Chapter 14

IoT technologies in unlicensed spectrum

Abstract

This chapter provides an overview of connectivity solutions for Internet of Things (IoT) applications and services. The focus lies on technologies that operates in unlicensed spectrum. The characteristics of using unlicensed spectrum are presented together with an introduction of short-range radio and low power wide area technologies that operate in unlicensed spectrum.

Keywords

BLE; Cellular IoT; CEPT; ETSI; FCC; IEEE 802.11; IEEE 802.15.4; Low power wide area (LPWA); Unlicensed spectrum; Wi-Fi; Wireless sensor networks; WSN

1. 14.1 Operation in unlicensed spectrum 
   1. 14.1.1 Unlicensed spectrum regulations 
   2. 14.1.2 Coexistence in unlicensed spectrum 
2. 14.2 Radio technologies for unlicensed spectrum 
   1. 14.2.1 Short-range radio solutions 
      1. 14.2.1.1 IEEE 802.15.4 
      2. 14.2.1.2 BLE 
      3. 14.2.1.3 Wi-Fi 
      4. 14.2.1.4 Capillary networks 
   2. 14.2.2 Long-range radio solutions 
      1. 14.2.2.1 LoRa 
      2. 14.2.2.2 Sigfox 
3. References 

14.1. Operation in unlicensed spectrum

14.1.1. Unlicensed spectrum regulations

Cellular Internet of Things (IoT) networks, such as EC-GSM-IoT, LTE-M, NB-IoT, LTE URLLC and NR URLLC operate in licensed spectrum. This means that mobile network operators have acquired long-term spectrum licenses from regulatory bodies in the country/region after, for example, an auction process. Such licenses provide an operator with exclusive spectrum usage right for a carrier frequency range. Such spectrum licenses may also be combined with an obligation to build out a network and provide network coverage and communication services in a certain area within a certain time frame. This obligation in combination with the cost of the license motivates mobile network operators to invest upfront into a network infrastructure. This exclusive spectrum usage right provides the prospect of good financial returns on the investment obtained via communication services within the lifetime of the license. There are also other spectrum bands, which do not abide to the rules of licensed spectrum. In unlicensed or license-exempt spectrum any device is entitled to transmit as long as it fulfills the regulation without requiring any player from holding a license. These regulatory requirements have the objective to harmonize and ensure efficient use of the spectrum.

Unlicensed spectrum bands differ for different regions in the world. In the following an overview of the usage of unlicensed spectrum is provided for two bands, one at around 900   MHz and one at 2.4   GHz. These are of particular relevance due to their ability to cater for IoT services and popular wireless communication standards such as Wi-Fi and Bluetooth have been specified for these bands. The sub-GHz range around 900   MHz provides attractive propagation characteristics in terms of facilitating good coverage. The 2.4   GHz range is interesting because it is considered to be a global band, which is important for systems targeting a global footprint. While the 2.4   GHz band is globally harmonized, the sub-GHz range has regional variations. However, most regions have some unlicensed spectrum allocation even if they differ in their specifics. A more detailed description is here provided for the US unlicensed spectrum at 902–928   MHz and the European unlicensed spectrum at 863–870   MHz. In Europe, some differences in the allocations of the 863–870   MHz band exist on a per country basis. There has been significant market traction for IoT connectivity solutions operating in the unlicensed frequency domain in these two regions. In other regions, the unlicensed spectrum allocation in the sub-GHz range varies for different countries. For example, the allocations in Korea and Japan are overlapping with the US spectrum region, and China has an allocation that is below the European spectrum allocation, see e.g., Ref. [1]. Radio technology standards that are addressing the unlicensed spectrum around 900   MHz, such as, IEEE 802.11ah, are typically designed in a way, in which they provide a common technology basis for different channelization options in this spectrum range; the detailed channelization is then adopted to the region where it is deployed, see e.g., Ref. [1] for the channelization of IEEE 802.11ah. IEEE 802.11ah is the basis upon which Wi-Fi HaLow is built.

For the United States, the usage of unlicensed spectrum for communication devices is regulated by the Federal Communications Commission and it is specified in Title 47 Code of Federal Regulations Part 15 [2]. For Europe, the spectrum rules are specified by the European Conference of Postal and Telecommunications Administrations (CEPT), which is a coordination body of the telecommunication and postal organizations within Europe. As of today, 48 countries are members of CEPT [3]. The CEPT recommendation for usage of short-range devices in unlicensed spectrum is described in Ref. [4]. This recommendation implements the European Commision decision on frequency bands used for short range devices [50]. It serves as the basis for European Telecom Standards Institute (ETSI) harmonized standards, which specify technical characteristics and measurement methods for devices that can be used by device implementers to validate their devices for conforming with the regulator rules. Such ETSI standards are as follows:

1. • ETSI standard EN 300   220 for short-range devices operating in the range 25   MHz–1   GHz [5,6],
2. • ETSI standard EN 300   440 for radio equipment to be used in the 1–40   GHz frequency range [7,8],
3. • ETSI standard EN 300   328 for data transmission equipment operating in the 2.4   GHz ISM band and using wide band modulation techniques. Direct-sequence spread spectrum (DSSS), frequency hopping spread spectrum, and Orthogonal Frequency-Division Multiplexing are considered to be wide band modulation techniques [9].

Some of the more relevant spectrum bands in the unlicensed spectrum range 863–870   MHz in Europe are given in Table 14.1 and for 902–928   MHz in the United States in Table 14.2. The tables present, e.g., the maximum allowed radiated power and requirements for interference mitigation. While the ETSI regulations in the band 863–870   MHz mandate the power in terms of Effective Radiated Power (ERP), i.e., the radiated power assuming a half-wave dipole antenna, the Federal Communications Commission sets requirements in terms of conducted power in combination with allowed antenna gain. Here we have converted these requirements to Equivalently Isotropically Radiated Power (EIRP), i.e., the radiated power assuming an isotropic antenna, according to the following equation:

EIRP = ERP + 2.15 dB

(14.1)

The 2.15   dB offset stems from the different reference antennas, i.e., dipole and isotropic, assumed for EIRP and ERP. It can be noticed that the permitted radiated power is higher for the unlicensed spectrum in the United States than in Europe. The power requirement is in some cases defined as a peak power requirement, and in some cases as an average power requirement. For a system using a modulation with a significant peak-to-average power ratio the peak power requirement will impact the average output power and the system coverage. This aspect is in detail discussed in Sections 15.2.3.3 and 15.3.3.2.

The European recommendations rely on duty cycle limitations, i.e., the percentage of time a transmitter may be active within a defined time span, while the US regulations defines maximum dwell times, i.e., the maximum continuous time by which a transmitting device may use a specific radio resource (e.g., a specific hopping channel). Regardless whether the limitations are on duty cycle or dwell time, these limits aim to avoid persistent interference. A short allowed dwell time can be understood to limit the coverage of a system, whereas a strict duty cycle requirement may limit coverage and system capacity. The duty cycle also limits the availability of downlink and uplink transmission opportunities which has a negative impact on service latency.

Table 14.1

European unlicensed spectrum at 863–870   MHz, for more details see Refs. [4,50].

Spectrum band	EIRP	Mitigation requirement	Bandwidth	Other
863–870   MHz	16.1   dBm	<0.1% duty cycle or LBT		FHSS
863–870   MHz	16.1   dBm	<0.1% duty cycle or LBT   +   AFA		DSSS and other non-FHSS wideband techniques
863–870   MHz	16.1   dBm	<0.1% duty cycle or LBT   +   AFA	≤ 100   kHz, for 1 or more channels; modulation bandwidth ≤ 300   kHz
868–868.6   MHz	16.1   dBm	<0.1% duty cycle or LBT   +   AFA
868.7–869.2   MHz	16.1   dBm	<0.1% duty cycle or LBT   +   AFA
869.4–869.65   MHz	29.1   dBm	<10% duty cycle or LBT   +   AFA
865.6–865.8   MHz	29.1   dBm	10% duty cycle for network access points ≤ 2.5% duty cycle for other SRDs	≤ 200   kHz	Adaptive power control
866.2–866.4   MHz
866.8–867.0   MHz
867.4–867.6   MHz
869.7–870   MHz	9.1   dBm 16.1   dBm	For 4.8   dBm: No requirements; For 11.8   dBm: ≤ 1% duty cycle or LBT   +   AFA

Table 14.2

US unlicensed spectrum at 902–928   MHz, for more details see Ref. [2].

Spectrum band	EIRP	Mitigation requirement	Bandwidth	Other
902–928   MHz	36   dBm	Dwell time per hopping channel: < 0.4   s/20   s	≤ 250   kHz	Frequency hopping with ≥50 hopping channels
902–928   MHz	30   dBm	Dwell time per hopping channel: < 0.4   s/10   s	200–500   kHz	Frequency hopping with ≥25 hopping channels
902–928   MHz	36   dBm		≥ 500   kHz	Digitally modulated

European recommendation allows to deviate from duty cycle limitations in case listen-before-talk (LBT) is used, i.e., when a transmitter first observes if a channel is not used by other devices before starting to transmit. Sometimes LBT needs to be combined with adaptive frequency agility, which is a method for avoiding transmission on already occupied channels (see e.g., Ref. [1]). LBT introduces an uncertainty in the systems transmissions that makes it difficult to provide high reliability and bounded latency.

The different rows in Tables 14.1 and 14.2 can be understood to cater for different types of applications and different types of equipment. For more details please see Refs. [2,4].

The spectrum usage recommendation for unlicensed spectrum at 2400–2483.5   MHz in Europe is given in Table   14.3 and the rules for the United States in Table   14.4. Also in this band the maximum allowed radiated power is higher in United States than in Europe. The European power limitation is given as EIRP. Higher radiated power is allowed for wide band transmission such as DSSS modulation, FHSS, and Orthogonal Frequency-Division Multiplexing under the condition of using a mitigation method such as LBT.

14.1.2. Coexistence in unlicensed spectrum

Different types of radio equipment and technologies can transmit on any frequency of the unlicensed spectrum. As depicted in Fig. 14.1, the simultaneous transmissions of different devices interfere with each other. As a result, this interference may lead to that one or both of the transmissions depicted in the figure fail. Spectrum coexistence mechanisms are mechanisms that limit the interference that a transmitter may cause on other nearby devices.

Fig. 14.1 Transmission of two different unlicensed devices to their corresponding receivers.

The simultaneous usage of unlicensed spectrum can occur between homogenous or heterogeneous types of devices, which means devices that use the same or different wireless communication technologies. A wireless communication technology designed for unlicensed spectrum, often, has some mechanism, which specifies how the spectrum is shared among different devices of that communication technology, so that each device has good transmission opportunities, while at the same time minimizing the interference to other devices. The coordination scheme typically follows the spectrum regulation introduced earlier but may also have additional technology specific features. While such a coordination scheme can be applied within one wireless communication technology, it can typically not provide the same level of interference mitigation in-between different wireless communication technologies.

The spectrum regulation for unlicensed spectrum provides requirements on devices, which shall provide technology neutral coexistence, i.e., independent of a particular wireless communication technology being used. Spectrum regulation has thereby mainly two types of communication devices in mind: adaptive devices and nonadaptive devices.

Nonadaptive devices are considered to transmit in unlicensed spectrum, while staying unaware of the other types of devices. To limit the amount of interference that can be generated, the devices are limited as follows:

1. • In the radiated power they may use, see Tables 14.1–14.4, sometimes further limited by an allowed power spectral density,
2. • By a duty cycle or dwell time, see Tables 14.1 and 14.2, which limit what fraction of time a device may transmit on a channel.

By these means the total amount of interference a device may cause on others is restricted.

In contrast, an adaptive device is aware of other devices in its vicinity, which also make use of the same channel. As a result, it can adapt its transmission, reducing the interference to other devices. This provides a device with the opportunity to transmit for a longer amount of time if there are few devices in the surrounding. In contrast, if many devices in the vicinity want to use the unlicensed spectrum, the adaptive device reduces the frequency of its transmissions and provides less interference. The common way to provide this type of adaptive device is LBT, where a device before a transmission listens on the radio channel and observes if other devices are communicating, and after a clear channel assessment it transmits for a limited time. In addition, technical standards consider a fair use of spectrum among devices, and thus consider that at a high utilization of the spectrum an adaptive device transmits less than in case of a low utilization.

Table 14.3

European unlicensed spectrum at 2400–2483.5   MHz, for more details see Ref. [9].

Spectrum band	EIRP	Mitigation requirement	Other
2400–2483.5   MHz	10   dBm
2400–2483.5   MHz	20   dBm	Spectrum sharing mechanism like LBT	For wide band data transmission and radio local area networks. For wide band modulations other than FHSS, the maximum EIRP-density is limited to 10   mW/MHz

Table 14.4

US unlicensed spectrum at 2400–2483.5   MHz, for more details see Ref. [2].

Spectrum band	EIRP	Mitigation requirement	Bandwidth	Other
2400–2483.5   MHz	30   dBm	Dwell time per hopping channel: < (0.4   s ∗ X), X   =   number of hopping channels	≤ 250   kHz	Frequency hopping with ≥15 and   <   75 non-overlapping hopping channels
2400–2483.5   MHz	36   dBm	Dwell time per hopping channel: < (0.4   s ∗ X), X   =   number of hopping channels	200–500   kHz	Frequency hopping with ≥75 non-overlapping hopping channels
2400–2483.5   MHz	36   dBm and max 8dBm/3   kHz PSD		≥ 500   kHz	Digitally modulated

The most common form of unlicensed spectrum usage is for short-range communication. One reason is that short-range communication provides a certain level of interference robustness. The amount of interference that a receiver is exposed to depends very much on the location of the interferer. Assuming that different transmitters in unlicensed spectrum use similar output powers, e.g., the maximum that is allowed by regulation, an interferer can be considered as a strong interferer if it is located closer to the receiver than the intended transmitter. Devices that jointly form a local network typically have some coordination or coexistence functionality provided by the wireless communication standard they are using, which avoids or reduces interference within the local network. However, other unlicensed radio technologies are typically not part of this interference coordination. In Fig. 14.2, it is depicted how different groups of devices use various unlicensed communication technologies. If these devices are separated in space, the interference is limited because it is typically significantly below the power levels of the communication within the group. However, if different unlicensed radio technologies operate at the same location, significant interference can occur.

Fig. 14.3 shows the challenge of long-range communication in unlicensed spectrum. With long-range communication it becomes more likely that an interferer is located closer to the receiver than the intended transmitter. The example of the figure shows a long-range system that is designed to cover a large path loss for transmission over several kilometers. There may exist several other local unlicensed networks using the same spectrum in vicinity of the long-range receiver. If the long-range receiver is, e.g., placed on the roof of a building, there may be some local unlicensed networks used in the same or neighboring buildings, e.g., for home automation. Because these devices are significantly closer to the long-range receiver, they may cause interference at the location of the long-range receiver, which is significantly higher, by e.g., several 10's of dB, than the strongly attenuated signal of the long-range transmitter, which is coming from far away. Furthermore, if we assume that the devices in the local network are adaptive devices, which e.g., use LBT to avoid interfering with other devices, this operation is likely to fail to adapt to long-range transmitters that are far away because the long-range signal is so strongly attenuated that it is below a sensitivity threshold used for clear channel assessment. As long as unlicensed spectrum is barely used, such interference situations may be unlikely. If it is anticipated that unlicensed IoT use cases (and other use cases) will drive the deployment of various local area networks using unlicensed spectrum the inference in unlicensed spectrum will increasingly play a role; long-range unlicensed radio technologies are more exposed to this interference.

Fig. 14.2 Coexistence among different groups of unlicensed devices. Inter-system interference is most severe if different groups are overlapping in space.

Fig. 14.3 Coexistence between long-range and short-range devices.

One aspect that is worth to mention for unlicensed spectrum, is about the maximum transmit power that may be emitted by the device. The maximum transmit power provided by spectrum regulation is typically given with respect to a certain reference configuration of emission. The reason is that different antenna configurations have different power emission patterns, which leads to that the antenna has different gains in different directions. Often maximum power levels are defined as either effective isotropically radiated power or as EIRP, as shown above for Tables 14.1–14.4. For EIRP an ideal isotropic antenna is assumed, whereas for ERP a half-wave dipole antenna is assumed, which has a 2.15   dB antenna gained compared with an isotropic antenna in direction of the highest antenna gain as specified in Eq. (14.1). When the maximum transmit power is specified as EIRP or ERP, a transmitter with any antenna configuration should not emit power in any direction that would exceed the maximum power that could be emitted in this direction with an isotropic or dipole antenna, respectively. A practical result of this is that if a real antenna has an antenna gain of X dB with a maximum allowed radiated power of Y dBm EIRP, then the transmitter must limit its conducted power at the antenna port to Y-X dBm. One consequence is that downlink performance can be substantially limited compared with uplink, if a base station antenna with significant antenna gain is used, see e.g. Ref. [51]. In uplink direction, the base station can use the antenna gain at the receiver side to improve the link budget. In the reverse downlink direction, the base station has to compensate the antenna gain by a reduced transmit power resulting in that the base station antenna gain does not improve the actual link budget, such as maximum path loss. This is one significant difference for a communication system operating in unlicensed spectrum compared with licensed spectrum, where the antenna gain could also be used in downlink direction and where higher maximum radiated powers are permitted in downlink.

Another cause for asymmetry of uplink and downlink in unlicensed spectrum can be found regarding capacity for nonadaptive devices that are limited by a duty-cycle. As an example, we assume a large number of N devices connected to a single base station, each device transmitting at a rate R _u and being limited by a maximum duty cycle of D _u . The maximum achievable uplink capacity is then limited by the maximum data rate and the duty cycle limitation for each device. Assuming an ideal situation of all transmissions being successful by neglecting all possible collisions and interference situations, the upper bound of maximum achievable uplink capacity C _u becomes in this case

C u = N · D u · R u ,

(14.2)

In downlink, the capacity of the base station is limited by its own maximum duty cycle. This duty cycle has to be used for the transmission to all N devices, in contrast to uplink where a duty cycle is valid per device. The upper bound of maximum achievable downlink capacity C _d under ideal error free transmissions becomes then

C d = D d · R d ,

(14.3)

for a downlink data rate R _d and a maximum duty cycle of D _d . Assuming the same uplink and downlink parameters, the downlink capacity is thus by a factor of N smaller compared with the uplink capacity, which leads to a significantly lower achievable effective downlink data rate per device. To some extent this can be compensated by configuring downlink transmissions to certain subbands that allow higher duty cycles (increasing D _d ), see e.g., Table 14.1.

14.2. Radio technologies for unlicensed spectrum

Unlicensed spectrum enables immediate market access for any new radio technology with minimal regulatory requirements. As a result, a very large number of different Machine-to-Machine (M2M) connectivity solutions have been developed and brought to the market making use of this spectrum. This section presents a set of established solutions for both long and short range communications. In addition, Chapter 15 provides an indepth description of LTE-M-U and and NB-IoT-U which are designed by the Multefire alliance to provide long range communication in unlicensed spectrum.

14.2.1. Short-range radio solutions

In this section we provide an overview of the most promising unlicensed short-range radio communication technologies for the IoT, which are IEEE 802.15.4, Bluetooth Low Energy (BLE) and Wi-Fi HaLow. The choice to focus on these is made based on the following properties of those technologies: they address the communication requirements of IoT, they target an IP-based IoT solution, and they are based on open standards and are expected to reach a substantial economy of scale. We also address the capillary network architecture where any of these solutions may be used to provide the short-range connectivity within a larger network context.

14.2.1.1. IEEE 802.15.4

IEEE 802.15.4 was standardized in 2003   at a time of intense research on wireless sensor networking technologies [13,14], and it is applicable for a wide range of IoT use cases, ranging from office automation, connected homes to industrial use cases. IEEE 802.15.4 was one of the early standards taking advantage of the adaptation framework IPv6 over low power wireless personal area networks (6LoWPAN) [10–12] which was specified in IETF to enable IP communication over very constrained wireless communication technologies (see also Chapter 17). Several application-specific protocol stacks have been developed, which build on parts of the IEEE 802.15.4 standard (mostly the physical layer and to some extent the medium access control (MAC)) [15], including ZigBee, WirelessHART, ISA-100, and Thread.

IEEE 802.15.4 has been specified for the three frequency bands of 868   MHz (for Europe), 915   MHz (for United States), and 2.4   GHz (global) [16–18]. In the 2.4   GHz band, IEEE 802.15.4 has 16 channels available, each of 2   MHz bandwidth. It uses offset quadrature phase shift keying with DSSS and a spreading factor of 8. A gross data rate of 250   kbps is achievable, see Refs. [16,18]. In 868   MHz one channel of 600   kHz is available, which uses differential binary phase shift keying modulation and DSSS with a spreading factor of 15. The achievable data rate at 868   MHz is 20   kbps [16,18]. In the 915   MHz band 10 channels are available with a gross data rate of 40   kbps, see Ref. [18].

IEEE 802.15.4 uses carrier-sense multiple access with collision avoidance (CSMA-CA) for access to the radio channel; this can be complemented with optional Automatic Repeat Request retransmissions. Typical coverage ranges are in the order of 10–20   m [11]. Two different topologies are supported: star topology and mesh (or peer-to-peer) topology. Two types of devices are defined: full-function devices that provide all MAC functionality and can act as network coordinator of the local network, and reduced function devices that can only communicate with a full-function device and are intended for very simple types of devices. The network can operate in beaconed mode, which allows a set of devices to synchronize to a superframe structure that is defined by the beacon transmitted by a local coordinating device. The channel access in this case is slotted CSMA-CA. In nonbeacon mode, unslotted CSMA-CA is applied. In case of direct data transmission, a device transmits data directly to another device. In indirect data transmission, data is transferred to a device, e.g., from a network coordinator. When beacon transmission is active, the network coordinator can indicate the availability of data in the beacon; the device can then request the pending data from the network coordinator. In nonbeaconed transmission, the network coordinator buffers data and it is up to the device to contact the network coordinator for pending data.

In Ref. [18] the performance of IEEE 802.15.4 has been evaluated in an experiment with an ideal link and devices placed at 1   m distance. It has been found that for a configuration at 2.4   GHz with a theoretical gross data rate of 250   kbps, a net data rate of 153   kbps was measured for direct transmission (from the device), and a net data rate of 66   kbps for indirect transmission (toward the device). Furthermore, it has been shown that the effective data rate and delivery ratio decrease with an increasing number of devices.

A major step for broader relevance of IEEE 802.15.4 for the IoT has been to address end-to-end IP-based communication. To this end the IETF working group 6LoWPAN has been chartered in 2005 and it has developed IETF standards for header compression and data fragmentation. The maximum physical layer payload size of 802.15.4 is limited to 127 bytes, which is further reduced by various protocol headers and optional security overhead and can leave as little as 81 bytes available for application data within an IEEE 802.15.4 frame. IETF has developed standards that provide header compression and IP packet fragmentation that enable the transmission of IPv6 over 802.15.4 networks [17,19–22]. In addition, the RPL routing protocol has been developed to enable IP mesh routing over IEEE 802.15.4 [17,22,23]. In 2014 the Thread group was formed with the objective to harmonize the usage of IEEE 802.15.4 together with 6LoWPAN for home automation.

IEEE 802.15.4 has been extended in IEEE 802.15.4g to address smart utility networks with an objective to improve coverage and support higher data rates [24–26]. To this end multiple new physical layers have been defined which can be used from a common MAC layer. Physical layer implementations are multirate frequency shift keying, multirate orthogonal frequency-division multiplexing, and multirate offset quadrature phase shift keying. Multirate frequency shift keying has a benefit of good transmit power efficiency because of constant signal envelop, MR-OFDM enables higher data rates for frequency selective fading channels, and multirate offset quadrature phase shift keying has the benefit of the original IEEE 802.15.4 modulation with cost-effective and easy design. A physical layer agnostic management protocol is based on a common signaling mode to allow a network configuration with interference coordination among multiple IEEE 802.15.4 transmitters [26].

14.2.1.2. BLE

Bluetooth has been developed as a technology for wireless short-range connectivity [27,28] and has established itself as a leading technology for personal area networking. With the release of the Bluetooth core specification 4.0 [29] in 2010 a novel transmission mode called Bluetooth Low Energy (BLE) was introduced, which considerably reduces power consumption compared with Bluetooth classic. BLE has been a significant first step to expand the Bluetooth ecosystem toward IoT.

BLE uses the 2.4   GHz ISM band. The spectrum is divided into 40 channels, with 2   MHz channel spacing, of which 37 are data channels and 3 are used as advertising channels. Frequency hopping is applied to mitigate the impact of interference. The modulation is based on Gaussian Frequency Shift Keying and a data rate of up to 1   Mbps can be achieved over-the-air. A master-slave architecture has been adopted to assign asymmetric roles to devices; peripheral devices perform only a minimum amount of functions to enable ultra-low power consumption, while central devices perform coordination functions. BLE has short connection setup and data transfer times so that applications can transfer authenticated data within a few milliseconds. BLE allows connection-oriented or connectionless communication. It supports fragmentation and reassembly of large data packets into small radio frames, which are then transmitted over the radio interface. This enables BLE to support data services with large packets (e.g., IP packets).

An analysis of BLE for building automation use cases has been performed in Refs. [16,30–32]. With a single-hop deployment, the range for BLE in an indoor deployment setup is in the order of 10   m, and around five BLE gateways are needed to provide coverage in a 1000   m² office floor [32].

In 2014 the Bluetooth Special Interest Group (BT SIG), the standardization forum for Bluetooth, published the Internet Protocol Support Profile [33], which enables IP connectivity for BLE devices. Further, IETF has developed a standard for end-to-end IPv6 connectivity over BLE [34], including header compression. This enables that end-to-end IP-based IoT services can be provided via BLE systems [35].

A further evolution of BLE has occurred with the launch of Bluetooth 5, the Bluetooth core specification 5.0 [36,37]. It comprises quadrupling of the communication range at low data rates (i.e., 125   kbps) and the doubling of the peak data rates (to 2   Mbps). Another important extension of BLE has been the introduction of mesh networking in 2017, which can significantly increase the range of BLE [38].

14.2.1.3. Wi-Fi

Wi-Fi based on IEEE 802.11 is one of the most used unlicensed radio access technologies and its focus is on providing high data rate services to a range of mainly consumer electronics devices. Very long battery life has not been in focus. Also the scalability of IEEE 802.11 has mainly addressed being able to provide a high total throughput to a number of connected devices in an area; from the early Wi-Fi version IEEE 802.11b to the version IEEE 802.11ac the theoretically achievable physical layer peak data rates have increased from 11   Mbps to 6.9   Gbps [39]. The typical operation of IEEE 802.11 is in the 2.4 and 5   GHz unlicensed spectrum bands.

IEEE 802.11ah is an amendment to the IEEE 802.11 standard that is focused on IoT applications. The Wi-Fi Alliance has chosen Wi-Fi HaLow as the marketing term to be used for the IEEE 802.11ah amendment. IEEE 802.11ah has some design targets that significantly differ from the high-data rate focused IEEE 802.11 variants. First, IEEE 802.11ah addresses the unlicensed spectrum below 1   GHz, which is in the range 902–928   MHz in the United States and 863–868   MHz in Europe; other regions also have unlicensed spectrum regions somewhere in the range 750–928   MHz [40]. Differences of the sub-1-GHz spectrum versus higher spectrum bands are as follows:

1. • The propagation conditions sub-1-GHz facilitate longer range. For a wide-area usage and spread of IoT devices, transmission range is a key property to provide sufficient coverage to IoT services with limited amount of access points. At the same time, the IoT devices are expected to transmit only limited amounts of data. For mobile broadband Wi-Fi usage, where devices are expected to transmit a lot of data, the extended range would mean that the channel is blocked for a longer time and the channel access time per device would reduce.
2. • There is less unlicensed spectrum available than at higher spectrum bands. This also means that the total capacity of data that can be provided within an area is lower than at higher spectrum bands. For mobile broadband focused Wi-Fi usage, this is a disadvantage because one focus is to provide high capacity in combination with high per user data rates. For IoT-focused Wi-Fi this is less of a problem, as the total amount of data transmitted even by a very large group of IoT devices is expected to remain modest.

The IEEE 802.11ah physical layer design is derived from the IEEE 802.11ac [1]. To address the lower spectrum bands, with less available bandwidth, and to enable robust long range transmission, the bandwidth of the IEEE 802.11ah has been scaled down by a factor of 10 compared to 802.11ac. That means that IEEE 802.11ah supports different carrier bandwidths of 2–16   MHz in comparison with the 20–160   MHz carriers of 802.11ac. In addition, an extra robust carrier configuration with 1   MHz bandwidth has been defined. Reference [1] describes a 24.5   dB link budget gain of IEEE 802.11   ah   at 900   MHz compared with 802.11n at 2.4   GHz. The gains stem from reduced path loss at low frequency (8.5   dB), reduced noise bandwidth due to narrower carriers (10   dB), further reduced noise bandwidth and repetition coding gains of the new robust 1   MHz carrier configuration (6   dB). The achievable data rates with IEEE 802.11ah are between 150   kbps and 347   Mbps. Several MAC features have been introduced to reduce power consumption for a client device and support more devices being connected to the same access point. IEEE 802.11 applies LBT in form of CSMA/CA. A larger number of connected devices lead to increased collision probabilities, which can be accentuated with the increased effect of hidden nodes with outdoor deployments [1]. To reduce the collision probability, the restricted access window (RAW) has been introduced. It divides the contention period into up to 64 RAW slots. Devices are allocated to particular RAW slots; and the number of devices, which are contending simultaneously for channel access, can be reduced to those devices being allocated to the same RAW slot. Device battery consumption can be significantly reduced, by enabling communication in uplink and downlink direction in new bidirectional transmission opportunities, where reverse link traffic can follow closely on forward link traffic. This enables long sleep cycles for devices. In addition, with a new target wake time the device and an access point can agree on certain fixed time periods, when data that the access point receives for a device shall be forwarded to the device. This reduces the amount of activity of a device to be able to receive data. Furthermore, the maximum idle period for a device has been extended in IEEE 802.11ah so that devices can be configured with sleep periods of up to around five years, and such devices only need to connect once every maximum idle period to the access point to avoid being automatically disassociated from the access point. IEEE 802.11ah also introduces new frame formats, which reduce the overhead of control information added in messages. This is significant for IoT traffic because the data payloads are often very small (e.g., a few bytes for a meter reading) and control info can quickly introduce significant overhead. For data transmission a short MAC frame format is added, and for control messages a null data packet has been introduced.

A more extensive description and evaluation of Wi-Fi IEEE 802.11ah can be found in Refs. [39–43]. The specification was published in early 2017 [52] and as of today no commercial chipsets are on the market.

14.2.1.4. Capillary networks

Short-range radio technologies provide the ability to build out connectivity efficiently to devices within a specific local area. Typically, these local, or capillary, networks need to be connected to the edge of a wide area communication infrastructure so that they have the ability, for example, to reach service functions that are hosted somewhere on the Internet or in a service cloud.

A capillary network needs a backhaul connection, which can be well provided by a cellular network. Their ubiquitous coverage allows backhaul connectivity to be provided practically anywhere, simply and, more significantly, without additional installation of network equipment. Furthermore, a capillary network might be on the move, as is the case for monitoring goods in transit, and therefore cellular networks are a natural solution. To connect a capillary network through a cellular network, a gateway is used between the cellular network and the capillary network, which acts just like any other cellular device toward the cellular network.

Fig. 14.4 illustrates an architecture, which comprises three domains: the capillary connectivity domain, the wide-area connectivity domain, and the data domain. The capillary connectivity domain spans the nodes that provide connectivity in the capillary network, and the wide-area connectivity domain spans the nodes of the cellular network. The data domain spans the nodes that provide data processing functionality for a desired service. These nodes are primarily the connected devices themselves as they generate and use service data through an intermediate node, such as a capillary gateway. The capillary gateway would also be included in the data domain if it provides data processing functionality (for example, if it acts as a CoAP mirror server).

Fig. 14.4 Capillary networks.

All three domains are separate from a security perspective, and end-to-end security can be provided by linking security relationships in the different domains to one another.

The ownership roles and business scenarios for each domain may differ from case to case. For example, to monitor the in-building sensors of a real estate company, a cellular operator might operate a wide-area network and own and manage the capillary network that provides connectivity to the sensors. The same operator may also own and manage the services provided by the data domain and, if so, would be in control of all three domains.

Alternatively, the real estate company might own the capillary network, and partner with an operator for connectivity and provision of the data domain. Or the real estate company might own and manage both the capillary network and the data domain with the operator providing connectivity only. In all these scenarios, different service agreements are needed to cover the interfaces between the domains specifying what functionality will be provided.

In large-scale deployments, some devices will connect through a capillary gateway, while others will connect directly. Regardless of how connectivity is provided, the bootstrapping and management mechanisms used should be homogenic to reduce implementation complexity and improve usability.

A more extensive discussion of IoT connectivity via capillary networks can be found in Ref. [44].

14.2.2. Long-range radio solutions

For unlicensed spectrum usage, short-range radio systems are most common. However, for IoT applications that require very low data rates, it is possible to trade lower data rate for a longer transmission range. Many technology concepts have been developed in recent years for unlicensed LPWAN. Many different variants of unlicensed LPWAN exist; some of which are more often referred to are as follows:

1. • LoRa
2. • Sigfox Ultra-Narrow Band (UNB)
3. • LTE-M-U
4. • NB-IoT-U

All of those have in common that they target wireless M2M/IoT communication over a long range of multiple kilometers, where devices transmit only infrequently very low amounts of data. Message sizes are small and there is often a focus on uplink transmission. Devices are desired to be simple and battery-powered operation should be possible over extended time periods. All these technologies are proprietary and not standardized in standards developing organizations.

In the following we provide a briefly overview of the LoRa and Sigfox technologies. LTE-M-U and NB-IoT-U are in detailed described in Chapter 15.

14.2.2.1. LoRa

LoRa is a network technology designed to provide long-range connectivity to battery operated devices; it is specified within an industry alliance. The LoRa Alliance claims to provide a Maximum Coupling Loss of 155   dB in the European 867–869   MHz band, and 154   dB in the United States 902–928   MHz band [45]. LoRa has the target to provide secure bidirectional communication. LoRa operates in the sub-GHz unlicensed frequency bands. The physical layer is based on Chirp spread-spectrum modulation, which is a technique using frequency modulation to spread the signal. A radio bearer is modulated with up and down chirps, where an up chirp corresponds to a pulse of finite length with increasing frequency, while a down chirp is a pulse of decreasing frequency. The channel bandwidth is mainly 125   kHz for European spectrum bands, and 125 or 500   kHz for US spectrum bands. Different data rates are supported and are reported to lie in the range of 300 bps–50   kbps. The selection of data rate is a trade-off between transmission duration, i.e., the time during which the message is transmitted over the air and range. The LoRa Alliance claims to provide a Maximum Coupling Loss of 155   dB in the European 867–869   MHz band, and 154   dB in the US 902–928   MHz band [46].

LoRa does not deploy LBT but instead uses the duty cycle restrictions required by regulation and the maximum dwell time (i.e., the maximum time that a device may continuously occupy a channel). The system architecture comprises LoRa end devices, LoRa gateways, and a network server. LoRa gateways correspond to base stations in a cellular network. Communication is between the end device and the network server. The communication between the network server and the gateway is based on IP communication; the communication between the end device and a gateway is based on LoRa specific protocols without IP. IP communication is terminated at the LoRa gateway. When a device transmits in uplink, the message can be received by one or more gateways.

For bidirectional communication a downlink transmission opportunity is provided after an uplink transmission. If downlink data arrives for a device in between uplink transmissions, the data needs to be buffered in the network and can only be transmitted during the devices downlink receive window, which follows on an uplink transmission by the device.

For more information on LoRa see Ref. [46].

14.2.2.2. Sigfox

Sigfox is a proprietary technology and of which until recentently no specifications have been publicly available. As of March 2019 Sigfox published a radio protocol specification describing the device operation. Some indicative properties of the so-called UNB communication scheme of Sigfox was however provided already in the Third Generation Partnership Project (3GPP) contribution [47] and technical reports of an industry specification group in ETSI [48,49].

UNB is targeted for operation in sub-GHz unlicensed spectrum bands. The supported channel bandwidths are 600   Hz and 100   Hz [49]. The uplink physical layer uses differential BPSK, which implies that the data rates are limited in the order of some few hundreds of bits per second. The maximum payload size for uplink data is 12 bytes. UNB does not use LBT but applies duty cycle limitations per transmitter. The channel access scheme is based on ALOHA, which starts to deteriorate at higher loads when the channel utilization exceeds around 15% [47]. Sigfox claims to support a maximum path loss, taking receive and transmit antenna gains into accout, of up to 163   dB [49].

The Sigfox network architecture comprises devices, which communicate with Sigfox servers. The radio communication is between the devices and Sigfox access points or base stations. Devices can transmit at any time without prior synchronization to the network. Typically, messages are transmitted on three different uplink channels, which are randomly selected. The base stations observe the entire system bandwidth to detect and decode uplink data. Messages can be received by different base stations, which provide selection diversity.

Downlink transmission is “piggybacked” onto uplink transmissions. After an uplink transmission a device maintains an open receiver window to receive downlink data for a certain time. The server sends buffered data to the device after receiving an uplink message. If the server has received uplink data via multiple base stations, it selects one of the base stations for downlink transmission.

References

[1]. Park M. IEEE 802.11ah: sub-1-GHz license-exempt operation for the internet of things.  Comput. Commun., . March 2015;Vol. 58:53–69.

[2]. Federal Communications Commission, . Title 47 of the Code of federal regulations, Part 15 on radio frequency devices. 2017 Available from:. https://www.fcc.gov/general/rules-regulations-title-47.

[3]. European Conference of Postal and Telecommunications Administrations (CEPT). Available from: http://www.cept.org/cept/.

[4]. CEPT Electronic Communications Committee (ECC).  ERC recommendation 70-03, relating to the use of short-range devices (SRD) . 2016 Edition of October 2016.

[5]. European Telecom Standards Institute, .  EN 300 220-1 V3.1.1, short range devices (SRD) operating in the frequency range 25 MHz to 1 000 MHz; Part 1: technical characteristics and methods of measurement . February 2017.

[6]. European Telecom Standards Institute, .  EN 300 220-2 V2.4.1 electromagnetic compatibility and radio spectrum matters (ERM); short range devices (SRD); radio equipment to Be used in the 25 MHz to 1 000 MHz frequency range with power levels ranging up to 500 mW; Part 2: harmonized EN covering essential requirements under article 3.2 of the R&TTE directive . May 2012.

[7]. European Telecom Standards Institute, .  EN 300 440-1 V1.6.1, electromagnetic compatibility and radio spectrum matters (ERM); short range devices; radio equipment to Be used in the 1 GHz to 40 GHz frequency range; Part 1: technical characteristics and test methods . August 2010.

[8]. European Telecom Standards Institute, .  EN 300 440 V2.1.1, short range devices (SRD); radio equipment to Be used in the 1 GHz to 40 GHz frequency range; harmonised standard covering the essential requirements of article 3.2 of directive 2014/53/EU . January 2017.

[9]. European Telecom Standards Institute.  EN 300 328 V2.1.1, wideband transmission systems; data transmission equipment operating in the 2,4 GHz ISM band and using wide band modulation techniques; harmonised standard covering the essential requirements of article 3.2 of directive 2014/53/EU . November 2011.

[10]. Gutierrez J.A, Naeve M, Callaway E, Bourgeois M, Milter V, Heile B. IEEE 802.15.4: a developing standard for low-power low-cost wireless personal area networks.  IEEE Network., . September/October 2001;Vol. 15(5):12–19.

[11]. Callaway E, Gorday P, Hester L, Gutierrez J.A, Naeve M, Heile B, Bahl V. Home networking with IEEE 802.15.4: a developing standard for low-rate wireless personal area networks.  IEEE Commun. Mag., . August 2002;Vol. 40(8):70–77.

[12]. Zheng J, Lee M.J. Will IEEE 802.15.4 make ubiquitous networking a reality? a discussion on a potential low power, low bit rate standard.  IEEE Commun. Mag., . June 2004;Vol. 42(6):140–146.

[13]. Akyildiz I.F, Su W, Sankarasubramaniam Y, Cayirci E. Wireless sensor networks: a survey.  Comput. Netw., . 2002;Vol. 38:393–422.

[14]. Akyildiz I.F, Su W, Sankarasubramaniam Y, Cayirci E. A survey on sensor networks.  IEEE Commun. Mag., . August 2002;Vol. 40(8):102–114.

[15]. Willig A. Recent and emerging topics in wireless industrial communications: a selection.  IEEE Transactions on Industrial Informatics, . May 2008;Vol. 4(2):102–124.

[16]. Langhammer N, Kays R. Performance evaluation of wireless home automation networks in indoor scenarios.  IEEE Transactions on Smart Grid, . December 2012;Vol. 3(4).

[17]. Palattella M.R, Accettura N, Vilajosana X, Watteyne T, Grieco L.A, Boggia G, Dohler M. Standardized protocol stack for the internet of (important) things.  IEEE Communications Surveys and Tutorials, . 2013;Vol. 15(3).

[18]. Lee J.-S. Performance evaluation of IEEE 802.15.4 for low-rate wireless personal area networks.  IEEE Trans. Consum. Electron., . August 2006;Vol. 52(3).

[19]. Internet Engineering Task Force (IETF), .  Request for comments 4944, transmission of IPv6 packets over IEEE 802.15.4 networks . September 2007.

[20]. Hui J.W, Culler D.E. IPv6 in low-power wireless networks.  Proc. IEEE, . November 2010;Vol. 98(11).

[21]. Hui J.W, Culler D.E.  Extending IP to low-power, wireless personal area networks . IEEE Internet Computing; July/August 2008. .

[22]. Ishaq I, Carels D, Teklemariam G.K, Hoebeke J, Van den Abeele F, De Poorter E, Moerman I, Demeester P. IETF standardization in the field of the internet of things (IoT): a survey.  J. Sens. Actuator Netw., . April 2013;Vol. 2(2):235–287.

[23]. Internet Engineering Task Force (IETF), .  Request for comments 6550, RPL: IPv6 routing protocol for low-power and lossy networks . March 2012.

[24]. Chang K.-H, Mason B. The IEEE 802.15.4g standard for smart metering utility networks. In:  Proceedings IEEE international conference on smart grid communications, Tainan city, Taiwan, . 5–8 November 2012.

[25]. Sum C.-S, Kojima F, Harada H. Coexistence of homogeneous and heterogeneous systems for IEEE 802.15.4g smart utility networks. In:  Proc. IEEE international symposium on dynamic spectrum access networks (DySPAN), Aachen, Germany, 3–6 May . 2011.

[26]. Sum C.-S, Harada H, Kojima F, Lu L. An interference management protocol for multiple physical layers in IEEE 802.15.4g smart utility networks.  IEEE Commun. Mag., . April 2013;Vol. 51(4):84–91.

[27]. Haartsen J. Bluetooth—the universal radio interface for ad hoc, wireless connectivity.  Ericsson Rev. Engl. Ed., . 1998;Vol. 75(3):110–117 Available from:. http://ericssonhistory.com/Global/Ericsson%20review/Ericsson%20Review.%201998.%20V.75/Ericsson_Review_Vol_75_1998_3.pdf.

[28]. Haartsen J.C. The bluetooth radio system.  IEEE Personal Communications, . August 2002;Vol. 7(1):28–36.

[29]. Bluetooth Special Interest Group, .  Bluetooth core specification v4.0 . June 2010.

[30]. Del Carpio L.F, Di Marco P, Skillermark P, Chirikov R, Lagergren K, Amin P. Comparison of 802.11ah and BLE for a home automation use case. In:  Proc. IEEE 27th Annual international symposium on personal, indoor, and mobile radio communications (PIMRC), . 4–8 September 2016.

[31]. L. F. Del Carpio, P. Di Marco, P. Skillermark, R. Chirikov, K. Lagergren. Comparison of 802.11ah and BLE in a home automation use case, to appear in springer international journal of wireless information networks.

[32]. Di Marco P, Chirikov R, Amin P, Militano F. Coverage analysis of bluetooth low energy and IEEE 802.11ah for office scenario. In:  Proc. IEEE int. Symposium on personal, indoor and mobile radio communications . Workshop M2M Communication; August 2015.

[33]. Bluetooth Special Interest Group, .  Internet protocol support profile (IPSP) . December 2014.

[34]. Internet Engineering Task Force (IETF), .  Request for comments 7668 . IPv6 over BLUETOOTH Low Energy; October 2015.

[35]. Bluetooth Special Interest Group, .  Internet gateways, Bluetooth white paper . February 2016.

[36]. Bluetooth Special Interest Group, .  Bluetooth core specification v5.0 . December 2016.

[37]. Di Marco P, Skillermark P, Larmo A, Arvidson P, Chirikov R. Performance evaluation of the data transfer modes in Bluetooth 5.  IEEE Communications Standards Magazine, . July 2017;Vol. 1(2):92–97.

[38]. Bluetooth Special Interest Group.  Bluetooth specification mesh profile, v1.0.1 . January 2019.

[39]. Sun W, Lee O, Shin Y, Kim S, Yang C, Kim H, Choi S.  Wi-fi could Be much more . IEEE Communications Magazine; November 2014.

[40]. Sun W, Choi M, Choi S. IEEE 802.11ah: a long range 802.11 WLAN at sub 1 GHz.  Journal of ICT Standardization, . July 2013;Vol. 1(1):83–108.

[41]. Adame T, Bel A, Bellalta B, Barcelo J, Oliver M. IEEE 802.11ah: the WiFi approach for M2M communications.  IEEE Wireless Communications, . December 2014;Vol. 21(6):144–152.

[42]. Khorov E, Lyakhov A, Krotov A, Guschin A. A survey on IEEE 802.11ah: an enabling networking technology for smart cities.  Comput. Commun., . March 1, 2015;Vol. 58:53–69.

[43]. Badihi B, Del Carpio L.F, Amin P, Larmo A, Lopez M, Denteneer D. Performance evaluation of IEEE 802.11ah actuators. In:  Proceedings 83rd IEEE vehicular technology conference, Nanjing, China . May 2016:15–18.

[44]. Sachs J, Beijar N, Elmdahl P, Melen J, Militano F, Salmela P.  Capillary networks – a smart way to get things connected . Ericsson Review; September 2014.

[45]. LoRA Alliance Technical Marketing Workgroup 1.0.  White paper, LoRaWAN, what is it?  A Technical Overview of LoRa® and LoRaWAN; November 2015.

[46]. LoRa Alliance.  LoRaWAN™ specification, V1.0 . January 2015.

[47]. Sigfox Wireless. GPC150052, coperative ultra narrow band technology for cellular IoT, 3GPP TSG GERAN1-GERAN2 ad'hoc meeting, sophia antipolis, France, 2–5; February 2015. .

[48]. ETSI Industry Specification Group on Low Throughput Networks.  (ISG LTN), GS LTN 002 V1.1.1, low throughput networks (LTN); functional architecture . September 2014.

[49]. ETSI Industry Specification Group on Low Throughput Networks.  (ISG LTN), GS LTN 003 V1.1.1, low throughput networks (LTN); protocols and interfaces . September 2014.

[50]. European Comission, Comission Decision of 9 November 2006 on harmonisation of the radio spectrum for use by short-range device (2006/771/EC), Edition of 18.8.2017.

[51]. S. Andreev, O. Galinina, A. Pyattaev, Gerasimenko, T. Tirronen M, J. Torsner, J. Sachs, M. Dohler, Koucheryavy Y. Understanding the IoT connectivity landscape: a contemporary M2M radio technology roadmap.  IEEE Commun. Mag., . September 2015;Vol. 53(9):32–40.

[52]. IEEE 802.11ah-2016 - IEEE standard for information technology--telecommunications and information exchange between systems - local and metropolitan area networks--specific requirements - Part 11: wireless LAN medium access control (MAC) and physical layer (PHY) specifications amendment 2: sub 1 GHz license exempt operation . May 2017.