describe how an interferer s signal power P Rx ( f) received at a certain frequency offset translates to effective co-channel interference I(0) experienced by the victim receivers. 3 This would be somewhat similar to the process which is applied to UK Boadband. Page 8/11
c) In a third approach, the licences directly specify the maximum permitted EIRP of the Tier-2 equipment. These would also depend on the operating frequency, location, and other deployment characteristics of the Tier-2 equipment. However, the location and deployment characteristics of equipment would be accounted for with a coarse granularity in deriving the regulatory EIRP limits. For example, locations would be described in the form of areas/zones, and typical nominal values would be assumed for characteristics such as height and antenna gain/angular discrimination. The above options have different implications in terms of efficient spectrum sharing, complexity, flexibility for Tier-2 licensees, and the burden on the regulator in implementing the spectrum sharing framework. These are summarised in the following table. a) b) c) Licences specify mechanism for regulator to specify restrictions for individual Tier-2 equipment Licences only specify maximum permitted received interference power levels at Tier-1 receivers Licences specify restrictions over wide geographic areas based on nominal characteristics of Tier-2 equipment Efficient use of spectrum Complexity of specifications Flexibility for Tier-2 licensees Burden on regulator for on-going operation of framework Our preference is for approach (a). We consider that the Tier-2 licensees (MNOs) are best placed to account for the specifics of their networks, and to ensure their most efficient use of the spectrum while complying with the regulatory restrictions for the avoidance of harmful interference to Tier-1 receivers. With the Tier-2 licensees responsible for compliance with regulatory limits on experienced interference at Tier-1 receivers, the burden on the regulator for the on-going implementation of the sharing framework is minimised. In approach (a), we prefer that the maximum permitted received interference is specified in the form of power levels at the input of the Tier-1 receivers (post antenna), rather than as electric field strengths over the air at the locations of the Tier-1 receiver antennas. This would allow the gain, angular pattern and pointing angle of individual Tier-1 receiver antennas to be exploited in the process of compliance with the regulatory limits, thereby resulting in less stringent restrictions (see Footnote 1). Figure (1) illustrates the information required for the derivation of the restrictions on Tier-2 equipment. In the case of Approach (a), the calculations would be performed by the Tier-2 licensees. It is important that the interfaces for the communication of the relevant information to the Tier-2 licensees are based on standardised technologies such as Licensed Shared Access being specified at ETSI. This would avoid the costs associated with bespoke implementations in different countries. Page 9/11
Figure (1): Information required for the derivation of regulatory restrictions on Tier-2 equipment. In Approach (a), the calculation of any restrictions would be performed by the Tier-2 licensees, with the possible involvement of the incumbents, and with ex ante qualification and/or ex post regulatory oversight. Finally, we note that the sharing framework between Tier-2 licensees and UK Broadband may be quite simple given that they will both use the spectrum for Electronic Communication Services. Question 4: Should a potential future tiered authorisation approach to spectrum access in the 3.8 GHz to 4.2 GHz band accommodate changes from incumbent services of the spectrum? I.e. should new licences or variations to existing fixed link and satellite earth station licences be allowed to continue on a first-come-first-served co-ordinated basis? We consider that certainty in relation to the stability of the spectrum sharing and interference environment for Tier-2 will be essential in promoting investment in the UK s 5G mobile network infrastructure. The absence of such certainty would hugely diminish the value of this band in providing future 5G services to the UK citizens and consumers. For this reason, we welcome Ofcom s proposal for Tier-2 which states that 3.7.2. Tier 2 / New Geographic Licence Layer would comprise geographic licences, as described in Section 3.3.1. Once a geographic licence has been acquired, future Tier 1 licences would need to be coordinated around these licences. We understand Ofcom s proposal to mean that once Tier-2 licences have been awarded, any new Tier-1 (incumbent) use of the band would only be authorised subject to the avoidance of harmful interference to the Tier-2 licensees use of the band. We re-iterate our view that Tier-2 should be authorised via nationwide licences (possibly with restrictions in specific geographic areas) rather than geographic licences. We believe Ofcom s proposal in 3.2.7 strikes a reasonable balance between the need for regulations to provide certainty to the Mobile industry for investment in 5G infrastructures, Page 10/11
and the possibility of some growth in incumbent Fixed and Fixed Satellite services in 3800-4200 MHz. Having said this, we consider that regulators and stakeholders should carefully examine the issue of spectrum sharing between mobile communications services (IMT) and the satellite services across a broad range of bands. Such assessment should account for the current and forecasted use of the bands by the two services, the social and economic benefits that could derive from the use of the bands by the services, as well as technology and ecosystem trends of the two services. Such assessment will allow for appropriate trade-offs to be made across the radio spectrum by identifying those bands where exclusive use by mobile communications services (IMT) is most beneficial and should be promoted, those bands which should be used by satellite services, and those bands where sharing between the two services is viable. Page 11/11