The IoT is a hugely important and rapidly growing market. The connectivity revolution powered by M2M and the emerging IoT is one of the most important trends in modern technology and is set to transform countless industries. Mobile services play an important role in the wide area M2M and IoT markets. According to the GSMA[35], the bulk of the M2M market uses short-range, unlicensed connections (e.g. WiFi, ZigBee etc.), however the wide area market is heavily reliant on mobile connectivity. Furthermore, as forecasts indicate that the number of IoT connected devices worldwide is set to rise dramatically and reach 26 billion by 2020, the need for IoT devices to be identifiable in the network will persist, as it does today for traditional voice and data devices, although IoT devices are fundamentally different from traditional devices.

However, many of the requirements associated with the use of E.164 (the traditional telephone numbers) ranges are inappropriate for the large majority of IoT connected services. Requirements related to current numbering regulation such as number portability are not relevant with respect to the IoT, as the service does not directly involve an individual and the connectivity element is just an enabler of the entire service wrap. Given that the majority of BEREC NRAs have both spectrum and numbering responsibilities, BEREC considers it important to consider these aspects with respect to any potential future monitoring of the IoT. Therefore, the following sections cover the responses to questions in the supplementary questionnaire on these matters.

The spectrum needs of IoT applications are determined by their throughput requirements, but also coverage, latency and reliability. For a given spectral efficiency (b/s/Hz), the lower the latency requirements the larger the bandwidth needed to send a given amount of data - this becomes very important for applications such as remote surgery. E-health applications often need ultra-reliable connections (security and privacy), and the combination of ultra-robust connections (heavy coding and retransmissions) with high throughput/low latency requirements requires large bandwidths. Spectrum bands suitable for IoT are determined by each IoT application’s range and coverage requirements as well as bandwidth needs of the applications. Range and coverage requirements also depend on deployment scenarios.

A number of NRAs indicated, in their response to the specific question as to what effects on spectrum policy do they expect the development of the IoT to have, that monitoring of spectrum management developments already takes place at an EU level with organisations other than BEREC (for example, CEPT, ETSI, ITU-R). However, while there may not be a necessity to launch a monitoring of technical issues of spectrum usage within BEREC, as this may overlap with these other organisations’ work, best-practice sharing among NRAs could be useful, especially in the case of future 5G-based M2M services, in order to find the best opportunities or solutions for the introduction of 5G.

Some NRAs pointed out that the development of the IoT could have consequences in terms of spectrum policy, for example:

- The allocation of appropriate frequency bands for IoT, either licensed (LTE-M, NB-IoT) or license exempt,

- The adaptation of the technical conditions of use of the frequency bands, to reflect the technological evolutions of IoT.

Regarding the first point above, the bulk of the M2M market uses short-range, unlicensed connections (for example, WiFi, ZigBee etc.). The wide area market is heavily reliant on mobile connectivity; there exist high quality of service guarantees over wide areas, as operators are not at risk of interference and can control usage levels. There is, therefore, a whole portfolio of different use cases and a whole range of different needs for different type of IoTs. In terms of spectrum requirements, provisions have to be made within both the licence exempt frequency band and also within the licensed frequency band.

Monitoring the development of IoT would help the NRAs to adjust their spectrum policy accordingly. One NRA indicated that it has already started the forward-looking process by identifying spectrum ranges suitable for narrowband, wideband, short or long-range IoT applications. Frequency options for IoT applications include public or private mobile networks as well as license exempt frequencies for short range devices. These frequency options enable a very flexible and adaptable environment for IoT applications to address their specific demand. It is expected that IoT applications are an essential driver for the implementation of 5G including network slicing within the limits of identified spectrum for public mobile operators. Monitoring the specific spectrum use of IoT applications may therefore be essentially to define future demands. Furthermore, for capacity planning, it is important to know the estimated amount of data traffic based on technologies where regulation is in place (in particular cellular/mobile technologies or other types of licensed spectrum).

In its supplementary questionnaire to NRAs, BEREC asked the following questions:

- With regard to the expected growth in the use of IoT devices, do you see the necessity for NRAs/BEREC to monitor these developments?

- Do you see the need to monitor which national numbers for IoT devices are used outside your territory (and vice-versa, which numbers assigned in other countries are used in your territory)?

Regarding the first question, in terms of the necessity to monitor the expected growth in the use of the IoT, the response by NRAs was unanimous, in that all indicated a need to monitor these developments. One NRA noted that the expected proliferation of IoT devices may lead to a high demand for national numbering and network resources, hence it is important for national NRAs to monitor developments in the IoT market.

Since numbers are a scarce resource and the use of IoT devices could increase dramatically in the coming years, monitoring these developments is important to identify expected demands as early as possible. NRAs will need to monitor the increased demand of numbering, prepare national plans accordingly and monitor the development of needs in this area.

Regarding the second question, some NRAs indicated the need for more evidence (i.e. the size of such markets) to assess this matter. Reasons pointed out by some NRAs against the necessity to monitor extraterritorial use of national numbers for IoT devices are that numbering rules apply for these numbers are the same whether they are assigned to users domestically or abroad. Reasons in favour of this monitoring indicated are that this information (i.e. information gleaned from monitoring which national numbers for IoT devices are used outside of a specific country’s territory) would be helpful in order to keep an overview of the geographic distribution of the resources.

In addition, such information would be useful with respect to M2M/IoT roaming, which would have implications for potential security issues nationally, in the EU and beyond. It should also be noted that in light of the new EECC, BEREC has been tasked with developing a database of numbering resources with a right of extraterritorial use within the EU. This database could be sufficient to achieve such monitoring. BEREC also asked NRAs as to the relevance of these matters (i.e. monitoring of expected growth of IoT and of extraterritorial numbers) for NRAs and/or for other national authorities. Responses to this question typically noted that while, for the former question, it is of utmost relevance to NRAs, for the latter it is of most relevance to authorities responsible for public security and criminal enforcement/IT security. However, it can also be considered relevant for NRAs for the purposes of statistical analysis, interpretation, and the operation of security-related information tasks.

Based on the elaboration of the boundaries of the IoT, set out above in Figures 2 and 3, some elements of the IoT market may well be considered outside of the scope of classical/current telecoms regulation, which has its roots in the opening of the formerly monopolised/state owned telecoms markets characterised by very high investment cost and high barriers to replication of infrastructure. Historically, the regulation was set up to create competition and to assure that each end-user had access to basic telecom services at affordable cost and at sufficient quality. As long as IoT does not create totally new end-user activities, but “only” refines/automates existing activities (car travel, heating/ventilation/lighting of buildings, medical appliances, assurance of public security, etc.) the existing authorities have to adapt to and deal with IoT in their respective field of activities.

Question 3.1:

In your opinion, what effects on spectrum policy is the development of the IoT expected to have, and do you think it’s necessary for NRAs to monitor, and BEREC to benchmark, these developments?

Answer to question 3.1:

Question 3.2:
With regard to the expected growth in the use of IoT devices, do you see the necessity for NRAs to monitor, and BEREC to benchmark, these developments, particularly with respect to numbering? If so, why?

Answer to question 3.2:

Question 3.3:

Do you see the need for NRAs to monitor which national numbers for IoT devices are used outside their domestic market/territory (and vice-versa, which numbers assigned in other countries are used in the NRA’s territory)? If so, please elaborate.

Answer to question 3.3:

Question 3.4:

In your opinion, in addition to NRAs, for which entities (EU and non-EU) are the following individual matters relevant:
(a) The effect of IoT on spectrum policy

(b) The effect of IoT on scarce resources, i.e. numbering

(c) The monitoring of national numbers for IoT devices used on an extraterritorial basis

Answer to question 3.4:

Between July-October 2018, NRAs submitted information, in response to two questionnaires on the IoT, to BEREC. These questionnaires focused on the current data collection processes of NRAs with respect to the IoT[36] as well as on why the IoT is of importance to BEREC[37], particularly on what type of indicators BEREC should look to collect data on going forward. The two sections below present information, at a high level, on the responses to those questionnaires, with particular emphasis on why NRAs consider such a focus on IoT indicators to be important for BEREC.

According to their responses to BEREC’s questionnaire on current data collection processes with respect to the IoT, most NRAs indicated that they do collect some statistical information on the supply side (i.e., from operators/service providers). However, statistical information collected is almost exclusively related to machine-to-machine (M2M). While one NRA began collecting data on M2M as far back as 2000, typically, the regular collection of statistical information of this type started in 2010, with some NRAs particularly being at the vanguard of such collection. In general, the M2M-related statistics that are collected by NRAs include the number of subscriptions, data volumes and revenues. In addition, some NRAs elaborated on statistical information on the IoT sourced from the demand side (i.e. from consumer or business surveys) but, again, this was limited to M2M type data.[38]

In responding to BEREC’s first questionnaire (see Annex 1) only a small cohort of NRAs indicated that they would collect statistical information on the IoT beyond that which is already being collected (i.e., as set out above, statistical information on M2M) in the short to medium term, i.e. during the next 12-24 months, before the deadline for transposition of the new Code. NRAs weren’t asked to comment on the longer term and their collection of such statistical information. The other responses received were either clearly negative (as in the NRA has no plans in the short-run to collect such information), or that there was uncertainty as to whether such a future collection of statistical information would take place. At the same time, when responding to a question regarding the need for BEREC to benchmark the IoT, a significant number of responding NRAs (more than 10) agreed that there is such a need.

Given that the collection of statistical indicators on the IoT currently focuses on rather general information on M2M, some of the responding NRAs suggested that a step forward would be to just expand the M2M statistics that are collected. This could potentially be achieved by collecting additional data on the specific sectors the M2M services are used within (for example, industry, health, automotive, agricultural etc.). By keeping this information at the sectoral level only, it would not include significant ‘sensitive’ information. However, it should be noted that M2M data presents a very narrow and perhaps random view of the development of the IoT.[39]

Most NRAs have, under current legislation, mainly the right to gather indicators from service providers in the field of electronic communications. Given the new EECC[40], it may be possible to gather data from adjacent sectors to the telecommunications sector. This could open up the possibilities to gather more useful indicators of IoT, beyond the current data covering M2M. The usefulness of such indicators of IoT should be reflected in requests for information, which should, in turn, be proportionate and sufficiently reasoned. In light of this, NRAs also provided responses to a more forward looking questionnaire circulated by BEREC, the responses to which are synthesized in the following section.

In responding to BEREC’s supplementary questionnaire on the IoT, several NRAs elaborated on their reasons for, and provided clarity on the benefit of, a BEREC common approach regarding the IoT. Such possible benefits of a BEREC common approach regarding statistical information on the IoT include:

- A global BEREC approach would be an asset at two levels:

- guide EU Member States in their choices related to players to question and data to collect,

- in having a harmonized approach regarding IoT market at a European level.

- Common approach to benchmarking of IoT indicators could be considered useful for harmonisation of data collection and for sharing experiences/best-practices among NRAs. This could help to make the IoT market more transparent and also could contribute to the development of IoT environment, especially in context of deploying 5G networks in EU.

- A common approach with respect to IoT data collection and benchmarking may prove beneficial since they can be compared on a ‘like with like’ basis across all member states. However, any IoT data requests should be kept as high level as possible.

- In general, it would be of benefit to achieve a common understanding of IoT in all EU Member States. Since IoT products presumably are used EU-wide or even world-wide, the evaluation of the development of the usage and distribution of IoT devices should be implemented at least on EU level. Further to compare IoT markets in relation to traditional ECS-markets, an EU benchmarking would be helpful for providers which are operating in more than one member state.

At the same time, some NRAs responded with some uncertainty on the importance for BEREC to have a common approach with respect to the IoT data collection. In particular, these uncertainties steam from the fact IoT have been outside of the typical competencies of telecoms regulators and, thus, the concrete purpose of collecting IoT indicators needs to be clearly defined (in particular, for those NRAs that would collect these data for the first time). Any data request might have to be kept as high level as possible, at least in the short/medium term. One of BEREC’s main tasks is to improve the consistency of the application of European telecom rules, and contribute to the development of the Digital Single Market and the European Gigabit Society.[41]

Given the possibilities under the new EECC, it would be beneficial to achieve a common understanding of the IoT in all EU Member States. There are, of course, other types of benefits with respect to such harmonisation, for example, common terminology, comparable data, and tracing transnational use of the IoT.

While it is difficult to identify and determine common indicators prior to knowing the extent which the regulation will actually affect the possibilities, this section identifies some suggested potential areas in which it would be beneficial to collect data, if the regulation allowed it. Figure 4 below graphically illustrates the general areas of IoT indicators proposed by multiple NRAs to BEREC, in response to its supplementary questionnaire on the matter.

One suggested area of indicators includes the types of network used for the IoT devices to communicate. Furthermore, several responses suggested that the number of devices, including types of users, are interesting to collect (for example, industrial use or residential devices) as well as which sectors or domains the devices are used in.

While capacity is pivotal for the network, IoT devices affect the network in several ways depending on the type of device. Hence, the total number of IoT devices alone is of less importance if the network impact is not included. This would add a competition perspective and identify how IoT devices strains the potential bottlenecks within the network.

In addition there are other types of volume data that could be interesting to collect beyond the scope of network capacity. Another relevant area for statistical indicators could be prices and price-models for the different IoT networks and services.


Figure 4: General areas of interest for future indicators according to NRAs. Source: BEREC.

According to NRAs, revenues are important for an assessment of the relevance of IoT markets in relation to traditional ECS-markets. Statistics on the number of users (broken down by usage) and the traffic generated are important to develop an understanding concerning the domains of IoT, and needed with respect to providing evidence in regulatory decisions made by NRAs. Statistical information relating to capacity could inform NRAs about connectivity needs.
The statistical indicators proposed for future collection are similar in nature to the suggested way forward with respect to the current M2M data that is collected. However, NRA responses suggest that, in time, there will exist a need for indicators on also machine-to-person (M2P) and person-to-person (P2P). P2P connections are characterized by collaborative solutions that leverage new and existing network infrastructure, devices, and applications. M2P connections mean that people can send information to technical systems and receive information from these systems (for example, receive data and analytics). All of these connections are transactional, which means the flow of information moves in both directions, from machines to people, from people to machines or from machines to other machines. However, it is unclear if such indicators would fall under the current remit of NRAs.

Overall, NRAs suggested that BEREC needs to gather such IoT-related statistical information for some of the following reasons:

- IoT is an emerging market with far reaching possibilities
- Interesting to learn what kind of value added layer emerges and to what extent MNOs try to cover all demand on their own
- Give a correct picture of the mobile markets (not to hide IoT in global SIM card data), to be able to calculate market shares and describe markets in a more detailed fashion
- Understand as to whether IoT could develop towards a European market, or if it (and to what extent) remains a national market
- Add to a basic set of information concerning data economy issues
- Progress towards data-based regulation, opening such data (after anonymisation when required) to external entities
- Having reliable, market-based knowledge is central to the strategy of NRAs.