Whitepaper: Narrowband Internet of Things

Narrowband Internet of Things
Whitepaper
As part of Release 13, 3GPP has specified a new radio interface, the Narrowband Internet of Things (NBIoT). NB-IoT is optimized for machine type traffic. It is kept as simple as possible in order to reduce device
costs and to minimize battery consumption. In addition, it is also adapted to work in difficult radio conditions, which is a frequent operational area for certain machine type communication devices. Although NBIoT is an independent radio interface, it is tightly connected with LTE, which also shows up in its integration in the current LTE specifications.
In this whitepaper we introduce the NB-IoT technology with an emphasis on the tight connection to LTE.
Note:
Visit our homepage for the most recent version of this application note (www.rohde-schwarz.com/appnote/
1MA266).
Whitepaper
J. Schlienz, D. Raddino
NarrowBand_IoT 1MA266_0e
Contents
Contents
1 Introduction............................................................................................ 3
2 Overview................................................................................................. 5
3 Physical Layer........................................................................................ 9
4 Cell Access........................................................................................... 24
5 Data Transfer........................................................................................ 33
6 Summary and Outlook.........................................................................39
7 References............................................................................................40
8 Additional Information.........................................................................41
9 Rohde & Schwarz.................................................................................42
Whitepaper NarrowBand_IoT 1MA266_0e
Introduction
1 Introduction
One of the characteristics of Machine Type Communication (MTC) is the broad spectrum of capabilities. For example surveillance cameras have to deliver a huge amount
of UL data while being almost stationary, whereas devices for fleet tracking have a
small amount of data while performing a lot of handovers.
Yet another class of devices has neither of these capabilities. Examples are devices
for meter reading like electricity, gas, or water consumption. They are often stationary,
thus need not an optimized handover. Only a small amount of data is usually transferred, which is even not delay sensitive. However, the number of these MTC devices
may become quite big, even up to several orders of magnitude compared to the traditional devices. Using existing LTE technology would lead to a network overload,
because despite of their small amount of user data the amount of signaling is about the
same. The first specification of NB-IoT focusses on this class of devices.
These devices are often installed at places without power supply. Consequently they
run completely on battery and it may be very expensive to change the battery, because
they may only be accessed by trained staff. Hence, in some cases the battery lifetime
can even determine the lifetime of the whole device. An optimized power consumption
is therefore essential for a proper operation. In addition, the coverage at these places
is often quite bad. Therefore, the indoor coverage has to be significantly improved, up
to 23 dB are regarded as necessary.
Due to their sheer amount of required devices, they have to be in the low cost range.
As a goal, each module shall be in the price range of less than 5 US$.
In order to evaluate possible solutions, a study item was discussed in 3GPP in the
GERAN TSG [1]. The main requirement in addition to the above ones was the coexistence with existing GSM, UMTS and LTE systems and the hardware used for those
technologies.
Figure 1-1: The three different solutions for specifying an optimized internet of things standard
Two solutions, the NB-IoT and the EC-GSM have been identified from this study,
where the latter is building upon the GSM standard. In parallel, also a pure LTE solu-
Introduction
tion, LTE-M, was brought into 3GPP. It continues the optimizations already done in
Release 12 with the introduction of a new device category cat-M1.
In this whitepaper, the NB-IoT is presented. Although it is integrated in the LTE standard, it can be regarded as a new air interface. Therefore, it is not backward compatible with LTE. The coexistence is realized by specifying the time and frequency resources used from the existing standards, or in the neighborhood thereof.
The whitepaper is structured in the following way: We first give an overview of the
more specific requirements and the network architecture and provide the details of the
physical layer. After having described the access to the cell, we show how the data
packets are transported over the air interface. Finally, an outlook to further developments of this technology is given.
Overview
Network
2 Overview
2.1 Requirements
From the general MTC requirements mentioned in the previous chapter, the following
standard specific requirements for NB-IoT were derived:
Minimize the signaling overhead, especially over the radio interface
Appropriate security to the complete system, including the core network
Improve battery life
Support delivery of IP and non-IP data [2, 3]
Support of SMS as a deployment option [4]
In order to fulfill these requirements, many advanced and even basic features of LTE
Release 8/9 are not supported [5]. The most striking example is the lack of handover
for UEs in the connected state. Only cell reselection in the idle state is supported,
which is even restricted to be within the NB-IoT technology. As there is no interaction
with other radio technologies, also the associated features are not supported. Examples are the lack of LTE-WLAN interworking, interference avoidance for in-device coexistence, and measurements to monitor the channel quality.
Most LTE-Advanced features are also not supported. This concerns e.g. Carrier Aggregation, Dual Connectivity, or device-to-device services. In addition, there is no QoS
concept, because NB-IoT is not used for delay sensitive data packets. Consequently,
all services requiring a guaranteed bit rate, like real time IMS, are not offered in NB-IoT
as well.
With these requirements, 3GPP uses a different approach than before. Instead of creating one air-interface for all types of applications, the air-interface for small non-delay
sensitive data packets is split off and optimized separately. UEs which support to work
on NB-IoT technology are tagged with the new UE category cat-NB1.
2.2 Network
2.2.1 Core Network

In order to send data to an application, two optimizations for the cellular internet of
things (CIoT) in the evolved packet system (EPS) were defined, the User Plane CIoT
EPS optimisation and the Control Plane CIoT EPS optimisation, see Figure 2-1. Both
optimisations may be used but are not limited to NB-IoT devices.
Overview
Network
Figure 2-1: Network for the NB-IoT data transmission and reception. In red, the Control Plane CIoT
EPS optimisation is indicated, in blue the User Plane CIoT EPS optimisation.
On the Control Plane CIoT EPS optimisation, UL data are transferred from the eNB
(CIoT RAN) to the MME. From there, they may either be transferred via the Serving
Gateway (SGW) to the Packet Data Network Gateway (PGW), or to the Service Capability Exposure Function (SCEF) which however is only possible for non-IP data packets. From these nodes they are finally forwarded to the application server (CIoT Services). DL data is transmitted over the same paths in the reverse direction. In this solution, there is no data radio bearer set up, data packets are sent on the signaling radio
bearer instead. Consequently, this solution is most appropriate for the transmission of
infrequent and small data packets.
The SCEF is a new node designed especially for machine type data. It is used for
delivery of non-IP data over control plane and provides an abstract interface for the
network services (authentication and authorization, discovery and access nework
capabilities).
With the User Plane CIoT EPS optimisation, data is transferred in the same way as the
conventional data traffic, i.e. over radio bearers via the SGW and the PGW to the application server. Thus it creates some overhead on building up the connection, however it
facilitates a sequence of data packets to be sent. This path supports both, IP and nonIP data delivery.
2.2.2 Access Network

On the overal access network architecture there is no difference to LTE [6]:
Overview
Frequency Bands
Figure 2-2: Network architecture towards the air--interface
The eNBs are connected to the MME and S-GW using the S1 interface, with the difference of carrying the NB-IoT messages and data packets. Even though there is no
handover defined, there is still an X2 interface between two eNBs, which enables a fast
resume after the UE goes to the idle state, see Chapter 4.5.1, "RRC Connection Establishment", on page 28 for details, even for the case that the resume process is to
another eNB.
2.3 Frequency Bands

For the frequency bands, the same frequency numbers as in LTE are used, with a subset defined for NB-IoT. In Release 13, these are the following bands [7]:
Band Number
Uplink frequency range / MHz
Downlink frequency range / MHz
1920 - 1980
2110 - 2170
1850 - 1910
1930 - 1990
1710 - 1785
1805 - 1880
824 - 849
869 - 894
880 - 915
925 - 960
12
699 - 716
729 - 746
13
777 - 787
746 - 756
17
704 - 716
734 - 746
18
815 - 830
860 - 875
19
830 - 845
875 - 890
20
832 - 862
791 - 821
26
814 - 849
859 - 894
28
703 - 748
758 - 803
66
1710 - 1780
2110 - 2200
Overview
Frequency Bands
It is worth mentioning that most frequencies are in the lower range of existing LTE
bands. This reflects that for machine type communications there are a lot of devices
expected in difficult radio conditions.
Physical Layer
Operation Modes
3 Physical Layer
3.1 Operation Modes
NB-IoT technology occupies a frequency band of 180 kHz bandwidth [8], which corresponds to one resource block in LTE transmission. With this selection, the following
operation modes are possible:
Stand alone operation. A possible scenario is the utilization of currently used GSM
frequencies. With their bandwidth of 200 kHz there is still a guard interval of 10 kHz
remaining on both sides of the spectrum
Guard band operation, utilizing the unused resource blocks within an LTE carrier's
guard-band
In-band operation utilizing resource blocks within an LTE carrier
These modes are visualized in the following figure:
Figure 3-1: Operation modes for NB-IoT
For the stand alone operation, the GSM carriers in the right part of the figure are only
shown as an example in order to indicate that this is a possible NB-IoT deployment. Of
course, this operation mode also works without neighboring GSM carriers.
In the in-band operation, the assignment of resources between LTE and NB-IoT is not
fixed. However, not all frequencies, i.e. resource blocks within the LTE carrier, are
allowed to be used for cell connection. They are restricted to the following values:
Table 3-1: Allowed LTE PRB indices for cell connection in NB-IoT in-band operation
LTE system
bandwidth
3 MHz
5 MHz
10 MHz
15 MHz
20 MHz
LTE PRB indices

for NB-IoT synchronization
2, 12
2, 7, 17,
22
4, 9, 14, 19,
30, 35, 40, 45
2, 7, 12, 17, 22,

27, 32, 42, 47,
52, 57, 62, 67, 72
4, 9, 14, 19, 24, 29, 34,

39, 44, 55, 60, 65, 70, 75,
80, 85, 90, 95
As indicated in this table, there is no support for in-band operation of an LTE band with
1.4 MHz bandwidth. A conflict between resources used by the LTE system like the cell
specific reference signals (CRS) or the downlink control channel at the start of each
subframe must be taken into account when resources are allocated for NB-IoT. This is
also reflected in Table 3-1 by not using the 6 inner resouce blocks, as these are allocated for the synchronization signals in LTE.
Physical Layer
Downlink
For the guard band operation, the UE only synchronizes to signals, for which the bands
are completely in the guard band.
In order to cope with different radio conditions, there may be up to 3 coverage
enhancement (CE) levels, CE level 0 to CE level 2. CE level 0 corresponds to normal
coverage, and CE level 2 to the worst case, where the coverage may be assumed to
be very poor. It is up to the network, how many CE levels are defined. A list of power
thresholds for the received reference signals is broadcasted in the cell for each CE
level. The main impact of the different CE levels is that the messages have to be
repeated several times.
For Release 13, FDD half duplex type-B is chosen as the duplex mode. This means
that UL and DL are separated in frequency and the UE either receives or transmits,
however not simultaneously. In addition, between every switch from UL to DL or vice
versa there is at least one guard subframe (SF) in between, where the UE has time to
switch its transmitter and receiver chain.
3.2 Downlink
For the DL, three physical channels
NPBCH, the narrowband physical broadcast channel
NPDCCH, the narrowband physical downlink control channel
NPDSCH, the narrowband physical downlink shared channel
and two physical signals
NRS, Narrowband Reference Signal
NPSS and NSSS, Primary and Secondary Synchronization Signals
are defined. These are less channels than for LTE, the physical multicast channel
PMCH is not included, because there is no MBMS service for NB-IoT.
The following figure illustrates the connection between the transport channels and the
physical channels:
Figure 3-2: Mapping of the transport channels to the physical channels
10
Physical Layer
Downlink
MIB information is always transmitted over the NPBCH, the remaining signaling information and data over the NPDSCH. The NPDCCH controls the data transfer between
UE and eNB.
The physical DL channels are always QPSK modulated. NB-IoT supports the operation
with either one or two antenna ports, AP0 and AP1. For the latter case, Space Frequency Block Coding (SFBC) is applied. Once selected, the same transmission
scheme applies to NPBCH, NPDCCH, and NPDSCH.
Like in LTE, each cell has an assigned physical cell ID (PCI), the Narrowband physical
cell ID (NCellID). Totally 504 different values for NCellID are defined. Its value is provided by the secondary synchronization signal NSSS, see Chapter 3.2.3, "Synchronization Signals", on page 13.
3.2.1 Frame and Slot Structure

In the DL, OFDM is applied using a 15 kHz subcarrier spacing with normal cyclic prefix
(CP). Each of the OFDM symbols consists of 12 subcarrier occupying this way the
bandwitdh of 180 kHz. Seven OFDMA symbols are bundled into one slot, so that the
slot has the following resource grid [9]:
Figure 3-3: Resource grid for one slot. There are 12 subcarriers for the 180 kHz bandwidth.
This is the same resource grid as for LTE in normal CP length for one resource block,
which is important for the in-band operation mode. A resource element is defined as
one subcarrier in one OFDMA symbol and is indicated in Figure 3-3 by one square.
Each of these resource elements carries a complex value with values according to the
modulation scheme.
These slots are summed up into subframes and radio frames in the same way as for
LTE:
11
Physical Layer
Downlink
Figure 3-4: Frame structure for NB-IoT for DL and UL with 15kHz subcarrier spacing
There are 1024 cyclically repeated radio frames, each of 10ms duration. A radio frame
is partitioned into 10 SFs, each one composed of two slots.
In addition to the system frames, also the concept of hyper frames is defined, which
counts the number of system frame periods, i.e. it is incremented each time the system
frame number wraps. It is a 10 bit counter, so that the hyper frame period spans 1024
system frame periods, corresponding to a time interval of almost 3 hours.
3.2.2 Narrowband Reference Signal

The narrowband reference signal (NRS) is transmitted in all SFs which may be used
for broadcast or dedicated DL transmission, no matter if data is actually transmitted or
not, see Chapter 3.2.5, "Dedicated Channels", on page 16 for more details.
Depending on the transmission scheme, NRS is either transmitted on one antenna port
or on two. Its values are created like the CRS in LTE, with the NCellID taken for the
PCI. The mapping sequence is shown in the following figure:
Figure 3-5: Basic mapping of reference signals to the resource elements. In blue, NRS transmitted on
antenna port 0, in magenta, NRS transmitted on antenna port 1.
The NRS mapping shown in Figure 3-5 is additionally cyclically shifted by NCellID mod
6 in the frequency range. When NRSs are transmitted on two APs, then on every
resource element used for NRS on AP0, the same resource element on AP1 is set to
zero and vice versa.
12
Physical Layer
Downlink
For the in-band operation, the LTE CRS are also transmitted in the NB-IoT bands for
the SFs which are not used for MBSFN. With the structure of the NRS there is no overlap between the LTE CRS and the NRS, however the CRS have to be taken into
account for rate matching and resource element mapping. All DL transmissions must
not use these resource elements and have to skip them.
An important point on the in-band operation concerns the NcellID. It may be the same
as the PCI for the embedding LTE cell or not. This is indicated by the opeartionMode
parameter in MIB-NB, see Chapter 3.2.4, "Narrowband Physical Broadcast Channel",
on page 14, which distinguishes between in-band operation with same PCI as true or
false. If this parameter is set to true, then NCellID and PCI are the same and the UE
may assume that the number of antenna ports is the same as in the LTE cell. The
channel may then be inferred from either reference signal set. Therefore, LTE CRS
port 0 is associated with NRS port 0, and CRS port 1 is associated with NRS port 1. If
same-PCI is set to false, the UE may not take any of these assumptions.
3.2.3 Synchronization Signals

For a first synchronization in frame and subframe and in order to determine the
NCellID, the LTE concept of Primary Synchronization Signal (PSS) and Secondary
Synchronization Signal (SSS) is reused. With these signals, also timing and frequency
estimation may be refined in the UE receiver.
In order to distinguish these signals from their LTE counterparts, they are denoted as
NPSS and NSSS, respectively. Their structure is depicted in the following figure:
Figure 3-6: Primary and secondary synchronization signals indicated in light blue and green, respectively. In violet, LTE CRS locations are shown. In this example we assume a 4 antenna
port CRS transmission. NRS are not transmitted in the NPSS and NSSS subframes.
The first 3 OFDM symbols are left out, because they may carry the PDCCH in LTE
when NB-IoT is operated in the in-band mode. Note that during the time when the UE
synchronizes to the NPSS and NSSS, it may not know the operation mode, consequently this guard time applies to all modes. In addition, both synchronization signals
are punctured by the LTE's CRS. It is not specified, which of the antenna ports is used
for the synchronization signals, this may even change between any two SFs.
13
Physical Layer
Downlink
A length 11 Zadoff-Chu sequence in frequency domain is taken for the sequence generation of the NPSS. This sequence is fixed and therefore carries no information about
the cell. It is transmitted in SF5 of each radio frame, so that its reception allows the UE
to determine the frame boundary.
The NSSS sequence is generated from a length-131 frequency domain Zadoff-Chu
sequence, binary scrambled and cyclically shifted depending on the radio frame number. NCellID is an additional input parameter so that it can be derived from the
sequence. Like in LTE, 504 PCI values are defined. NSSS are transmitted in the last
SF of each even numbered radio frame.
For the in-band operation, transmission of the NPSS and NSSS, as well as the NPBCH
described in the next section, may only be done on PRBs as indicated in Table 3-1.
The carrier selected for receiving this information is called anchor carrier.
Using this construction, the UE can not confuse the NB synhronization signals with
those transmitted by the LTE system. Consequently, there is no danger of a false
detection and UEs with either technology are automatically routed to the correct frequency range.
3.2.4 Narrowband Physical Broadcast Channel

NPBCH carries the Narrowband Master Information Block (MIB-NB). The MIB-NB contains 34 bits and is transmitted over a time period of 640ms, i.e. 64 radio frames. The
following information is provided therein:
4 bits indicating the most significant bits (MSBs) of the System Frame Number
(SFN), the remaining least significant bits (LSBs) are implicitely derived from the
MIB-NB start
2 bits indicating the two LSBs of the hyper frame number
4 bits for the SIB1-NB scheduling and size
5 bits indicating the system information value tag
1 bit indicating whether access class barring is applied
7 bits indicating the operation mode with the mode specific values
11 spare bits for future extensions
Figure 3-7 shows its mapping to physical resources:
14
Physical Layer
Downlink
Figure 3-7: NPBCH mapping to the subframes
After physical layer baseband processing, the resulting MIB-NB is split into 8 blocks.
The first block is transmitted on the first subframe (SF0) and repeated in SF0 of the
next 7 consecutive radio frames, respectively. In SF0 of the following radio frame, the
same procedure is done for BL2. This process is continued until the whole MIB-NB is
transmitted. By using SF0 for all transmissions, it is avoided that NPBCH collides with
a potential MBSFN transmission on LTE, if NB-IoT is deployed as in-band operation.
The SF structure of the NPBCH is shown in the following figure:
Figure 3-8: Resource element occupation of the NPBCH indicated in yellow. In magenta, the NRS are
shown, and in violet, the CRS.
Symbols are mapped around the NRS and the LTE CRS, where it is always assumed
that two antenna ports are defined for NRS and 4 antenna ports for CRS. This
assumption is necessary, because the UE gets the actual antenna port information
only from reading the MIB-NB. The reference signal location in the frequency range is
given by the NCellID, provided by the NSSS. Although the NCellID may be different to
the PCI in the in-band operation, its range is restricted so that it points to the same frequency locations, hence the CRS's cyclic shift in the frequency range is known to the
UE. Again, the first 3 OFDM symbols are left out in order to avoid a possible conflict
with the LTE's control channel.
15
Physical Layer
Downlink
3.2.5 Dedicated Channels

The principle of control and shared channel also applies for NB-IoT, defining the Narrowband Physical Downlink Control Channel (NPDCCH) and the Narrowband Phsyical
Downlink Shared Channel (NPDSCH). Not all SFs may be used for transmission of the
dedicated DL channels. In RRC signaling, a bitmap of either 10 or 40 bit indicating the
valid SFs may be signaled, which is applied in a periodic way. For the case that a SF is
not indicated as valid, dedicated DL channel transmission is postponed until the next
valid SF.
3.2.5.1
Control Channel NPDCCH

The NPDCCH indicates for which UE there is data in the NPDSCH, where to find them
and how often they are repeated. Also, the UL grants are provided therein, showing the
resources the UE shall use for data transmission in the UL. Finally, additional information like paging or system information update is contained in the NPDCCH as well.
The NPDCCH subframe design is depicted in the following figure:
Figure 3-9: Resource elements used by NPDCCH (light and dark green), LTE CRS (violet) and NRS
(blue). This example shows the mapping for an in-band operation assuming a single
antenna port in the LTE cell and two antenna ports in NB-IoT.
Resource elements used for NPDCCH are indicated in green color. They have to be
mapped around the NRS shown in blue, and for the in-band operation also around the
CRS (violet). A parameter lNstart, the control region size signaled by NB-SIB1, indicates
the OFDM start symbol. This way conflict with the LTE control channel can be avoided
for the in-band operation. For the guard-band and standalone operation modes, the
control region size is by default 0, which provides more resource elements for the
NPDCCH.
On each SF, two narrowband control channel elements (NCCEs) are defined, NCCE0
and NCCE1. They are indicated in Figure 3-9 with dark green color (NCCE0), and light
green (NCCE1). Two NPDCCH formats are defined to use them:
NPDCCH format 0 taking one NCCE. Consequently, two of them can be transmitted within a SF.
NPDCCH format 1 taking both NCCEs.
In order for the UE to find the control information with a reasonable amount of decoding complexity, NPDCCH is grouped into the following search spaces:
16
Physical Layer
Downlink
Type-1 common search space, used for paging
Type-2 common search space, used for random access
UE specific search space
Each NPDCCH may be repeated several times with an upper limit configured by the
RRC. In addition, Type-2 common search space and UE specific search space are
provided by RRC, whereas the Type-1 common search space is given by the paging
opportunity SFs, see Chapter 4.7, "Paging", on page 32.
Different radio network temporary identifier (RNTI) are assigned to each UE, one for
random access (RA-RNTI), one for paging (P-RNTI), and a UE specific identifier (CRNTI) provided in the random access procedure. These identifiers are implicitely indicated in the NPDCCH's CRC. So, the UE has to look in its search space for that RNTI,
and, if found, decodes the NPDCCH.
Three DCI formats are defined in Release 13, DCI format N0, N1 and N2 [10]:
DCI Format
SIze / bit
Content
N0
23
UL grant
N1
23
NPDSCH scheduling
RACH procedure initiated by NPDCCH order
N2
15
Paging and direct indication
When the UE receives the NPDCCH, it can distinguish the different formats in the following way: DCI format N2 is implicitely indicated in the way that the CRC is scrambled
with the P-RNTI. If the CRC is scrambled with the C-RNTI, then the first bit in the message indicates whether DCI format N0 or N1 is contained. For the case that the CRC is
scrambled with the RA-RNTI, the content is a restricted DCI format N1 including only
those fields required for the RACH response.
Included in the DCI formats N0 and N1 is the the scheduling delay, i.e. the time
between the NPDCCH end and the NPDSCH start or the NPUSCH start. This delay is
at least 5 SFs for the NPDSCH and 8 for the NPUSCH. For DL transmission via DCI
format N2, the scheduling delay is fixed to 10 SFs.
3.2.5.2
Traffic Channel NPDSCH

An NPDSCH SF has the same structure as for the NPDCCH shown in Figure 3-9. It
starts at a configurable OFDM symbol lNstart and is mapped around the NRS and, for inband operation, the LTE CRS. lNstart is provided by RRC signaling for the in-band operation, and is 0 otherwise.
A maximum transport block size (TBS) of 680 bit is supported. The mapping of a transport block spans NSF SFs. The transport block is repeated providing NRep identical copies, using an SF interleaving for an optimized reception at the UE. Both values, NSF
and NRep are indicated in the DCI. The resulting SF sequence is mapped to NSFNRep
consecutive SFs defined for NPDSCH.
For the DL there is no automatic acknowledgement to a transmission, the eNB indicates this in the DCI. If this is done, the UE transmits the acknowledgement using
17
Physical Layer
Uplink
NPUSCH format 2, see Chapter 3.3.2, "Physical Uplink Shared Channel",

on page 19. The associated timing and subcarrier is indicated in this DCI as well.
There is multi-carrier support for all operation modes, see Chapter 5.2.2, "Multi Carrier
Configuration", on page 35, which means that another carrier may be used when the
UE is in the connected state. In the idle state, the UE camps on the NB-IoT carrier from
which it received the synchronization signals and broadcast information, i.e. the anchor
carrier. It waits there for paging or starts access for mobile originated data or signaling,
both by transmitting a preamble in the associated UL carrier provided in SIB2-NB.
SIB1-NB Transmission
SIB1-NB is transmitted over the NPDSCH. Its has a period of 256 radio frames and is
repeated 4, 8 or 16 times. The transport block size and the number of repetitions is
indicated in the MIB-NB. 4, 8 or 16 repetitions are possible, and 4 transport block sizes
of 208, 328, 440 and 680 bits are defined, respectively. The radio frame on which the
SIB1-NB starts is determined by the number of repetitions and the NCellID. SF4 is
used for SIB1-NB in all radio frames transmitting SIB1-NB. As the other transmission
parameters are also fixed, there is no associated indication in the control channel.
SIB1-NB content may only be changed on each modification period, which has a
length of 4096 radio frames, i.e. 40.96 seconds. This corresponds to 4 SFN periods,
which is why the 2 LSBs of the hyper frame number are indicated in MIB-NB. If such a
modification occurs, it is indicated in the NPDCCH using DCI format N2.
Although sent over the NPDSCH, SIB1-NB resources are mapped like the MIB-NB
shown in Figure 3-8, i.e. leaving out the first 3 OFDM symbols. This is necessary,
because the UE knows the start of the resource mapping from SIB1-NB, so it needs to
decode this SIB first.
3.3 Uplink
For the uplink (UL), the two physical channels
NPUSCH, the narrowband physical uplink shared channel
NPRACH, the narrowband physical random access channel
and the
DMRS, Demodulation Reference Signal
are defined. The connection between the physical channels and the associated transport channels is depicted in the following figure:
18
Physical Layer
Uplink
Figure 3-10: Mapping between the transport channels and the physical channels in UL
Except for RACH transmission, all data are sent over the NPUSCH. This includes also
the UL control information (UCI), which is transmitted using a different format. Consequently there is no equivalent to the PUCCH in LTE.
3.3.1 Slot Structure

In the UL, Single Carrier Frequency Division Multiple Access (SC-FDMA) is applied,
either with a 3.75 kHz or 15 kHz subcarrier spacing. eNB decides which one to use.
The resource grid for the UL is the same as for the DL, when the 15kHz subcarrier
spacing is applied, see Figure 3-3. For the 3.75kHz subcarrier spacing, the resource
grid for a slot has a modified structure:
Figure 3-11: Resource grid for 3.75 kHz subcarrier spacing. There are 48 subcarrier for the 180 kHz
bandwidth
Again there are 7 OFDM symbols within a slot. According to the OFDM principles, the
symbol duration for 3.75 kHz subcarrier spacing has four times the duration compared
to 15 kHz, which results in a slot length of 2 ms.
3.3.2 Physical Uplink Shared Channel

On the physical UL shared channel NPUSCH, two formats are defined. NPUSCH format 1 is for UL transport channel data over the UL-SCH, with transport blocks not big-
19
Physical Layer
Uplink
ger than 1000 bits. NPUSCH format 2 carries UL control information (UCI), which in
Release 13 is restricted to an acknowledgement of a DL transmission.
The smallest unit to map a transport block is the resource unit (RU). Its definition
depends on the PUSCH format and subcarrier spacing.
For NPUSCH format 1 and 3.75 kHz subcarrier spacing, an RU consists of 1 subcarrier
in the frequency range, and 16 slots in the time range, i.e. an RU has a length of 32
ms. On the 15 kHz subcarrier spacing there are 4 options:
Number of
subcarriers
Number of
slots
RU Duration
16
8 ms
4 ms
2 ms
12
1 ms
For NPUSCH format 2, the RU is always composed of one subcarrier with a length of 4
slots. Consequently, for the 3.75 kHz subcarrier spacing the RU has an 8 ms duration
and for the 15 kHz subcarrier spacing 2 ms.
For NPUSCH format 2, the modulation scheme is always BPSK. The allowed modulations for NPUSCH format 1 depend on the selected RU:
For RUs with one subcarrier, BPSK and QPSK may be used
For all other RUs, QPSK is applied
A grant for UL-SCH transmission is indicated in the NPDCCH via DCI format N0. The
start time of the NPUSCH, the number of repetitions, the number of RUs used for one
transport block, and the number of subcarriers including their position in the frequency
range are indicated in this DCI. Also the MCS index is contained, providing the modulation scheme for the one subcarrier RUs and in addition, together with the number of
RUs, the transport block size.
Finally, the time signal is created by applying an inverse fourier transrformation and
prepending a cyclic prefix (CP). For the 15kHz subcarrier spacing, this CP is the same
as for LTE using normal CP, while for 3.75kHz it is 256 samples, corresponding to 8.3
s. For the latter case, a period of 2304 samples (75s) at the end of each slot remains
empty, which is used as a guard interval. For the in-band operation, this guard interval
may be used to transmit Sounding Reference Signals in the LTE system.
Contrary to the DL transmission, where it is configurable whether a transmission shall
be acknowledged, there is always an acknowledgement in the associated DL.
3.3.3 Reference Signals

In UL, demodulation reference signal (DMRS) is defined. It is multiplexed with the data
so that it is only transmitted in RUs containing data transmission. There is no MIMO
transmission defined for the UL, consequently all transmissions use a single antenna
port.
20
Physical Layer
Uplink
Depending on the NPUSCH format, DMRS is transmitted in either one or three SCFDMA symbols per slot. For NPUSCH format 1 these are the symbols indicated in red
in Figure 3-12.
Figure 3-12: Resource elements used for demodulation reference signal in NPUSCH format 1. This
figure shows the example of an RU occupation with 6 subcarrier for the 15 kHz subcarrier spacing.
As one can see from this figure, the SC-FDM symbols for DMRS transmission depend
on the subcarrier spacing. This is also true for NPUSCH format 2 shown in Figure 3-13:
Figure 3-13: Resource elements used for demodulation reference signals in NPUSCH format 2. In this
format, the RU generally occupies only one subcarrier.
DMRS symbols are constructed from a base sequence multiplied by a phase factor.
They have the same modulation as the associated data. For NPUSCH format 2 DMRS
symbols are spread with the same orthogonal sequence as defined for the LTE
PUCCH formats 1, 1a and 1b.
3.3.4 Random Access Channel

In the Random Access Channel, NPRACH, a preamble is transmitted. The associated
random access procedure, described in Chapter 4.4, "Random Access Procedure",
on page 27, may be used to signal to the cell that the UE is camping on it and wants
to get access.
The preamble is based on symbol groups on a single subcarrier. Each symbol group
has a cyclic prefix (CP) followed by 5 symbols. The following structure shows this
sequence:
Figure 3-14: Preamble symbol group
Two preamble formats are defined, format 0 and format 1, which differ in their CP
length. The five symbols have a duration of TSEQ = 1.333 ms, prepended with a CP of
TCP = 67 s for format 0 and 267 s for format 1, giving a total length of 1.4 ms and 1.6
ms, respectively. The preamble format to be used is broadcasted in the system information.
21
Physical Layer
Uplink
The preamble is composed of 4 symbol groups transmitted without gaps. Frequency

hopping is applied on symbol group granularity, i.e. each symbol group is transmitted
on a different subcarrier. By construction, this hopping is restricted to a contiguous set
of 12 subcarrier. Depending on the coverage level, the cell may indicate that the UE
shall repeat the preamble 1, 2, 4, 8, 16, 32, 64, or 128 times, using the same transmission power on each repetition.
NPRACH resources are provided for each CE group separately. They consist in the
assignment of time and frequency resources and occur periodically, where NPRACH
periodicities between 40 ms and 2.56 s may be configured. Their start time within a
period is provided in system information. The number of repetitions and the preamble
format determine their end.
In the frequency range, subcarrier spacing of 3.75 kHz is applied. NPRACH resources
occupy a contiguous set of either 12, 24, 36 or 48 subcarriers and are located on a discrete set of subcarrier ranges. Depending on the cell configuration, the resources may
be further partitioned into resources used by UEs supporting multi-tone transmission
for msg3 and UEs that do not support it.
Figure 3-15 shows an example of a preamble repeated at least 4 times. Here, each
blue rectangle describes one preamble symbol group as depicted in Figure 3-14,
hence a preamble repetition consists of four rectangles.
Figure 3-15: Preamble sequence in a frequency range between subcarrier 12 and 23
Among the 12 possibilities, the UE selects the subcarrier for the transmission of the
first preamble symbol group, if not provided by the eNB for the case of an ordered preamble transmission. The next 3 symbol groups are determined by an algorithm which
depends only on the location of the first one. For the subcarrier selection of the first
symbol group of the next repetition, a pseudo-random hopping is applied, where
NCellID and the repetition number are used as input. The subcarrier selection of the
following symbol groups again depend only on this result.
This frequency hopping algorithm is designed in a way that different selections of the
first subcarrier lead to hopping schemes which never overlap. Hence there are as
many different congestion free preambles as there are subcarrier allocated to the
22
Physical Layer
Power Control
NPRACH. No further partitioning is done for NB-IoT, i.e. there is no concept like the
preamble indices applied in LTE.
The preamble sequence is built upon a Zadoff-Chu sequence, which depends on the
subcarrier location. Modulation and upconversion to the carrier frequency is done in
the same way as for LTE.
3.4 Power Control
3.4.1 Uplink
In the UL, the transmitted power depends on a combination of cell specific parameters,
the selected RU and UE measured parameters [11]. For the case that there are maximally 2 repetitions, the power on slot i is given by
If there are more than two repetitions, the transmission power is generally given by
PCMAX,c(i).
PCMAX,c(i) is the cell specific maximum transmit power on slot i. With the above construction, the transmitted power may never exceed this threshold. MNPUSCH,c depends
on the bandwidth of the selected RU and the subcarrier spacing, PO_NPUSCH,c a combination of different parameters signaled by the RRC, which depends on whether the
transport block is for UL-SCH data (j=1) or for the RACH message (j=2). PLc is the
path loss estimated by the UE. This factor is weighted by c (j), which for NPUSCH format 1 is provided by RRC, otherwise the fixed value of 1 is applied. In other words, this
factor indicates how strong the path loss shall be compensated.
3.4.2 Downlink
DL transmission power refers to the NRS transmission power. Its value is indicated to
the UE in order to estimate the path loss. It is constant for all resource elements carrying the NRS and all SFs.
For the NPBCH, NPDCCH and NPDSCH the transmit power depends on the transmission scheme. If only one antenna port is applied, the power is the same as for the
NRS, otherwise it is reduced by 3dB.
A special case occurs if the in-band operation mode is used and the samePCI value is
set to true. Then the eNB may additionally signal the ratio of the NRS power to the
CRS power, enabling the UE to use the CRS for channel estimation as well.
23
Cell Access
Protocol Stack and Signaling Bearer
4 Cell Access
When a UE accesses a cell, it follows the same principle as for LTE: It first searches a
cell on an appropriate frequency, reads the associated SIB information, and starts the
random access procedure to establish an RRC connection. With this connection it registers with the core network via the NAS layer, if not already done. After the UE has
returned to the RRC_IDLE state, it may either use again the random access procedure
if it has mobile originated data to send, or waits until it gets paged.
4.1 Protocol Stack and Signaling Bearer

The general principle for the protocol layers is to start with the LTE protocols, reduce
them to a minimum and enhance them as needed for NB-IoT. This way, the proven
structures and procedures are re-used while overhead from unused LTE features is
prevented. Consequently, the NB-IoT technology can be regarded as a new air interface also from the protocol stack point of view, while being built on a well established
fundament.
One example thereof is the bearer structure. Signaling radio bearer are partly re-used
from LTE. There is the SRB0 for RRC messages transmitted over the CCCH logical
channel, and the SRB1 for RRC messages and NAS messages using the DCCH logical channel. However, there is no SRB2 defined.
In addition, a new signaling radio bearer, the SRB1bis is defined. It is implicitely configured with SRB1 using the same configuration, however without the PDCP. This channel takes the role of the SRB1 until security is activated, then SRB1bis is not used anymore. This also implies that for the Control Plane CIoT EPS optimisation, only SRB1bis
is used at all, because there is no security activation in this mode.
The protocol stacks are the same as for LTE with functionalities optimized for NB-IoT:
Figure 4-1: Protocol stacks for NB-IoT
24
Cell Access
System Information
4.2 System Information

Like in LTE, system information is used to broadcast information which is valid for all
UEs within a cell. As broadcasting system information takes resources and causes battery consumption for each UE, it is kept to a minimum, as well in its size as in its occurence.
Consequently, a set of System Information Blocks (SIBs) is defined, which is a subset
of the SIBs defined for LTE. These are shown in the following table:
System Information Block
MIB-NB
Content
Essential information required to receive further system information
SIBType1-NB
Cell access and selection, other SIB scheduling
SIBType2-NB
Radio resource configuration information
SIBType3-NB
Cell re-selection information for intra-frequency, interfrequency
SIBType4-NB
Neighboring cell related information relevant for intrafrequency cell re-selection
SIBType5-NB
Neighboring cell related information relevant for interfrequency cell re-selection
SIBType14-NB
Access Barring parameters
SIBType16-NB
Information related to GPS time and Coordinated Universal Time (UTC)
SIBs are indicated with the suffix NB. Each of these SIBs is defined with a reduced and
modified set of information elements, however, the type of content is the same as in
LTE, e.g. SIB16-NB describes the time information. UEs exclusively use these SIBs
and ignore those from LTE, even in the case of in-band operation.
It is always mandatory for a UE to have a valid version of MIB-NB, SIB1-NB and SIB2NB through SIB5-NB. The other ones have to be valid if their functionality is required
for operation. For instance, if access barring (AB) is indicated in MIB-NB, the UE needs
to have a valid SIB14-NB.
System information acquisition and change procedure is only applied in the RRC_IDLE
state. The UE is not expected to read SIB information while being in the RRC_CONNECTED state. If a change occurs, the UE is informed either by paging or direct indication. The eNB may also release the UE to the RRC_IDLE state for the purpose of
acquiring modified system information.
4.2.1 Scheduling
The MIB-NB and SIB1-NB are transmitted as indicated in Chapter 3.2, "Downlink",
on page 10. Scheduling of the remaining SIB information is done in an analogous way
to LTE: SIB-NB messages are grouped into SI messages, which are then transmitted
in separate SI-windows. The SI windows of different SI messages do not overlap. Their
25
Cell Access
Cell Selection and Mobility
length is indicated in SIB1-NB and is the same for all SI messages. The maximum size
of each SIB and SI message is 680 bit.
Within an SI window, the SI messages are sent over 2 or 8 consecutive valid DL SFs,
depending on their transport block size, and may be repeated several times. Scheduling information is indicated in SIB1-NB, consequently there is no indication in the
NPDCCH necessary and there is no SI-RNTI needed.
4.2.2 Change Notification

As already mentioned in "SIB1-NB Transmission" on page 18 system information may
only change at the time boundaries of a modification period and is indicated with a
paging message. An exception is the change in the access barring (AB) indicated via
SIB14-NB, this information may be changed at any given point in time. The reason
behind this exception is that AB parameters may be required to change in a much
shorter time scale. Of course, SIB16-NB is not indicated by a paging notification, as it
changes its content regularly.
An additional way to indicate changes in SIB1-NB or in SI messages is the concept of
the value tag. The associated field systemInfoValueTag is contained in MIB-NB. This
concept is used for UEs returning from an out-of-coverage location back to coverage
and for returning from a longer DRX cycle. In these cases the UE could not receive the
paging message, so it checks the value tag. If there were SIB changes, the value tag is
modified. However, the UE must read the system information anyway if it was out-ofcoverage for more than 24 hours.
4.2.3 Summarizing SIB Acquisition

The UE first obtains the NCellID from the NSSS. By reading the schedulingInfoSIB1 in
MIB-NB it knows SIB1-NB size and number of repetitions, and can infer its starting
position. In SIB1, the location of the other SIB-NB messages are indicated. Finally, with
help of the 2 LSBs of the hyperframe number obtained from MIB-NB, the UE knows
when to check SIB updates, if a SIB change is indicated either by a modified value tag
or by paging.
4.3 Cell Selection and Mobility

NB-IoT is designed for infrequent and short messages between the UE and the network. It is assumed that the UE can exchange these messages while being served
from one cell, therefore, a handover procedure during RRC_CONNECTED is not needed. If such a cell change would be required, the UE has first go to the RRC_IDLE
state and re-select another cell therein.
For the RRC_IDLE state, cell re-selection is defined for both, intra frequency and inter
frequency cells [12]. Inter frequency refers here to the 180 kHz carrier, which means
that even if two carriers are used in the in-band operation embedded into the same
LTE carrier, this is still referred to as an inter frequency re-selection.
26
Cell Access
Random Access Procedure
In order to find a cell, the UE first measures the received power and quality of the NRS.
These values are then compared to cell specific thresholds provided by the SIB-NB.
The S-criteria states that if both values are above these thresholds, the UE considers
itself to be in coverage of that cell. If the UE is in coverage of one cell, it camps on it.
Depending on the received NRS power, the UE may have to start a cell re-selection.
The UE compares this power to a re-selection threshold, which may be different for the
intra-frequency and the inter-frequency case. All required parameters are received
from the actual serving cell, there is no need to read SIB-NBs from other cells.
Among all cells fulfilling the S-criteria, the UE ranks the cells with respect to the power
excess over another threshold. A hysteresis is added in this process in order to prevent
too frequent cell reselection, and also a cell specific offset may be applied for the intra
frequency case. Contrary to LTE, there are no priorities for the different frequencies.
The UE finally selects the highest ranked cell which is suitable, i.e. from which it may
receive normal service.
When the UE leaves RRC_CONNECTED, it does not necessarily select the same carrier to find a cell to camp on. The RRCConnectionRelease messsage may indicate the
frequency on which the UE first tries to find a suitable cell. Only if the UE does not find
a suitable cell on this frequency, it may also try to find one on different frequencies.
4.4 Random Access Procedure

The RACH procedure has the same message flow as for LTE, however, with different
parameters [13]:
Figure 4-2: Message flow for the RACH procedure. Each of these messages is repeated according to
the UEs coverage enhancement level.
For NB-IoT the RACH procedure is always contention based and starts with the transmission of a preamble as outlined in Chapter 3.3.4, "Random Access Channel",
on page 21. After the associated response from the eNB, a scheduled message, msg3,
is transmitted in order to start the contention resolution process. The associated contention resolution message is finally transmitted to the UE in order to indicate the succesful completion of the RACH procedure.
Upon transmission of the preamble, the UE first calculates its RA-RNTI from the transmission time. It looks then in the PDCCH for the DCI format N1 scrambled with the RA-
27
Cell Access
Connection Control
RNTI, in which the Random Access Response message is indicated. The UE expects
this message within the Response Window, which starts 3 SFs after the last preamble
SF and has a CE dependent length given in SIB2-NB.
If the preamble transmission was not successful, i.e. the associated Random Access
Response (RAR) message was not received, the UE transmits another one. This is
done up to a maximum number, which again is depending on the CE level. For the
case that this maximum number is reached without success, the UE proceeds to the
next CE level, if this level is configured. If the total number of access attempts is
reached, an associated failure is reported to the RRC.
With the RAR, the UE gets in addition to a temporary C-RNTI the timing advance command. Consequently, the following msg3 is already time aligned, which is necessary
for transmission over the NPUSCH. Further, the RAR provides the UL grant for msg3,
containing all relevant data for msg3 transmission.
The remaining procedure is done like in LTE, i.e. the UE sends an identification and
upon reception of the Contention Resolution indicating this identification the random
access procedure is successfully finalized.
4.5 Connection Control

With a system not supporting handover to a different technology , the state model of
the RRC becomes quite simple (Figure 4-3).
Figure 4-3: Model of the RRC states and their transitions
As in LTE, there are only two states, RRC_IDLE and RRC_CONNECTED. However,
there are no transitions to the associated UTRA and GSM states, because handovers
to these technologies are not supported. There is also no handover to LTE, because
LTE is regarded as a different RAT.
4.5.1 RRC Connection Establishment

The RRC Connection Establishment has the same message flow as for the LTE system:
28
Cell Access
Connection Control
Figure 4-4: RRC Connection Establishment. The message flow is like in LTE, however the content is
different
With the RRCConnectionRequest the UE indicates that it wants to connect to the network and for what purpose. This Establishment Cause is restricted to mobile originated
signaling, mobile originated data, mobile terminated access and exceptional reports.
There is no establishment cause for delay tolerant traffic, because in NB-IoT all traffic
is assumed to be delay tolerant. In addition to the establishment cause, the UE also
indicates its capability to support multi-tone traffic and multi carrier support. Although
the capabilities are generally signaled in an own procedure, see Chapter 4.6, "UE
Capability Transfer", on page 31, these capabilities have to be signaled already here
so that the eNB can apply them for following UL grants in this procedure.
Upon response with the RRCConnectionSetup message the eNB provides configuration of the signaling radio bearer (SRB1), up to 2 data radio bearer (DRB) and the protocols. Finally, in the RRCConnectionSetupComplete message the UE includes its
selected PLMN and MME, and can piggyback the first NAS message.
After the connection is set up for User Plane CIoT EPS optimisation, security and RRC
connection reconfiguration procedures are done in the same way as for LTE with functionalities restricted to NB-IoT. Also the RRC connection re-establishment procedure is
defined for this case. For Control Plane CIoT EPS optimisation, these procedures are
not applied.
When the eNB releases the connection, see Chapter 4.5.2, "RRC Connection
Release", on page 31, it may also suspend the UE. In this case, the UE transits to
the RRC_IDLE state and stores the current AS context. It may resume later the
RRC_CONNECTED state with that context. Radio bearer are automatically set up, and
security is activated with updated keys. In addition, parts of the AS context may be
changed. Obviously, this saves considerable signaling overhead for the transmission of
infrequent small data packets.
The RRC Connection Resume is depicted in the following figure:
29
Cell Access
Connection Control
Figure 4-5: RRC Connection Resume request accepted by the eNB
The resume request may only be applied, when the UE is configured for User Plane
CIoT EPS optimisation and is configured with at least one DRB. Upon reception of the
RRCConnectionResumeRequest, the eNB decides whether it accepts this request or
whether a conventional RRC Connection Setup shall be started. If the eNB does not
accept the resume request, it switches back to the connection request:
Figure 4-6: RRC Connection Resume request not accepted by the eNB
In this case, the UE releases the stored AS context and it is not possible anymore to
resume this AS context for a later connection.
When the eNB indicates in the MIB-NB that access is subject to access class barring
and broadcasts SIB14-NB, the UE has first to do an access barring check when it
wants the connection for mobile originated signaling or data, before it tries to establish
or resume an RRC connection. Commercially available UEs have an access class from
0 to 9. In SIB14-NB there is a associated bitmap containing one bit for each access
class. If the bit associated to the access class is set, then access to that cell is barred.
The UE has then to wait for an update of SIB14-NB to check again the actual barring
status. Note that this access barring check may be skipped for some exceptional data,
depending on the SIB14-NB settings.
If the request for connection or resume request is rejected, e.g. because there are no
free resources anymore, the eNB replies with an RRCConnectionReject instead. Then,
the UE has to wait for an amount of time provided by the reject message. This way, the
eNB can prevent an excessive jam when by any reason too many UEs start network
connection simultaneously. If the reject is for a resume procedure, the eNB indicates
30
Cell Access
UE Capability Transfer
whether the current UE context shall be released or kept further stored for a following
resume request.
4.5.2 RRC Connection Release

The RRC connection release is initiated by the eNB and is depicted in Figure 4-7.
Figure 4-7: RRC connection release, always triggered by the eNB
For User Plane CIoT EPS optimisation the eNB may indicate here the suspension of
the connection with the rrcSuspend flag. In this case, the UE stores the AS context and
may request an RRC connection resume as described above, otherwise the AS context is deleted and the UE may only get another RRC connection using the complete
RRC connection setup.
After this procedure has been finalized, the UE enters the RRC_IDLE state.
4.6 UE Capability Transfer

When the UE connects to the network, the eNB does neither know on which release
the UE is built upon, nor which of the optional features defined therein it supports. In
order to get this information the UE Capability Transfer procudure is defined, which is
shown in Figure 4-8:
Figure 4-8: UE capability transfer
UE capability transfer is always initiated by the eNB, because the UE cannot know
whether the eNB has this information already from the network or from a previous session.
The capabilities include the release the UE is built upon, the UE category, the list of
supported bands and the capability to set up multiple bearer. In addition, the UE may
indicate whether it supports multi carrier operation, and multi tone transmission in UL.
Also the maximal number of RoHC context sessions and the supported profiles may be
contained.
31
Cell Access
Paging
This message is usually considerably smaller than the corresponding LTE message,
because all LTE features which are not supported in NB-IoT, like further access technologies or carrier aggregation, are left out.
4.7 Paging
Paging is used to trigger an RRC connection and to indicate a change in system information for UE in RRC_IDLE mode.
A paging message is sent over the NPDSCH and may contain a list of UEs to be
paged and the information, whether paging is for connection setup or whether system
information has changed. Each UE which finds its ID in this list forwards to its upper
layer that it is paged, and may receive in turn the command to initialize an RRC connection. If system information has changed, the UE starts to read SIB1-NB and may
obtain from there the information, which SIBs have to be read again.
The UE in the RRC_IDLE state only monitors some of the SFs with respect to paging,
the paging occasions (PO) within a subset of radio frames, the paging frames (PF). If
coverage enhancement repetitions are applied, the PO refers to the first transmission
within the repetitions. The PFs and POs are determined from the DRX cycle provided
in SIB2-NB, and the IMSI provided by the USIM card. DRX is the discontinous reception of DL control channel used to save battery lifetime. Cycles of 128, 256, 512 and
1024 radio frames are supported, corresponding to a time interval between 1.28s and
10.24s.
Due to the fact that the algorithm to determine the PFs and POs also depends on the
IMSI, different UEs have different paging occasions, which are uniformly distributed in
time. It is sufficient for the UE to monitor one paging occasion within a DRX cycle, if
there are several paging occasions therein, the paging is repeated in every one of
them.
The concept of extended DRX (eDRX) may be applied for NB-IoT as well. This is done
using the hyper frames outlined in Chapter 3.2.1, "Frame and Slot Structure",
on page 11. If eDRX is supported, then the time interval in which the UE does not monitor the paging messages may be considerably extended, up to almost 3 hours. Correspondingly, the UE must know on which HFN and on which time interval within this
HFN, the paging time window (PTW), it has to monitor the paging. The PTW is defined
by a start and stop SFN. Within a PTW, the determination of the PFs and POs is done
in the same way as for the non-extended DRX.
32
Data Transfer
User Plane CIoT EPS optimisation
5 Data Transfer
As described in Chapter 2.2.1, "Core Network", on page 5 , there are two ways for the
data transfer, the Control Plane CIoT EPS optimisation and the User Plane CIoT EPS
optimisation. The MME indicates support for each of those optimizations. For mobile
originated data, the UE may select among the supported options. For mobile terminated data the MME choses the optimization and may take the UE preference into
account, which is signaled to the network in the attach procedure.
5.1 Control Plane CIoT EPS Optimisation

For the Control Plane CIoT EPS optimisation, data exchange between the UE and the
eNB is done on RRC level. In the DL, data packets may be piggybacked in the
RRCConnectionSetup message or in the UL in the RRCConnectionSetupComplete
message. If this is not sufficient, data transfer may be continued using the two messages DLInformationTransfer and ULInformationTransfer:
Figure 5-1: NAS dedicated information transfer between the UE and the eNB
Contained in all these messages is a byte array containg NAS information, which in
this case corresponds to the NB-IoT data packets. Consequently it is transparent to the
eNB, and the UE's RRC forwards the content of the received DLInformationTransfer
directly to its upper layer. Between the eNB and the MME, the dedicatedInfoNAS is
exchanged via the S1-MME interface.
For this data transfer method, security on AS level is not applied. As there is also no
RRC connection reconfiguration, it may immediately start after or during the RRC connection setup or resume procedure, respectively. Of course, the RRC connection has
to be terminated afterwards with the RRC connection release.
5.2 User Plane CIoT EPS optimisation

In the User Plane CIoT EPS optimisation data is transferred over the conventional user
plane through the network, i.e. the eNB forwards the data to the S-GW or receives it
33
Data Transfer
from this node. In order to keep the UE complexity low, only one or two DRB may be
configured simultaneously.
Two cases have to be distinguished: When the previous RRC connection was released
with a possible resume operation indicated, see Chapter 4.5.2, "RRC Connection
Release", on page 31, the connection may be requested as a resume procedure as
shown in Figure 4-5. If this resume procedure is successful, security is established with
updated keys and the radio bearer are set up like in the previous connection. If there
was no previous release with a resume indication, or if the resume request was not
accepted by the eNB, security and radio bearer have to be established as shown in the
next section.
5.2.1 Establishment and Configuration of the Data Connection

After having set up the RRC connection as shown in Figure 4-4 or Figure 4-6, the first
step is to establish AS level security. This is done via the Initial security activation procedure:
Figure 5-2: Establishment of AS level security
In the SecurityModeCommand message, the eNB provides the UE with the ciphering
algorithm to be applied on the SRB1 and the DRB(s), and the integrity protection algorithm to protect the SRB1. All algorithms defined for LTE are also included in NB-IoT.
With this message, the SRB1bis automatically changes to the SRB1, which is used for
the following control messages.
After the security is activated, DRBs are set up using the RRC connection reconfiguration procedure:
Figure 5-3: RRC connection reconfiguration procedure
In the reconfiguration message, the eNB provides the UE with the radio bearer, including the configuration of the RLC and the logical channels. The latter includes a priority
used to balance the data transmission according to the actual requirements. PDCP is
only configured for DRBs, because the SRB only uses the default values.
34
Data Transfer
In the included MAC configuration, the configuration for buffer status report (BSR),
scheduling request (SR), time alignment and DRX are provided. Finally, the physical
configuration provides the necessary parameters for mapping the data to the slots and
frequencies.
5.2.2 Multi Carrier Configuration

The RRCConnectionReconfiguration may contain the settings for an additional carrier
in UL and DL, the non-anchor carrier.
When a non-anchor carrier is provided in DL, the UE shall receive all data on this frequency. This excludes the synchronization, broadcast information and paging, which
are only received on the anchor carrier. A bitmap may be provided indicating the
allowed DL SFs. The non-anchor carrier may contain considerable more SFs for data,
since it does not require synchronization and broadcast information.
Once the non-anchor carrier is configured, the UE solely listens to this one while it is in
the RRC_CONNECTED state. Consequently the UE requires only one receiver chain.
In UL the same principle applies. If an additional UL carrier is configured, the UE only
takes this one for data transmission, there is no simultaneous transmission in this carrier and the anchor carrier.
For both, DL and UL, the UE returns to its anchor carrier when it is released to the
RRC_IDLE state.
Figure 5-4 shows a scheduling example:
Figure 5-4: Scheduling example for 3 UEs. Every square denotes one subframe.
UE1 is configured with the anchor carrier, UE2 with other carrier in DL and UL, and
UE3 with a different carrier only on DL. For simplicity, this diagram neither considers
the NPDCCH period explained in the next section nor the SFs which are not allowed
for DL data. It shall only be interpreted in a schematic way.
Even in the in-band operation, the assigned DL carrier is not restricted to the values
shown in Table 3-1. This restriction is only for those carrier transmitting the NPSS,
NSSS and NPBCH, i.e. which may be used as anchor carrier. For a provided DL frequency, all carriers are allowed.
This structure allows to roll out an NB-IoT broadband network, although each UE only
has one transmitter / receiver chain with a narrow bandwidth available. The overhead
of synchronization, broadcast in DL, and the NPRACH resources in the UL may be
35
Data Transfer
restricted to one or a view pairs of carrier, while the other ones may be completely
used for data transmission. As the reception and transmission is never done simultaneously and always restricted to one band, respectively, it is sufficient that the UE only
has one transmitter / receiver chain with a bandwidth of 180 kHz.
5.2.3 Receiving the Control Channel

In the RRC connected state, the UE only monitors the UE specific search space (USS)
to obtain its UL grants and DL assignments. The reconfiguration message contains the
maximum number of repetitions, which ranges from 1 up to 2048 in powers of 2. However, the actual number of repetitions may be smaller, as indicated in the following
table:
Max nr. of repetitions Rmax
Actual nr. of repetitions
1, 2
1, 2, 4
>= 8
Rmax/8, Rmax/4, Rmax/2, Rmax
For the case that the actual number of repetitions is smaller than its maximum number,
the remaining SFs may be used to send a different NPDCCH to another UE. For example, if the maximum number of repetitions is 4, then all SFs may contain the DCI for
one UE, or two SFs may be used for each of two UEs, or each SF for a different UE,
respectively. Of course, the UE has to monitor all these candidates.
The reconfiguration message also contains a parameter to describe the NPDCCH
period. With respect to the start of this period, the USS starts either at the beginning or
at an offset of 1/8, 1/4 or 3/8 of the period length.
5.2.4 Transmitting the UL Data Channel

The DCI format N0 indicates an UL grant for transmission on the NPUSCH, including
all relevant parameters. One transport block can be repeated several times. The
arrangement of the repetitions depends on the number of subcarriers for one RU, the
subcarrier spacing and the number of repetitions. This shall be first illuminated with an
example shown in Figure 5-5:
36
Data Transfer
Figure 5-5: Example of an arrangement for NPUSCH transmission with repetitions. For the case of no
repetitions, the slot sequence shown in (b) would be transmitted.
For the case of a 15 kHz subcarrier spacing, a transport block, named test word (TW),
is transmitted on two RUs (a), where each RU has the format of 3 subcarrier over 8
slots (b). A total number of 8 repetitions is applied. In Figure 5-5 Tn denotes the n-th
slot of the first RU, Wm the m-th slot of the second RU.
In a first step, the two slots T1 and T2 are transmitted. This pair is repeated three more
times, so that there are 4 transmissions of these slots. Then the same procedure is
done with the next two slots. This is continued until the slots W7 and W8 are pairwise
transmitted four times. Finally, as there are now 4 repetitions of the TW, the transmission sequence is repeated once again, reaching this way the 8 repetitions.
In the general case, the first repetition of two slots is always done for a subcarrier
spacing of 15 kHz. On the 3.75 kHz subcarrier spacing it is done for every slot separately. The total number of first repetitions is half the number of total repetitions with an
upper limit of four, if the RU has more than one subcarrier, or one if the RU has only
one subcarrier. On the above example this would mean that if there would be 32 repetitions, the sequence generation would be like in Figure 5-5 (c), however the total
sequence would be repeated 7 additional times.
Usually, the sequence is mapped to a contiguous set of slots. An exception occurs on
a larger number of repetitions. To be more precisely, after a transmission of 256 ms, a
gap of 40 ms is created before the NPUSCH transmission is continued. This gap is
necessary, because when the UE transmits on the NPUSCH, it cannot simultaneously
receive the DL channel and may so lose the synchronization to the eNB. During this
gap, synchronization is fine-tuned again.
5.2.5 Receiving the DL Data Channel

The DCI format N1 indicates a DL assignment describing where and how the data
symbols are transmitted on the NPDSCH. The principle is essentially the same as for
the UL, see the example shown in Figure 5-5, however the data packets are not grou-
37
Data Transfer
ped first into RUs. If there are no repetitions, the data packets are consecutively mapped into slots and sent in NPDSCH SFs.
For the case that there are repetitions, the mapping is similar like in multi-tone UL.
Data for two slots are first mapped into one SF, which is now repeated with the total
number of repetitions, again with an upper limit of four. Then the mapping is continued
the same way until all SFs are transmitted. Finally this whole structure is repeated until
the desired number of repetitions is reached.
There is an exception, when the DL data contains SIB-NB information. In this case, the
slots are transmitted consecutively like in the example in Figure 5-5 (b), and then this
whole block is repeated the appropriate number of times.
Transmission gaps may be configured by the RRC for NPDSCH transmissions with a
large number of repetitions. The RRCConnectionReconfiguration message may provide the information how big the number of repetitions shall be in order to trigger such
gaps, the periodicity and the length of their occurences. When the gap occurs, the
NPDSCH transmission is postponed to the next available SF after the gap. These gaps
do not apply for NPDSCHs carrying MIB-NB or SIB-NB information.
38
Summary and Outlook
6 Summary and Outlook

With the NB-IoT technology specified in Release 13, 3GPP has created a new cellular
air interface which is fully adapted to the requirements of typical machine type communications. It is optimized to small and infrequent data packets and abstains from cellular features not required for that purpose. This way, the UE can be kept in a cost efficient way and needs only a small amount of battery power.
Data transmission is kept to a small frequency band of 180 kHz. However, due to the
multi-band construction a broad spectrum of frequencies may be used. The signaling
part may be reduced to one or only a few NB-IoT carrier, whereas the remaining ones
may be fully utilized for data transfer. This way, a considerable amount of bandwidth is
used for data transfer, although the single UEs have only a comparatively narrowband
transmitter and receiver.
With Release 14, the development of NB-IoT will continue [14]. According to the current plans, NB-IoT will be extended to include positioning methods, multicast services
required e.g. for software update or for messages concerning a whole group, mobility
and service continuity, as well as further technical details to enhance the field of applications for the NB-IoT technology. We will provide the associated technology description, so please check our website for a new version of this document.
39
References
7 References
[1] 3GPP TR 45.820 V13.1.0, November 2015; Cellular system support for ultra-low
complexity and low throughput Internet of Things (CIoT)
[2] 3GPP TS 23.272 V13.3.0, March 2016; Circuit Switched (CS) fallback in Evolved
Packet System (EPS)
[3] 3GPP TS 23.401 V13.6.1, March 2016; General Packet Radio Service (GPRS)
enhancements for Evolved Universal Terrestrial Radio Access Network
[4] 3GPP TS 23.682 V13.5.0, March 2016; Architecture enhancements to facilitate
communications with packet data networks and applications
[5] 3GPP TS 36.331 V13.2.0, June 2016; Radio Resource Control (RRC); Protocol
specification
[6] 3GPP TS 36.300 V13.4.0, June 2016; Overall description
[7] 3GPP TS 36.101 V13.4.0, June 2016; User Equipment (UE) radio transmission and
reception
[8] 3GPP TS 36.201 V13.2.0, June 2016; LTE physical layer; General description
[9] 3GPP TS 36.211 V13.2.0, June 2016; Physical channels and modulation
[10] 3GPP TS 36.212 V13.2.0, June 2016; Multiplexing and channel coding
[11] 3GPP TS 36.213 V13.2.0, June 2016; Physical layer procedures
[12] 3GPP TS 36.304 V13.2.0, June 2016; User Equipment (UE) procedures in idle
mode
[13] 3GPP TS 36.321 V13.2.0, June 2016; Medium Access Control (MAC) protocol
specification
[14] 3GPP RP-161324 3GPP Work Item Description Enhancements of NB-IoT
40
Additional Information
8 Additional Information
Please send your comments and suggestions regarding this white paper to
[email protected]
41
Rohde & Schwarz
9 Rohde & Schwarz

Rohde & Schwarz is an independent group of companies specializing in electronics. It
is a leading supplier of solutions in the fields of test and measurement, broadcasting,
radiomonitoring and radiolocation, as well as secure communications. Established
more than 80 years ago, Rohde & Schwarz has a global presence and a dedicated
service network in over 70 countries. Company headquarters are in Munich, Germany.
Sustainable product design
Environmental compatibility and eco-footprint
Energy efficiency and low emissions
Longevity and optimized total cost of ownership
Certied Quality Management
ISO 9001
Certied Environmental Management
ISO 14001
Regional contact
Europe, Africa, Middle East

Phone +49 89 4129 12345
[email protected]
North America
Phone 1-888-TEST-RSA (1-888-837-8772)
[email protected]
Latin America
Phone +1-410-910-7988
[email protected]
Asia/Pacific
Phone +65 65 13 04 88
[email protected]
China
Phone +86-800-810-8228 / +86-400-650-5896
[email protected]
Headquarters
Rohde & Schwarz GmbH & Co. KG
Mhldorfstrae 15 | D - 81671 Mnchen
+ 49 89 4129 - 0 | Fax + 49 89 4129 13777
www.rohde-schwarz.com
This application note and the supplied programs may only be used subject to the conditions of use set forth
in the download area of the Rohde & Schwarz website.
R&S is a registered trademark of Rohde & Schwarz GmbH & Co. KG. Trade names are trademarks of the
owners.
42

Whitepaper: Narrowband Internet of Things

Uploaded by

Copyright:

Available Formats

Whitepaper: Narrowband Internet of Things

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Whitepaper: Narrowband Internet of Things

Uploaded by

Copyright:

Available Formats

Narrowband Internet of Things

Whitepaper NarrowBand_IoT 1MA266_0e

Whitepaper NarrowBand_IoT 1MA266_0e

Whitepaper NarrowBand_IoT 1MA266_0e

Minimize the signaling overhead, especially over the radio interface

Appropriate security to the complete system, including the core network

Improve battery life

Support delivery of IP and non-IP data [2, 3]

Support of SMS as a deployment option [4]

2.2.1 Core Network

Whitepaper NarrowBand_IoT 1MA266_0e

2.2.2 Access Network

Whitepaper NarrowBand_IoT 1MA266_0e

Figure 2-2: Network architecture towards the air--interface

2.3 Frequency Bands

Uplink frequency range / MHz

Downlink frequency range / MHz

Whitepaper NarrowBand_IoT 1MA266_0e

Whitepaper NarrowBand_IoT 1MA266_0e

In-band operation utilizing resource blocks within an LTE carrier

These modes are visualized in the following figure:

Figure 3-1: Operation modes for NB-IoT

LTE PRB indices

2, 7, 12, 17, 22,

4, 9, 14, 19, 24, 29, 34,

Whitepaper NarrowBand_IoT 1MA266_0e

NPBCH, the narrowband physical broadcast channel

NPDCCH, the narrowband physical downlink control channel

NPDSCH, the narrowband physical downlink shared channel

and two physical signals

NRS, Narrowband Reference Signal

NPSS and NSSS, Primary and Secondary Synchronization Signals

Figure 3-2: Mapping of the transport channels to the physical channels

Whitepaper NarrowBand_IoT 1MA266_0e

3.2.1 Frame and Slot Structure

Whitepaper NarrowBand_IoT 1MA266_0e

3.2.2 Narrowband Reference Signal

Whitepaper NarrowBand_IoT 1MA266_0e

3.2.3 Synchronization Signals

Whitepaper NarrowBand_IoT 1MA266_0e

3.2.4 Narrowband Physical Broadcast Channel

2 bits indicating the two LSBs of the hyper frame number

4 bits for the SIB1-NB scheduling and size

5 bits indicating the system information value tag

1 bit indicating whether access class barring is applied

11 spare bits for future extensions

Figure 3-7 shows its mapping to physical resources:

Whitepaper NarrowBand_IoT 1MA266_0e

Figure 3-7: NPBCH mapping to the subframes

Whitepaper NarrowBand_IoT 1MA266_0e

3.2.5 Dedicated Channels

Control Channel NPDCCH

NPDCCH format 1 taking both NCCEs.

Whitepaper NarrowBand_IoT 1MA266_0e

Type-1 common search space, used for paging

Type-2 common search space, used for random access

UE specific search space

Paging and direct indication

Traffic Channel NPDSCH