NB-IoT

Abstract

This chapter presents the design of Narrowband Internet of Things (NB-IoT). The first part of this chapter describes the background behind the introduction of NB-IoT in the Third Generation Partnership Project (3GPP) specifications and the design principles of the technology.

The second part of this chapter focuses on the physical channels with an emphasis on how these channels were designed to fulfill the objectives that NB-IoT was intended to achieve, namely, deployment flexibility, ubiquitous coverage, ultra-low device cost, long battery lifetime, and capacity sufficient for supporting a massive number of devices in a cell. Detailed descriptions are provided regarding both downlink and uplink transmission schemes and how each of the NB-IoT physical channels is mapped to radio resources in both the frequency and time dimensions.

The third part of this chapter covers NB-IoT idle and connected mode procedures and the transition between these modes, including all activities from initial cell selection to completing a data transfer.

NB-IoT since its first introduction has been further enhanced both in terms of performance and features. This chapter highlights the improvements accomplished in Release 14 and Release 15 of the 3GPP specifications. As NB-IoT devices are expected to have long life cycles, how a network operator migrates from Long-Term Evolution (LTE) to Fifth Generation (5G) New Radio (NR) while continuing to honor its contracts toward NB-IoT service providers becomes an important aspect. At the end of this chapter, we describe NB-IoT and NR coexistence.

Keywords

Access barring; Blind decoding; Cell selection; Channel coding; Channel raster; Connected mode; Extended coverage; Frame structure; Idle mode; Master information block; Mobility; Modulation; Multicarrier; Multicast; NB-IoT; NPBCH; NPDCCH; NPDSCH; NPRACH; NPSS; NPUSCH; NSSS; Operation modes; Paging; Positioning; Power control; Power Saving Mode; Random access; Scheduling; Search space; Synchronization; System information block; Early data transmission; Wake-up signal; Power head room; New Radio; 5G; 5G migration; Coexistence

7.1. Background

7.1.1. 3GPP standardization

In early 2015, the market for low-power wide area networks (LPWAN) was rapidly developing. Sigfox was building out their Ultra Narrowband Modulation networks in France, Spain, the Netherlands, and the United Kingdom. The LoRa Alliance, founded with a clear ambition to provide IoT connectivity with wide-area coverage, released LoRaWAN R1.0 specification in June 2015 [1]. The alliance at that point quickly gathered significant industry interest and strong membership growth. Up until then, Global System for Mobile Communications/General Packet Radio Service (GSM/GPRS) had been the main cellular technology of choice for serving wide-area IoT use cases, thanks to it being a mature technology with low modem cost. This position was challenged by the emerging LPWAN technologies that presented an alternative technology choice to many of the IoT verticals served by GSM/GPRS.

Anticipating the new competition, 3GPP (see Section 2.1) started a feasibility study on Cellular System Support for Ultra-low Complexity and Low Throughput Internet of Things [2], referred to as the Cellular IoT study for short in the following sections. As explained in Section 2.2.5, ambitious objectives on coverage, capacity, and battery lifetime were set up, together with a more relaxed objective of a maximum system latency. All these performance objectives offer major improvements over GSM and GPRS, as specified at that time, toward better serving the IoT verticals. One additional objective was that it should be possible to introduce the IoT features to the existing GSM networks through software upgrade. Building out a national network takes many years and requires substantial investment up front. With software upgrade, however, the well-established cellular network can be upgraded overnight to meet all the key performance requirements of the IoT market.

Among the solutions proposed to the Cellular IoT study, some were backward compatible with GSM/GPRS and were developed based on the evolution of the existing GSM/GPRS specifications. EC-GSM-IoT described in Chapters 3 and 4 is the solution eventually standardized in 3GPP Release 13.

Historically, the group carrying out the study, 3GPP TSG GERAN (Technical Specifications Group GSM/EDGE Radio Access Network), had focused on the evolution of GSM/GPRS technologies, developing features for meeting the need of GSM operators. Certain GSM operators, however, at that point considered refarming their GSM spectrum for deploying the Long-Term Evolution (LTE) technology as well as to LPWAN dedicated for IoT services. This consideration triggered the study on non–GSM backward compatible technologies, referred to as clean-slate solutions. Although none of the clean-slate solutions in the study were specified, it provided a firm ground for the Narrowband Internet of Things (NB-IoT) technology that emerged after study completion and was standardized in 3GPP Release 13. As described later in this chapter, the entire NB-IoT system is supported in a bandwidth of 180 kHz, for each of the downlink and uplink. This allows for deployment in refarmed GSM spectrum as well as within an LTE carrier. NB-IoT is part of the 3GPP LTE specifications and employs many technical components already defined for LTE. This approach reduced the standardization process and leveraged the LTE ecosystem to ensure a fast time to market. It also possibly allows NB-IoT to be introduced through a software upgrade of the existing LTE networks. The normative work of developing the core specifications of NB-IoT took only a few months and was completed in June 2016 [3].

Since its first release in 2016, NB-IoT up to 2018 has gone through two additional releases, i.e. 3GPP Releases 14 and 15. These later releases continued to improve device energy efficiency. Furthermore, features for improving system performance and for supporting new use cases and additional deployment options were also introduced. Table 7.1 provides a summary. These enhancements further improve NB-IoT's position as a superior LPWAN technology. Like before, all the Releases 14 and 15 features can be enabled through a software upgrade of the existing LTE or NB-IoT networks. In Release 15, 3GPP evaluated NB-IoT against a set of agreed Fifth Generation (5G) performance requirements defined for the massive machine-type communications (mMTC) use case [4]. As shown in Section 8.9, NB-IoT meets these requirements with margins and in all relevant aspects qualifies as a 5G mMTC technology.

The GSM Association, which is an organization that represents the interests of mobile network operators worldwide, tracks the status of commercial launches of NB-IoT. Since the completion of its first release in June 2016, there had been 80 NB-IoT launches in 45 markets as of June 2019, according to GSM Association [5]. On the device side, the Global Mobile Suppliers Association published a research report in 2018 [6] stating that as of August 2018, there were 106 modules supporting NB-IoT, 43 of which also support LTE-M. According to Ref. [7], the global NB-IoT chipset market is projected to have compound annual growth rate higher than 40% in the next decade. The device ecosystem has been well established and is expected to have strong momentum and growth in the coming years.

Table 7.1

New NB-IoT features introduced in 3GPP Releases 14 and 15.
Motivation	Release 14 (2017)	Release 15 (2018)
Device energy efficiency improvement	Category NB2 device (Sections 7.2.3, 7.2.4.6, 7.2.5.2) Support of 2 HARQ processes (Section 7.3.2.2.3)	Early data transmission (Section 7.3.1.7.3) Wake-up signal (Sections 7.2.4.8 and 7.3.1.4) Quick RRC release procedure Relaxed cell reselection monitoring (Section 7.3.1.3) Improved scheduling request (Section 7.3.2.5) Periodic buffer status reporting (Section 7.3.2.5) RLC unacknowledged mode
System performance improvement	Multicast (Section 7.3.1.11) Non-anchor carrier paging (Section 7.3.1.10) Non-anchor carrier random access (Section 7.3.1.10) Mobility enhancement DL channel quality reporting during random access procedure (Section 7.3.1.8)	Improved access barring (Section 7.3.1.9) Improved system information acquisition (Section 7.3.1.2.1) Improved measurement accuracy (Section 7.3.1.3) Improved UE power headroom reporting (Section 7.3.2.3.2) UE differentiation
New use case	14 dBm device power class (Section 7.1.2.1) Positioning (Section 7.3.2.6)
Support for more deployment options		TDD Support (Section 7.2.8) Support of small cell deployment Extend cell radius up to 120 km (Section 7.2.5.1) Flexible uses of stand-alone carrier (Section 7.3.2.7)

7.1.2. Radio access design principles

NB-IoT is designed for ultra-low-cost mMTC, aiming to support a massive number of devices in a cell. Low device complexity is one of the main design objectives, enabling low module cost. Furthermore, it is designed to offer substantial coverage improvements over GPRS as well as for enabling long battery lifetime. Finally, NB-IoT has been designed to give maximal deployment flexibility. In this section, we will highlight the design principles adopted in NB-IoT to achieve these objectives.

7.1.2.1. Low device complexity and cost

Device modem complexity and cost are primarily related to the complexity of baseband processing, memory consumption, and radio-frequency (RF) requirements. Regarding baseband processing, the two most computationally demanding tasks are synchronization during initial cell selection and demodulation and decoding during data reception. NB-IoT is designed for allowing low-complexity receiver processing in accomplishing these two tasks. For initial cell selection, a device needs to search for only one synchronization sequence for establishing basic time and frequency synchronization to the network. The device can use a low sampling rate (e.g., 240 kHz) and take advantage of the synchronization sequence properties to minimize memory requirement and complexity.

During connected mode, low device complexity is facilitated by restricting the downlink transport block (TB) size (TBS) to be no larger than 680 bits for the lowest device category and relaxing the processing time requirements compared with LTE. For channel coding, instead of using the LTE turbo code [8], which requires iterative receiver processing, NB-IoT adopts a simple convolutional code, i.e., the LTE tail-biting convolutional code (TBCC) [8], in the downlink channels. In addition, NB-IoT does not use higher-order modulations or multilayer multiple-input multiple-output transmissions. Furthermore, a device needs to support only half-duplex operation and is not required to receive in the downlink while transmitting in the uplink, or vice versa.

Regarding RF, all the performance objectives of NB-IoT can be achieved with one transmit-and-receive antenna in the device. That is, neither downlink receiver diversity nor uplink transmit diversity is required in the device. NB-IoT is designed for allowing relaxed oscillator accuracy in the device. For example, a device can achieve initial acquisition when its oscillator inaccuracy is as large as 20 parts per million (ppm). During a data session, the transmission scheme is designed for the device to easily track its frequency drift. Because a device is not required to simultaneously transmit and receive, a duplexer is not needed in the RF front end of the device. The maximum transmit power level of an NB-IoT device is either 20 or 23 dBm in 3GPP Release 13. Release 14 introduced a lower device power class with a maximum transmit power of 14 dBm. A device power level of 14 or 20 dBm allows on-chip integration of the power amplifier (PA), which can contribute to the device cost reduction. The 14 dBm device power class allows batteries with lower peak current draw, and thus facilitates the adoption of batteries of smaller form factors.

Economy of scale is yet another contributor to cost reduction. Thanks to its deployment flexibility and low minimum system bandwidth requirement, NB-IoT is already being globally available in many networks. This will help to increase the economy of scale of NB-IoT.

7.1.2.2. Coverage enhancement

Coverage enhancement (CE) is mainly achieved by trading off data rate for coverage. Like EC-GSM-IoT and LTE-M, repetitions are used to ensure that devices in coverage challenging locations can still have reliable communications with the network, although at a reduced data rate. Furthermore, NB-IoT has been designed to use a close to constant envelope waveform in the uplink. This is an important factor for devices in extreme coverage-limited and power-limited situations because it minimizes the need to back off the output power from the maximum configurable level. Minimizing the power backoff helps preserve the best coverage possible for a given power capability.

7.1.2.3. Long device battery lifetime

Minimizing power backoff also gives rise to higher PA power efficiency, which helps extend device battery lifetime. Device battery lifetime, however, depends heavily on how the device behaves when it does not have an active data session. In most use cases, the device actually spends the vast majority of its lifetime in idle mode as most of the IoT applications only require infrequent transmission of short packets. Traditionally, an idle device needs to monitor paging and perform mobility measurements. Although the energy consumption during idle mode is much lower compared with during connected mode, significant energy saving can be achieved by simply increasing the periodicity between paging occasions (POs) or not requiring the device to monitor paging at all. As elaborated on in Section 2.2, 3GPP Releases 12 and 13 introduced both extended discontinuous reception (eDRX) and power saving mode (PSM) to support this type of operation and optimize device power consumption. In essence, a device can shut down its transceiver and only keep a basic oscillator running for the sake of keeping a rough time reference to know when it should come out of the PSM or eDRX. The reachability during PSM is set by the tracking area update (TAU) timer with the maximum settable value exceeding 1 year [9]. eDRX can be configured with a cycle just below 3 h [10].

During these power-saving states, both device and network maintain device context, saving the need for unnecessary signaling when the device comes back to connected mode. This optimizes the signaling and power consumption when making the transition from idle to connected mode.

In addition to PSM and eDRX, NB-IoT also adopts connected mode DRX as a major tool for achieving energy efficiency. In Release 13, the connected mode DRX cycles were extended from 2.56 s to 10.24 s for NB-IoT [11].

7.1.2.4. Support of massive number of devices

NB-IoT achieves high capacity in terms of number of devices that can be supported on one single NB-IoT carrier. This is made possible by introducing spectrally efficient transmission scheme in the uplink for devices in extreme coverage-limited situation.

Shannon's well-known channel capacity theorem [12] establishes a relationship between bandwidth, power, and capacity in an additive white Gaussian noise channel as

C = W log 2 ( 1 + S N ) = W log 2 ( 1 + S N 0 W ) ,

(7.1)

where C is the channel capacity (bits/s), S is the received desired signal power, N is the noise power, which is determined by the product of noise bandwidth (W) and one-sided noise power spectral density (N ₀). The noise bandwidth is identical to the signal bandwidth if Nyquist pulse shaping function is used. At extreme coverage-limited situation, S N ≪ 1

. Using the approximation ln(1 + x) ≈ x, for x ≪ 1, it can be shown that the channel capacity in the very low signal-to-noise-power ratio (SNR) regime is

C = S N 0 log 2 ( e ) .

(7.2)

In this regime, the bandwidth dependency vanishes, and therefore channel capacity, in terms of bits per second, is only determined by the ratio of S and N ₀. Thus, in theory the coverage for a target data rate R = C depends only on the received signal power level, and not on the signal bandwidth. This implies that because the data rate at extreme coverage-limited situation does not scale according to the device bandwidth allocation, for the sake of spectral efficiency, it is advantageous to allocate small bandwidth for devices in bad coverage. NB-IoT uplink waveforms include various bandwidth options. While a waveform of wide bandwidth (e.g., 180 kHz) is beneficial for devices in good coverage, waveforms of smaller bandwidths are more spectrally efficient from the system point of view for serving devices in bad coverage. This will be illustrated by the coverage results presented in Chapter 8.

7.1.2.5. Deployment flexibility

To support maximum flexibility of deployment and prepare for refarming scenarios, NB-IoT supports three modes of operation, stand-alone, in-band, and guard-band.

7.1.2.5.1. Stand-alone mode of operation

NB-IoT may be deployed as a stand-alone carrier using any available spectrum with bandwidth larger than 180 kHz. This is referred to as the stand-alone deployment. A use case of the stand-alone deployment is for a GSM operator to deploy NB-IoT in its GSM band by refarming part of its GSM spectrum. In this case, however, additional guard-band is needed between a GSM carrier and the NB-IoT carrier. Based on the coexistence requirements in Ref. [13], 200 kHz guard-band is recommended, which means that a GSM carrier should be left empty on one side of an NB-IoT carrier between two operators. In case of the same operator deploying both GSM and NB-IoT, a guard-band of 100 kHz is recommended based on the studies in Ref. [2], and hence an operator needs to refarm at least two consecutive GSM carriers for NB-IoT deployment. An example is illustrated in Fig. 7.1. Here, NB-IoT bandwidth is shown as 200 kHz. This is because NB-IoT needs to meet the GSM spectral mask when deployed using refarmed GSM spectrum and the GSM spectral mask is specified according to 200 kHz channelization.

7.1.2.5.2. In-band and guard-band modes of operation

NB-IoT is also designed to be possible for deployment in the existing LTE networks, either using one of the LTE physical resource blocks (PRBs) or using the LTE guard-band. These two deployment scenarios are referred to as in-band and guard-band deployments, respectively. An example is illustrated in Fig. 7.2. An LTE carrier with a number of PRBs is shown. NB-IoT can be deployed using one LTE PRB or using the unused bandwidth in the guard-band. The guard-band deployment makes use of the fact that the occupied bandwidth of the LTE signal is roughly 90% of channel bandwidth [14], leaving roughly 5% of the LTE channel bandwidth on each side available as guard-band. It is therefore possible to place an NB-IoT carrier in the guard band of LTE when the LTE carrier bandwidth is 5, 10, 15, or 20 MHz.

Yet another possible deployment scenario is to have NB-IoT in-band deployment on an LTE carrier that supports LTE-M features. The concept of narrowband used in LTE-M is explained in Section 5.2.2.2. Some of these LTE-M narrowbands are not used for transmitting LTE-M system information (SI) Block Type 1 (SIB1) and thus can be used for deploying NB-IoT. More details about this deployment scenario are given in Section 7.2.1.1.

7.1.2.5.3. Spectrum refarming

NB-IoT is intended to offer flexible spectrum migration possibilities to a GSM operator. An operator can take an initial step to refarm a small part of the GSM spectrum to NB-IoT as the example shown in the top part of Fig. 7.3. Thanks to the support of LTE in-band and guard-band deployments, such an initial migration step will not result in spectrum fragmentation to make the eventual migration of the entire GSM spectrum to LTE more difficult. As illustrated in Fig. 7.3, the NB-IoT carrier already deployed in the GSM network as a stand-alone deployment may become an LTE in-band or guard-band deployment when the entire GSM spectrum is migrated to LTE. This high flexibility was also envisioned to facilitate NB-IoT deployments when LTE is later refarmed to the 5G New Radio (NR) technology. Indeed, as it will be described in Section 7.4 NB-IoT can be deployed together with NR, and in such a deployment, NR and NB-IoT achieve superior coexistence performance.

7.1.2.6. Coexistence with LTE

When designing a new access technology, the degree of freedom is higher compared with when basing the design on an existing technology. Therefore NB-IoT could with limited restrictions be designed from ground-up with the intention to follow the radio access design principles outlined thus far.

The ambition to provide a radio access technology with high deployment flexibility and the capability to operate both in a refarmed GSM spectrum and inside an LTE carrier did, however, impose certain guiding principles onto the design of the technology.

While the stand-alone deployment in the GSM spectrum, discussed in Section 7.1.2.5, is facilitated by the introduction of a guard-band between the NB-IoT and GSM carriers, the expectations on close coexistence with LTE were higher. NB-IoT deployment inside an LTE carrier was hence required to be supported without any guard-band between NB-IoT and LTE PRBs. To minimize the impact on the existing LTE deployments and devices, this imposed requirements on the NB-IoT physical layer waveforms to preserve orthogonality with the LTE signal in adjacent PRBs. It also implies that NB-IoT should be able to share the same time-frequency resources as LTE the same way as different LTE physical channels share time-frequency resources. Last but not least, because legacy LTE devices will not be aware of the NB-IoT operation, NB-IoT transmissions should not collide with essential LTE transmissions.

Among the essential LTE transmissions are the physical channel signals transmitted in the downlink control region, including Physical Control Format Indicator Channel (PCFICH), Physical Hybrid ARQ Indicator Channel, and Physical Downlink Control Channel (PDCCH). PCFICH is used for indicating the number of orthogonal frequency-division multiplexing (OFDM) symbols in a subframe that can be used for PDCCH. In LTE, starting from the beginning of a subframe, up to 3 OFDM symbols may be used for PDCCH transmissions. A PDCCH transmission may carry scheduling information, paging indicators, etc. As illustrated in Fig. 7.4, the resource elements (REs) of the first three OFDM symbols in an LTE subframe may therefore not be used by NB-IoT downlink channels.

Fig. 7.4 An illustration of how NB-IoT physical channels of the in-band mode share REs with LTE in the downlink.

Additional essential LTE physical channels and signals are the Cell-specific Reference Signal (CRS) and Multicast-Broadcast Single-Frequency Network (MBSFN) signal. The REs used by these channels and signals are also preserved and not mapped to any NB-IoT physical channels. MBSFN signal, for example, spans one subframe in the time dimension and all PRBs in the frequency dimension, and thus if one subframe is configured as an LTE MBSFN subframe, that subframe is not used by NB-IoT. Fig. 7.4 illustrates such an example.

Furthermore, NB-IoT also avoids using resources mapped to LTE Primary Synchronization Signal (PSS), Secondary Synchronization Signal (SSS), and Physical Broadcast Channel (PBCH). Since these transmissions use the middle six PRBs, in the case of 1.4, 10, and 20 MHz LTE carrier bandwidths, or middle seven PRBs, in the case of 3, 5, and 15 MHz LTE carrier bandwidths, an NB-IoT carrier cannot use any of these middle LTE PRBs in an in-band deployment.

Following these guiding principles will naturally mean that the physical layer to a large extent is inspired by LTE. At the same time, changes are required to meet the aforementioned design objectives.

Interestingly, the objective for achieving superior coexistence performance with LTE paves the way for NB-IoT to also achieve superior coexistence performance with 5G NR, after the LTE to NR migration. NB-IoT coexistence with NR will be described in Section 7.4.