Professional Documents
Culture Documents
5G New Radio - Explained in A Nutshell - Understand The Latest 5G Radio Specs in A Story Telling Way (5G System Architecture Book 1) - Ali A. Esswie
5G New Radio - Explained in A Nutshell - Understand The Latest 5G Radio Specs in A Story Telling Way (5G System Architecture Book 1) - Ali A. Esswie
5G New Radio - Explained in A Nutshell - Understand The Latest 5G Radio Specs in A Story Telling Way (5G System Architecture Book 1) - Ali A. Esswie
Authored by
Ali A. Esswie
1
Table of Contents
Preface ............................................................................................................................................................................... 3
Introduction ...................................................................................................................................................................... 5
5G Spectrum ............................................................................................................................................................... 33
2
Preface
This book introduces the major conceptional principals of the state-of-the-art fifth generation
(5G) radio system design, including its main driving technology pillars, radio interface design, and
the disruptive quality of service architectures. In a simple and very easy-to-digest way, this book
brings down the latest 5G radio standard specifications; however, with an effortless story telling
style, that would be easily accessible by undergrads, grads and early researchers. Accordingly,
the 3GPP specs of the 5G new radio are ultimately simplified, with in-text illustrative examples
3
Author Biography
Ali A. Esswie is an expert wireless research engineer, equipped with distinctive experience of industrial and
academic technical projects. He is currently with Nokia Bell Labs as a wireless research standardization
engineer. During 2013-2016, he was with Intel Labs and Huawei performance group, respectively, engaging
obtained his BSc. degree from Cairo University with academic honors. Later, he got his MSc. degree from
Memorial University of Canada when he won the Fellow of the graduate school for research excellence.
Currently, Ali is pursuing his PhD degree at the electronic systems department at Aalborg University. Ali
owns many high quality academic publications, published in world flagship academic conferences and
journals, in addition to invention intellectual reports which span the fundamental 4G/5G system design
principals.
4
Reading Keys
‘e.g.,’ denotes: for example, or for instance. It simply provides an example of the respective text.
‘i.e.,’ denotes: that is, or in other sense, in other way. It simply introduces another way of expressing the
former text.
‘−‘ implies: a side note or further explanation or review of prerequisite knowledge of former text.
‘BS’ or ‘base-station’ or ‘eNodeB’ or ‘gNB’: all denote the base-station end of the 4G and 5G radios.
5
Introduction
Since 1970s, a new cellular mobile generation tends to appear every 10 years period. Since then, our global
societies have been dramatically influenced by the advances in the wireless communications, pushing
towards a more modernized way of life. The first mobile generation (1G) was dictated by analogue
communications. Later, second generation (2G) has introduced the digital audio communications as well
as text messaging during 1989. The third generation (3G) accordingly has been developed mainly to meet
the wave of the data communication demand at this time and was finally frozen in standards by the end
of 2007. More specifically, the 3G technology has witnessed five different major updates, i.e., standard
releases, each was majorly concerned by several system design improvements to support higher data
rates. At the end of the 3G era, it was encompassed by what is globally re-known as the universal mobile
telecommunication system (UMTS) – simply put, the truly mature version of the 3G cellular
technology. Well then, first let me give you a quick glimpse on how the state-of-the-art 5G
technology generically adopts a different set of conceptional design principals, that are quite
Mainly, the 5G technology comes with a so-called new radio. The 5G new radio implies that the
radio interface becomes much more agile and simply programmable in time, equipped by further
sophisticated radio communications over the air as well as numerous technology variations that
were not available with the former 4G and 3G systems. In Table I, the main differences between
the major cellular standards are precisely presented – let’s lightly go through them one by one.
6
Table 1: Major differences between 3G, 4G and 5G cellular systems.
Item 3G 4G 5G
Downlink waveform CDMA OFDM OFDM
Uplink waveform CDMA SC-OFDM SC-OFDM
Channel coding Turbo Turbo LDPC (data) – Polar (control)
Transmission time interval 10/2 ms 1 ms Flexible {0.143, 0.5, 1, …} ms
Beamforming NA Data Data + Control
Bandwidth 5 MHz 0.4-20 MHz Up to 100/400 MHz
Quality of service Bearer based Bearer based Flow/packet based
Network slicing NA NA Supported
Fast transmissions NA NA Connectionless
Cloud services NA NA In-Built
Twenty years ago, the 3G networks adopted the code division multiple access (CDMA) to multiplex
different users over time, where each user is assigned a specific sequence code, which it should
use to decode its data. Although, CDMA has been proven to provide a limited capacity and
users. Then, the 4G technology later upgraded the multiplexing technology to the orthogonal
frequency division multiple access (OFDMA) where users are multiplexed over sereval time and
– imagine the whole bandwidth is divided into many smaller pipelines, each user may be assigned
bandwidth pipelines.
7
However, with the 5G new radio, it was decided – up to the current standardization activity, that
OFDM is still the best fit for the sake of multiplexing different and concurrent user transmissions.
That was in fact quite disappointing to many international researchers where the research line of
However, one of the major changes that is being introduced with the 5G new radio is the adoption
of a new channel coding – well, not really new within the academic community. The Turbo
encoding was extensively utilized within the 3G and 4G systems. Generally, the Turbo coding is
– a forward error correction technique is a way to correct unintended errors in the received signal,
mainly by adding some redundant data to the actual transmitted information, e.g., a transmitter
desires to transmit a ‘1’, Turbo encoder shall send it as ‘110’ with two additional bits just for error
Furthermore, with the 4G radio technology, the Turbo process can be dynamically reconfigured
at the receiver side to adjust the receiver performance and decoding complexity with the
predicted size of the communication error, basically coming from the varying wireless channel in
between, i.e., increase decoding complexity and time only when needed to maintain a desired
reception performance. Well, by now, I feel that I may have convinced you why 5G should use
Turbo decoding as well? In fact, it will NOT. The 5G new radio is being decided to adopt the low-
8
density parity check code (LDPC) for data transmissions and Polar codes for control plane, i.e.,
– LDPC is foreseen to address some problems of the Turbo coding and which the 5G new radio
significantly relies on, especially towards the 5G broadband use cases and reliability targets (– no
worries, continue reading and all shall be addressed later). So, LDPC provides a higher decoding
throughput with much less decoding complexity compared to the Turbo decoding. Also, LDPC
– well, the rate matching of a code implies its ability to match the number of bits of one code
block, i.e., one coded segment of data after adding redundant bits, to the maximum allowed
Accordingly, LDPC is highly beneficial when used with broadband services over the 5G new radio.
On the other side, for control channels, Polar codes are quite mature in literature and mostly are
known to be the only coding technique that approaches the theoretical optimal capacity limit. The
intuition behind using Polar codes in the control plane of the 5G new radio consolidates the
– imagine a broadband system with significantly degraded control communications, this means
that both users and network are able to achieve very high data rates, but they are not well
9
Stressing on such fact, the antenna beamforming capabilities coming with the 5G new radio are
also extended to include the control channels instead of only the data channels with the 4G
systems.
Furthermore, the 5G new radio is envisioned to have an ultimate timing flexibility. You may have
already heard that 5G is to support flexible numerologies in time and frequency. This means that
the 5G technology can be fit to different quality of service requirements such as the latency-
critical and broadband communications, respectively, over a single radio spectrum. Well, let’s
consider the time domain for now. Both 3G and 4G systems support quite a large and rigid
– well first, a transmission time interval is the time periodicity where the system performs either
seconds. With 4G, a transmission time interval is set to 1 milli-second. In that sense, when a
latency-critical traffic, that needs to be delivered as soon as possible, arrives at the transmitter
during a random point of time, such traffic will be buffered until the next transmission time
opportunity which may be due in 1 or 2 milli-seconds later. That is quite too much wasted time
towards the 5G latency requirements, and here comes the power of the 5G new radio. 5G utilizes
variable transmission time durations that can be as down to a duration of a single OFDM symbol,
i.e., 0.0715 milli-seconds, 2-OFDM symbols, i.e., 0.143 milli-seconds, and conventional 14-OFDM
10
symbols, i.e., 1 milli-second. This way, latency-critical users can be served by smaller transmission
time intervals while data-rate-hungry applications are only supported with the longer durations.
Moreover, for such latency sensitive applications, 5G new radio supports connectionless
communications for short pack transmissions, i.e., users can immediately transmit their
respective data payload without waiting for control signaling such as the scheduling grant. These
schemes are under the umbrella of the grant-free transmission technology, especially in the
uplink direction.
Additionally, network slicing and flow-based quality of service are two major breakthroughs of
the 5G radio technology. Generally, network slicing is a very similar approach to the software
defined networking and network functions virtualization techniques in fixed networks. With
network slicing, the physical network structure can be virtually partitioned, i.e., into different
virtual sub-networks or slices, based on the specific quality of service needs. It is mainly
– one simple example to demonstrate the 5G network slicing would be if we have two applications
running on a single cell phone, one is latency-critical such as vehicle to vehicle tracking app and a
broadband live streaming of a football game. The former requires a strict radio latency but not
necessarily a high throughput which is completely the opposite to the latter application. Thus, the
network can be virtually optimized such that the latency-critical traffic gets through the fastest
network paths, i.e., but not required to be of high bandwidth, while the other one can be a little
11
In conclusion, the 5G new radio comes with a revolutionary technology spanning across all its
system components. However, the key improvement of the 5G technology lies in its generic
flexibility, i.e., multiplexing of different quality of service demands on same spectrum, adaptation
of the time and frequency transmission numerology based on the instantaneous need, and the
However, I bet that you may have a fundamental question about why could not we just evolve
the 4G radio technology?, magically to get the 5G radio standards. We have done it before from
Well, each mobile generation is usually driven by its killer applications and use cases, which implies
why this world does need a new mobile technology, – simply, put. The 1G was motivated by voice
communications. Later, 2G was driven by improved voice and text communications. 3G then
comes with a further better voice and affordable data connectivity, that occurred when people
and technology were both ready for data connectivity. Though, 3G did not satisfy such hunger of
the internet connectivity. Thus, 4G followed the trend to offer ultimately higher capacity mobile
networks. That is, the main aim was to exponentially increase the network and user data rates. In
other sense, it is the data connectivity as the main and single 4G target to pursue. This led the 4G
standardization to combine a newly improved core network structure, – that is so called as the
evolved packet core (EPC) as well as evolved radio technology. As the name suggests, it was an
evolution to the 3G technology, i.e., improving transmission techniques, core hardware and
12
interfaces, and more bandwidth allocation; however, without changing the baseline design of the
However, before answering why 5G is different? – let’s first discuss what are the driving
5G introduces the term: over the top (OTT) vertical industries. Well, OTTs denote the industrial
partners, i.e., companies, segments, etc, which are recently becoming concerned by the 5G
technology standardization; although, they currently run businesses independent from the
cellular communications.
– think of food factories, health care facilities, automotive manufactures, gaming and virtual
These partners are highly involved in the 5G standardization process with a single target to
transform the physical wired connectivity into a purely wireless one, with ultimate flexibility,
reliability, and broadband rates. This imposes a lot of challenges against the 5G radio technology,
since currently, we do not simply have a single objective anymore to optimize the network
towards.
– imagine a surgical doctor is performing a remote operation on a patient through a haptic robotic
arm, connected through the 4G radio interface to the control unit at the doctor’s office. With such
a critical use case, the 4G radio interface is required to have a minimal radio latency, and an
13
ultimately high link reliability, i.e., data communication link is guaranteed without intermediate
simultaneously achieved on a single radio interface. If you recall, an optimal system capacity is
the Shannon limit; however, this comes with the assumption of an infinite communication latency.
Surprisingly, this set of mentioned requirements is exactly what the 5G standards promise. They
URLLC applications require extremely fast and stochastic transmissions of small data payloads
– think of a robotic machine that is randomly in time transmitting small update packets towards
a controller to indicate its positions, seeking instructions on the next movements. If either these
packets are being dropped due to the wireless channel in between or arrived quite late, an outage
occurs at this machine. Herein, outage simply means that such machine is out of service, i.e., not
aware about the next operation, due to lack of communication reliability to the controller. This is
mainly what URLLC service class is concerned about. That is, how does the 5G radio interface offer
14
Enhanced mobile broadband (eMBB)
eMBB service class is quite a conventional class to what we formally had with the 4G systems,
except that 5G targets require more and more of the eMBB data rates. eMBB denotes the
applications that demand extremely high, i.e., broadband, communication speeds or data rates
such as video streaming services, online gaming, internet teleconferencing, and many more.
Basically, the dominant factor towards the 5G eMBB targets is the channel bandwidth, – more
Such 5G use case is a crucial and clear differentiation from the 4G technology. eMTC characterizes
a massive number of cheaper, low-power and low-energy connected devices to the radio
interface, all transmitting data payloads with random arrivals and for a very short period of time.
– imagine we have a multi sensor network setup, where thousands of sensors, equipped with
radio frequency modules, are exchanging, through 4G links, some updates about air temperature,
pressure, humidity, etc, with a controller center at the other side of the globe. These sensors are
of extremely low power capability and need to transmit their payload as quickly as possible to
return idle (sleepy) again. Herein, the radio interface needs to introduce special access techniques,
resource allocations and scheduling algorithms in order to cope with such newly introduced
targets. Although the very recent versions of the 4G networks have introduced some system
enhancements for mMTC such as narrow band resource allocations for mMTC and internet of
15
things (IoTs) devices, some other improvements are not even feasible due to conflicts with
– as an example, when a device attempts accessing the radio interface of the 4G network, it must
perform the well-known random-access procedure. First, it waits until the first available random-
access opportunity, announced by the network. Note, during this period, the device is running in
full battery. Then, it passes through the four-step random-access operation of the 4G radio
technology, which eventually may or may not succeed. Upon failure, the device repeats the access
attempt after some random back-off time. This implies, devices must first connect to the radio
interface before being able to transfer their respective data. With mMTC, the data payload to be
transmitted may be extremely of small size, i.e., several bytes, while the 4G signaling overhead
for random-access may take over hundreds of bytes. That is then a significant loss of the network
resources and device energy consumption, respectively. However, that is how the 4G technology
works up to time. With the 5G new radio, it offers connectionless packet transmissions where
devices just wake up, transmit their payload and immediately fall back a sleep again, – that simple,
Thus, if you still recall the excited question in this part, i.e., why simply cannot we evolve the 4G
networks to reach the 5G targets?, the long-awaited answer is: NOT AT ALL, we cannot evolve
the 4G technology to reach the requirements of the 5G new radio. The main driving applications
of the 5G are completely different from that of the former 4G. The set of 5G requirements are
16
simply challenging the baseline system design of the 4G systems, for which it becomes not
Well, let’s organize things more here. Next, the formal standard 5G requirements are presented
There are several global and continental alliances which are considered pioneers to define the 5G
targets. Some of these are operator-led, others are hybrid between operators, network vendors
Next generation mobile networks (NGMN): it is an operator-led specification alliance that really
targets the needs of the participating telecom operators by setting the appropriate requirements
seeks to ensuring that we communicate over interference-free spectrum. In other sense, its major
task is to efficiently propose the best practices of using the available spectrum of each mobile
some help from the former 4G architecture so we may call it as pre-5G lunch.
The 3rd Generation Partnership Project (3GPP): that is the widely adopted standardizing partner
of each mobile generation. More details about 3GPP shall follow in the next Section. 3GPP unites
17
the seven-major telecommunication standard development entities to develop a stable 5G operation
As you can expect by now, these organizations may have set different requirements for 5G. Well,
that is right, but not with extreme variations though. Herein, we only go through the official 3GPP
Well, most of our populated 4G mobile networks usually offer radio latencies between 200 to 500
ms. The 5G new radio apparently sets a tougher radio latency requirement. But first, let me explain
– the user plane latency means the difference in time between the point a data packet arrived
at the buffers of either the base-station or the user (which needs to be transmitted) to the
point when such packet is successfully decoded by its intended receiver. Such latency
includes the scheduling delay, i.e., some data packets may not get transmitted instantly,
processing and hardware delays at transmitter and receiver, propagation delay, and most
importantly the re-transmissions delays, i.e., due to the time-variable wireless fading
channels, some packets may not go through from first time, thus, transmitters tend to
transmit these packets once again (can be of several times) and with same or different
transmission configuration as the first transmitted packet, until the packet gets either
successfully decoded or dropped if all attempts have failed. Although, the control-plane
latency implies the difference in time from the point a device is in idle mode (sleepy) until it
gets active. If you recall, in 4G systems, this delay shall contain multiple message
exchanges between base-station and the user for random access, security check, quality of
18
Thus, these latency requirements cannot be fundamentally satisfied by the 4G technology
due to its rigid frame structure as we will follow up later in more detail.
- 5G peak data rate: 20 Gbps for downlink and 10 Gbps for uplink
Several specification groups called for 1000x times peak data of 5G than 4G. Recalling that the latest
enhanced version of the 4G systems can maximally reach a peak data rate of 1-3 Gbps. Then, a 5G
system should be capable to reach 1000 – 3000 Gbps of peak data rate, which is a very hypothetical
case. Thus, 3GPP has adopted a more realistic view to define the target downlink and uplink
maximum data rate. But, first, how is the standard peak data rate being estimated?
– the peak data rate indicates the theoretical, error-free, maximum-possible user data rate that a
device should achieve when it has been granted all the system radio resources, e.g., a single active
user case.
Such extensive data rates can only be achieved by distributive capacity-boosting technologies such
exponentially increased; however, this comes with a highly reduced radio coverage and
The mobility interruption time implies the shortest time supported by the system during which a
device cannot exchange any data packets, for instance, during a handover from a serving BS to a
19
new one. That means that 5G new radio specifies that devices should be always connected to the
In 4G, the mobility interruption time is usually between 10 – 15 ms since a device fully disconnects
first from the serving BS and attempts a ‘reliable and pre-configured’ connection to the secondary
BS.
– the intuition behind such 4G handover procedure was to minimize the control overhead of the user
Aligned with the former 3G specifications, 5G new radio re-introduces the make-before-break
concept, as the device establishes its secondary connection before it terminates its primary one,
given that user devices have become of more signal processing and power capabilities.
– as you observe, an extensive care has been enforced in every way possible to achieve a very low
radio latency during the different user states, i.e., idle mode, active mode, and handover transitions.
This denotes the 5G new radio needs to be a reliable mean of wireless transmission regardless of
the system and/or device state. This defines the ultimate target of the 5G new radio as the reliability.
The most challenging requirement of the 5G new radio is to reach a reliability target of 99.999%
within 1 ms. However, for completeness, lets first define what really reliability is?
decoded at its intended receiver, conditioned that it should be decoded during a given latency bound.
Thus, the 5G new radio requires that during 99.999% of time, the latency-critical, i.e., URLLC,
packets get received and successfully decoded within just 1 ms. This one 5G system requirement
challenges the whole engineering baseline of the prior 4G radio technology. For more completeness,
some researchers also define the radio reliability by the outage probability as: 1 – 0.99999 = 0.00001,
where in a simple sense, it denotes how many packets can be easily tolerated by the system. Thus,
20
5G requires that a maximum of a single packet every 100000 transmitted packets, either gets
dropped or even successfully decoded but NOT within the 1 ms deadline, – that is really challenging.
But, do we really need such tight wireless reliability? answer is YES, since we are evolving to build
our daily life activities on cellular communication as being explained in detail over the previous sub-
sections. Although, the tough question now is how can actually we achieve it over wireless
connectivity? – well, we will find out over the next Sections. Below, several other self-explanatory
5G new radio requirements are presented, where they are all concerned about having more of
So, to wrap up, I would like to consolidate on why the 4G technology is not able to reach the 5G
targets. As we have gone through the 5G requirements, there are some targets which are just more
of what was there with the 4G; however, there are other applications and requirements that either
not being specified or cannot be supported by the 4G radio technology, in a sense that a newly
Over the upcoming sections, we will go through understanding the 5G radio protocols,
transmission strategies, system architectures and many more, with a moderate level of details,
combined with concrete conclusions. Although, lets’ start by explaining which entity does actually
21
3GPP Standardization of the 5G New Radio
3GPP has been created since 1998 as a joint standardization activity between USA, Europe, Japan,
and Korea. As of now, the 3GPP has become the leading standardization body for cellular
project coordination group (PSG), where its members manage three major technical specification
groups (TSGs), – think of a TSG as a specialized committee that is concerned only about
standardizing a certain system partition. Accordingly, each TSG is divided to several work groups
(WGs) which look closer in a more specialized and narrow way into standardizing specific system
aspects of the belonging TSG. Furthermore, as you may have heard the term work items (WIs)
before, they define the technical work and studies that will be performed within each WG. In
general sense, 3GPP TSGs usually hold 4 plenary meetings per year, typically in March, June,
September and December while there are 1-2 meetings of the corresponding WGs between every
TSG two successive meetings. So, WGs propose and study the technical aspects of the concerned
system design principals within their scope and provide recommendations to the parent TSGs for
final decisions.
22
Source: 3GPP
With this, let’s go back again to the main question here: does 3GPP alone standardize our cellular
networks? Answer is NO, 3GPP only produces technical reports (TRs) and technical specifications
– a TR is just a feasibility study that a given 3GPP WG has performed. TR can be seen as an
academic paper that evaluates the performance of a given system design aspect. Although, a TS
denotes a certain set of the system specifications which should be adopted to reach the
Well then, 3GPP does have several regional standardization partners as ARIB (Japan), ATIS (North
America), CCSA (China), ETSI (Europe), TSDSI (India), TTA (Korea), and TTC (Japan). These partners
shall transpose and slightly modify the 3GPP TSs to be well aligned with the regional demands.
–for that very specific reason, you may have experienced that a mobile phone, manufactured in
Europe may have inflicted some feature degradation while being connected to a Korean network
23
for example, simply because both standards are slightly variant. However, to achieve a global and
transparent connectivity, that should not affect the major mobile device operation.
Finally, every 1-3 years of the hard work of all 3GPP WGs, 3GPP issues a reliable release of the
– a 3GPP release indicates a full system specification set, which is approved by 3GPP, concerning
all aspects of the current mobile generation such as the radio design, core design, interfaces, IP
connectivity, etc.
For completeness, the 3GPP specification documents of the 4G usually take over the numbering
starting by 36.xxx, while the 5G TS documents shall start by 38.xxx. Examples can be as below.
By now, the first concrete 3GPP release for 5G has been out during September 2018, i.e., release-
15. Release-15 has identified the main building blocks of the 5G system architecture (which we
basically will go through over the next Sections), and that will operate in a standalone mode.
Although, the early 5G deployments, envisioned during 2019 and early 2020, shall rely on the
existing 4G structures.
24
Main Building Blocks of the 5G New Radio Architecture
As of now, we have answered why simply we do need new radio and core structures for the 5G
as scalable as possible, as efficient as possible, and as reliable as possible. Flexible denotes that
the 5G radio and core design should be both adaptable and programmable in time to meet the
diverse, and sometimes conflicting, requirements of the supported use cases. Some examples,
as being discussed in previous Sections, include the support for broad band transmissions with
large payload sizes and extreme data rates, while supporting a massive number of connectionless
devices with sporadic and very low payload sizes. Scalability is characterized by the system ability
to host a very large number of connected devices at an arbitrary time. This may seem easily
achieved by just expanding the operating spectrum, i.e., communication bandwidth; however,
having different types of traffic patterns, transmitted packet sizes, connection-based and
connectionless communications, and many more use cases set a fundamental limitation of the
system capacity.
– as an example of this limitation, having very time sporadic traffic patterns over the radio
interface does in fact disturb the link adaptation (LA) process. First, LA means that the radio
interface dynamically in time adjusts the transmission configurations such as the modulation and
coding schemes based on each link quality and radio conditions. Although, with such random and
small-payload traffic type where transmissions are very short and sporadic in time, link qualities
can be changing quite aggressively from a time unit just to the next one. Thus, LA should be very
25
fast and more accurate accordingly. Otherwise, LA process shall be dictated by outdated link
information that would not be valid at the time such LA enforcement is applied.
On another side, efficiency of the 5G new radio implies exactly the same meaning as being defined
with the 4G technology. Having efficient radio and core interfaces is still essential to achieve
– spectral efficiency is a measure of how much the system radio resources, i.e., spectrum, is well
utilized. High spectral efficiency denotes that the radio interface is capable to achieve high
Finally, the reliability is a key design principal for 5G communications. It basically changes the way
we assess the performance of the 5G system then. For instance, with 4G technology, we majorly
cared more about the system average coverage and block error rates, while with the 5G systems,
we mostly care about a specific percentile of users which could or could not achieve a certain
reliability bound, regardless of the coverage and channel conditions of each user. This adds the
requirement of an ultimate radio flexibility in order to achieve the same reliability level for all
Over the upcoming sub-sections, we will go through in more technical details of the main blocks
26
5G Radio Access Network: Radio Protocols
In a generic sense, the 5G new radio follows a relatively similar radio access network (RAN)
structure as the 4G one. The gNB, defined as the general NodeB, is the 4G-alike enhanced NodeB
(eNB) or base-station that offers the 5G new radio connectivity on both the user and control
planes. As being referred in previous Sections, dual connectivity with 4G is essential during the
early 5G deployments; why? because it would be too cost inefficient to start a full 5G system at
once, especially with the lack of its main use cases at the mean time. Thus, ng-eNB and eLTE-eNB
terms are defined as the next generation (NG) base-stations which offer the connectivity to the
The gNBs are then inter-connected through the logical Xn interfaces, which correspond to the
– a logical Xn interface denotes that connectivity may not be through a direct connection but
As will be explained in greater detail in the Second note series of this book, the gNBs are
connected to the 5G core through the NG core interfaces. The first core device to connect to is
the access and mobility management function (AMS) through the NG-control (NG-C) interface and
the user plane function (UPF) through the NG-user (NG-U) interface.
– for now, do not worry about why these 5G core hardware devices are entitled as functions
27
Thus, on the radio side, a 5G gNB is responsible to provide the end radio connections, admission
control, handovers, lower level network slicing, connectivity of multi-radio access technologies
On the core side, the AMS, as the name suggests, is the first core entity a 5G user shall connect
to in order to have its connection authorized, registered within the network, and ciphered.
However, the UPF is the core entity responsible for the user packet routing and forwarding
between the core ingress and egress points, and the handling of each user quality of service
requirements.
– think of it, as the AMS manages the control plane of the users and the UPF administrates their
respective user data plane transmissions. The 5G core structure will be explained in depth through
Additionally, the 5G technology further consolidates the importance of the centralized approach,
where the RAN architecture is split into a centralized base-band units (CUs) and many radio
– CU represents the administrative unit that processes the data received from or transmitted by
its associated DUs, while DUs can either have only the responsibility of the antenna transmissions
or a minor set of signal processing as well. Each CU can be associated to several DUs accordingly.
28
Such basic cloud RAN approach is well-known since the 3G and 4G systems; however, it did not
get a significant success due to the shortage of the application demands at this time, that made
So, why such centralized approach is quite important with the 5G new radio? Basically, because it
enables vital performance-improving techniques towards the main use cases associated with the
5G system. Having a centralized unit which has access to the information’s from multiple DUs,
and potentially hundreds of users accordingly at the same time, shall make it much easier for
interference coordination, exchanging channel state information between DUs, and supporting
much more efficient coordinated user scheduling operations. These techniques could not be
supported with the 4G radio interface because it would be too costly to share such control
information between eNBs. – here, too costly implies it requires a significant amount of control
signaling overhead.
Thus, with 5G new radio, a diversity of functional splitting options has been introduced between
– a functional split means that a specific set of the radio processing jobs can be moved and
processed at different spots of the network, either over the DUs or CUs. For example, with
original LTE standards, the baseline communication structure is the fully distributed one, i.e., no
CUs. It does mean that everything from having received an analogue signal on the antenna array,
channel estimation, packet retransmissions, resource and user scheduling and more, occur at the
29
LTE eNB processing units. Although, with the 5G functional split approach, a single gNB can be
divided into a single CU and multiple DUs, and all are interconnected through high data rate
interfaces, e.g., Fiber connections. Then, the latency-critical traffic can be fully processed at the
DUs for minimum end-end latency; however, latency-relaxed traffic can be processed at the CU
for better coordination performance among all involved DUs. This defines two main options for
the functional split as: the lower and higher layer split. The former indicates that the Physical layer
of the 5G new radio is located at the DUs, while the other layers from the protocol stack are
processed at the CU. However, the latter option implies that both layer-1 (Physical) and a sub-set
of Layer-2 (medium access control (MAC)) functionalities are processed at the DUs. The
advantage of option 2 is that it requires much lower transport overhead between the CU and its
– The physical layer (Layer-1) of the protocol stack usually exhibits the significant portion of the
control overhead.
So far, we have been discussing the functional layer split of the 5G new radio. Accordingly, lets
define what is a layer? and how the data and control information payloads are carried out through
You have probably heard of the terms: access and non-access stratum (AS and NAS) from LTE
standards. The NAS implies the control information flows between a 5G user, 5G new radio, and
the 5G core interfaces, as well as how such information is being understood through the different
30
parts of the network. However, the AS denotes the same functionality as the NAS but only for
the data flows over the radio interface. In other sense, the AS and NAS define how the data and
control planes shall work, including all the network processing units that should be involved within
each functionality. Then, the term protocol stack shows up. A protocol stack then describes the
necessary and ordered steps with which a user or network device shall perform to carry and
understand either a control or data flow. The 5G users hold both the radio protocol stack (for data
transmissions over radio) and the NAS protocol stack (for control transmissions over radio).
herein, we will briefly and solely go through the radio protocol stack.
Layer-1: the physical layer of the protocol stack is basically responsible for the data initial
reception and transmission including bit modulation, encoding, power control, error detection and
Layer-2: the medium access control (MAC) layer of the protocol stack performs data
retransmissions upon failed receptions, user and resource scheduling, and quality of service
Layer-3: the radio link control (RLC) of the protocol stack offers data segmentation or
Layer-4: the packet data converge protocol (PDCP) of the protocol stack applies radio ciphering,
31
Layer-5: the service data adaptation protocol (SDAP) is a 5G newly introduced layer. It has a key
function in order for the 5G new radio to support a diversity of quality of services at the same
time. SDAP maps and translates a certain quality of service requirement for a given data flow of
an arbitrary user into parametrized radio transmissions, that at the end would satisfy such
requirement.
It is worth noting that some researchers consider the MAC, RLC and PDCP layers to be part of
Layer-2. Well then, on the control radio plane, a user has two main layers as follows:
Control Layer-1: the radio resource control (RRC) handles the control information exchange over
the radio interface including the mobility control, and measurement reporting configurations.
Control Layer-2: the NAS layer is responsible for the user control connection establishment up
to the 5G core entities, i.e., establish a secure connection with the core interface.
Well, I know it may be a lot of boring protocol stack function details so far. So, let’s have an
– the user is informed by the network control signaling (RRC messaging) about how to measure
its operating bandwidth in order to identify the BS/gNB with the best receptive coverage, and on
which periodicity it should do so? So, when it gets awake, it connects to such BS accordingly. Then,
assume that a downlink data arrives for such user, such as an already-started video streaming,
then the BS starts from the SDAP layer to check for the required quality of service level requested
by such user. Generally, a user with a higher subscription level will get a better quality of service.
32
Thus, the SDAP translates that requirement into certain transmission parameters such as the
number of resources, service time, etc and feeds-back such information to Layer 2: the MAC layer
for later processing. Then, the PDCP layer compresses the data payloads, and applies radio
encryption for secure communications. Further, the MAC layer schedules such traffic over the
network resources, considering the input feedback from the SDAP layer to meet the quality of
service level of this user. Finally, the physical layer modulates, encodes, and transmits data
5G Spectrum
Currently, there is a progressive increase in data traffic volumes through our mobile
data networks. This introduces a vital technological question of how to resolve our
expected cellular capacity problems in future. For sure, there are a diversity of
network capacity.
–well, more specifically, the system capacity is simply restricted by the available
environments.
33
Thus, in order to meet the promised 5G new radio capacity specifications, a larger
amount of spectrum is necessary. Being said that, the 5G new radio is the first cellular
sub 6 GHz up to 100 GHz with scalable bandwidth ranges from 5 MHz up to 400 MHz.
–let’s recall that the maximum communication bandwidth, supported by the former
4G systems, is 20 MHz. 5G new radio simply supports up to 20x times more bandwidth
then. However, this imposes a lot of device hardware challenges though. For example,
mobile phones shall be required to scan the entire bandwidth at the times of special
measurements or cell selections, which is incredibly large with the 5G new radio, and to
do that as fast as possible. So, the RF chain of the mobile phone should be of extreme
high quality, that is not really an efficient requirement in sense of the cost metrics.
So, lets first start by introducing how the spectrum is shared between several telecom
operators in each country. The most common strategy which you may have probably
chunk or portion of the available spectrum. As a result, each operator shall solely have
the exclusive access to such spectrum. Honestly, this is the most appreciated
34
spectrum sharing strategy yet in 3GPP. However, with the 5G new radio, some other
For instance, the authorized licensed shared access aims to allow operators to have
partial and constrained access to some portion of the spectrum carriers that have
applications and use cases are not actually using the spectrum over all geo-locations
or at all times, e.g., may be only during military exercises. Thus, operators could gain
access to such spectrum, which is conditioned to certain areas or times. This way, the
the fact that the majority of our cellular traffic comes from the indoor environments
such as the malls, homes, etc. So, an efficient traffic steering can significantly relax
networks is not newly coming with the 5G new radio, but unlike the 4G standards, it is
The idea is that a 5G new radio mobile phone can steer its broadband traffic from the
35
the WIFI networks. Given that users are highly likely to be short distanced from the
WIFI access points, they can perceive a huge capacity improvement in addition to
offloading the congested 5G new radio networks from such significant amount of
traffic.
–honestly, there is another reason to offload traffic towards the WIFI networks. The
standardized 5G bands are quite large, e.g., 100 MHz for spectrum below 6 GHz. As
you know, such spectrum is highly congested as of now. It simply confines our former
2G, 3G and 4G communication systems. Thus, it is fairly difficult to have 100 MHz of
contiguous carrier components over such spectrum. Simply put, WIFI bands (2.4 to 5
Being said that, the WIFI spectrum is basically unlicensed and can be an insecure way
of communication for cellular connectivity as well, –as the WIFI access point could have
access to some of the private 5G network configurations as well as to the inflicted co-
channel interference. Thus, it is not a reliable way of communications for 5G new radio.
Accordingly, the 3GPP sets a great interest in such interoperability between the 5G
new radio and the unlicensed communications, by standardizing novel core interfaces
that simply maintain a secure inter-connectivity between 3GPP (5G/4G) and non-3GPP
well, –simply put, when and how shall users decide to offload traffic towards non-
3GPP RANs?
36
So well so far, the early 5G commercial deployments, expected during 2020, will be
allocated over the 3.3 to 4.9 GHz, with the time division duplexing (TDD) transmission
mode, and the 24 to 28 GHz for higher carrier bandwidths. Lower spectrum range
So, more specifically, two types of frequency ranges are defined for the 5G new radio
Frequency range 2 (FR2): all spectrum carriers between 24.25 to 52.6 GHz.
For each FR, the maximum bandwidth and sub-carrier spacing are set in standards.
transmission. With LTE, the sub-carrier spacing is fixed, i.e., does not depend on
operating spectrum, and always equals to 15 KHz. The sub-carrier spacing of the 5G
new radio can however be scalable and variant based on the operating bandwidth, and
Thus, the maximum 5G new radio bandwidth, supported within FR1, is 100 MHz;
however, it is 400 MHz within FR2. The sub-carrier spacings of 15 and 30 KHz,
37
respectively, can only be adopted in FR1; however, the sub-carrier spacing of 120 KHz
is only used within FR2. Furthermore, the 60 KHz sub-carrier spacing can be used over
Generically, and as depicted in below tables, the 5G new radio bands are classified into
The TDD bands imply that the downlink and uplink spectrums are not associated to
each other’s, basically because both transmission directions are not activated at the
same time, i.e., the system is either uplink or downlink. Thus, a single TDD band can
fulfill an individual TDD system. On another side, the FDD bands are so called as paired
bands, since an FDD downlink band requires a corresponding FDD uplink band to form
an operating FDD radio network. Finally, the supplementary downlink and uplink bands
are then introduced with the 5G new radio and are of a significant importance to boost
the end user experience. Think of the SUL and SDL bands as extra available bands that
can be freely utilized and appended to the operating spectrum in order to enhance
either the downlink or uplink capacity. For instance, assume that we have many
simultaneous and rich content transmissions over an FDD downlink band. This way,
38
the entire downlink band is fully occupied by just 50% of the offered traffic. Then, the
leading to a sub-optimal capacity accordingly. With the introduction of the SDL bands,
the radio network can configure some of the buffered users, with their associated
downlink traffic, to get instantly transmitted over an additional SDL, may be spanning
a different band.
the SUL bands, is concerned about the way a TDD system works. TDD denotes a radio
downlink and uplink, both over the same spectrum. The issue herein is the availability
of the TDD spectrum in standards is quite higher than, these used with the 4G
systems, i.e., 3.5 GHz band compared to (1.2-2.5 GHz) band with 4G. This means that
the TDD users need more transmission power to compensate for the additional
this issue, a 5G new radio user in TDD transmission can be configured with the
standard TDD band over 3.5 GHz band for the downlink direction; however, during the
uplink transmission opportunities, it can use a lower SUL band such as the 1.7 GHz
–a side note, the integration of the SUL and SDL bands is quite different from the
band aggregation of the 4G systems. For the latter, two or more bands are combined
39
together, and users can transmit over all involved carriers at the same time. Although,
with the 5G new radio SUL and SDL bands, users can only transmit over a single band
at a time.
So, to conclude this section, lets quickly skim over the available bandwidth levels of
each carrier range. Well, within FR1, there are 5, 10, 15, 20, 25, 40, 50, 60, 70, 80, and
100 MHz of bandwidth allocations. Although, for FR2, there are 50, 100, 150, and 200
–as a final note, opposed to the carrier aggregation of the 5G new radio spectrum, the
called as: bandwidth parts and will be discussed in greater details over upcoming
Sections. In a very simple sense, bandwidth-parts technology means that some users
may have the perception that system bandwidth is for instance 10 MHz, while the
actual system bandwidth is 100 MHz. A basic advantage of such technique is a great
reduction of the mobile phone processing overhead, as of now, users need only to
scan and measure smaller bandwidth allocations, despite that the 5G new radio bands
40
41
42
5G Flexible Frame Structure
This is one of the most differentiating aspects of the 5G new radio. But, let’s first quickly recall
that the frame structure of the LTE standards is fixed in both the time and frequency domains.
That is, with a frame length of 10 ms while the sub-carrier spacing is always set to 15 KHz.
–well, such inflexible frame design was quite sufficient towards the 4G use cases, but why is not
it satisfying the 5G new radio requirements? the answer gets us back to the diverse use cases of
the 5G new radio, for which latency-critical and latency-tolerant applications as well as large and
very small packet transmissions should be integrated over the same radio interface. Accordingly,
–let me give you a clearer example of this towards the latency critical URLLC traffic. As being
discussed before, such traffic arrives randomly in time at the transmitter side and needs to get
transmitted by the quickest possible way within just 1 ms of radio latency from data arrival at
transmitter until successful decoding at receiver. And as you may know that already, the 4G/5G
radio systems transmit data signals on the air over quantized steps in time. This is so called as
the transmission time interval (TTI). In other sense, every TTI period, the 5G new radio transmits
43
the scheduled data payloads over all radio frequency resources. With LTE, a single TTI is worth 1
ms. Thus, imagine that a URLLC sporadic packet arrives just after the start of the current system
TTI, then it will be buffered or queued at least until the next available TTI instance, that comes 1
ms later, –this is entitled as the frame alignment delay where packet arrivals are buffered until
they are aligned in time with the next TTI opportunity. This is definitely NOT satisfying the URLLC
latency limits. With the 5G new radio, the TTI can be significantly reduced down to 0.1 ms for
latency-critical traffic and as large as 1 or 2 ms for latency-tolerant traffic. Again, these variable
TTI durations are supported over a single radio interface but for different applications, users or
traffic types.
So, the 5G new radio is said to support multiple numerologies. A numerology denotes a reference
sub-carrier spacing of the radio interface. This corresponds to a certain OFDM symbol time and
hence, a certain TTI size, –the TTI duration with the 5G new radio is always expressed by a number
of successive OFDM symbols. Thus, as the 5G new radio supports 15, 30, 60, and 240 KHz of sub-
carrier spacings, the TTI length or duration can significantly vary accordingly.
With 5G, a radio frame is worth 10 ms in time. This is divided into 10 subframes, each is 1 ms. Then,
each subframe is always 14 OFDM symbols; however, it can be divided into a scalable number of
radio slots.
–important note, as being debated, the TTI duration is the actual transmission periodicity of the
radio interface. Then, you may be probably wondering why do we have the frame, and subframe
44
periodicities as well, despite that radio transmissions always occur per TTI? The answer is simple,
it is designed that way to allow for several system periodicities. As an example, user scheduling
for latency-tolerant traffic can occur per subframe basis; however, system broadcast information
can be transmitted per radio frame periodicity. Traffic for URLLC users can be processed,
scheduled and transmitted per TTI or slot basis for lower radio latency.
–you may probably also be wondering about why not to unify all system periodicities to the lowest
one which is the TTI duration. Once more, answer is to save up the radio control overhead. For the
previous example, if the system broadcast information is transmitted every TTI for instance, that
would result in a significant and not-necessarily needed radio overhead, greatly degrading the
–a slot or a mini-slot is also referred to as the TTI duration. Thus, with the 5G new radio, if a single
subframe is divided into 4 mini-slots, it implies the radio interface is able to transmit data signals
4 times during every subframe duration (1 ms). Recall that with LTE, in such case, system can only
–to unify acronyms herein, and for the rest of this book, consider the slot as the radio subframe.
Then, an arbitrary subframe or slot can contain a scalable number of mini-slots. A mini-slot is just
transmission opportunity of the radio interface. Thus, a mini-slot is the actual TTI accordingly.
–well, now let me give you an inclusive example about the frame flexibility of the 5G new radio.
Consider we have two users: one with latency-critical traffic such as URLLC, and other is latency-
45
tolerant and capacity-hungry user such as eMBB. The URLLC user is scheduled with a TTI duration
of just 2 OFDM symbols; however, the eMBB user is scheduled with a TTI of 14 OFDM symbols. In
other way, the system shall transmit or receive data towards/from the URLLC user every 0.14 ms
while for the eMBB user per 1 ms. You see the difference.
–thus, unlike 4G systems, with the 5G new radio, the radio interface can be transmitting data
users.
Well, massive MIMO is not quite a 5G specific technology. It has been in fact matured in academia
since many years ago. However, it did not yet have a chance to fly in 3GPP standards. There are
several reasons for this yet to be presented; though, lets first define what massive MIMO really
is?
–the wireless channel is composed of several paths from transmitter to receiver in the spatial
domain, i.e., in the free space. Some of these paths are truly stronger than others, e.g., some
paths may be weak (i.e., highly faded) due to for example, being blocked by an object among the
transmitter and receiver pair, or by several reflections, thus, energy towards these paths will be
almost scattered and lost. The optimal transmission strategy is indeed to direct the transmission
energy into these strongest paths towards the intended receiver. The fundamental problems
here are how to identify these paths at the transmitter side and how accurately the transmitter
46
can direct such transmission energy towards these paths? the answer is the channel state
information and massive MIMO communications, respectively. The channel state information
implies that the mobile phones simply inform their serving transmitters, i.e., base stations, about
their preferred channel paths using a quantized feedback. There are tons of proposals of how to
best efficiently exchange such quantized channel information. On another side, the massive
MIMO denotes that the transmission radio system is equipped with hundreds of antennas at the
transmitter and receiver sides. This simply enables more precise direction of the transmission
energy towards the strongest channel paths, being known from the user channel state
information.
By now, I should have motivated you to ask the question: why the hell did not massive MIMO
communications make it in 3GPP standards yet? Answer is the implementation complexity and
required use cases. With the 4G radio technology, massive MIMO had fundamental limitations
against its implementation. First, the required channel feedback overhead size from cell phones
hundreds of antennas, such overhead may consume the radio capacity of the control uplink
channels accordingly. Secondly, the 4G radio systems are almost likely deployed on sub 6 GHz
bands. Commonly, they are deployed over the 1.2 to 2.5 GHz bands to achieve a good tradeoff
between capacity and coverage. Thus, as we know from antenna design theory, the size of the
antenna arrays, and specifically the physical spacing between antenna elements, is a function of
the operating carrier frequency. The lower the operating band, the larger antenna spacing limits.
47
Thus, to have fairly large antenna arrays with the 4G radio, the physical size of the massive MIMO
arrays may go up so large, which is practically not feasible. Moreover, the need drive for the
massive MIMO was not really sufficient at this time, as the 4G radio has already offered a
So, you may be probably wondering: what is the drive for massive MIMO with the 5G new radio
then? First, it is the higher supported spectrum allocation, i.e., FR2 from 20 to 60 GHz spectrum.
These bands are very dispersive in nature. This means the channel shall have way too many spatial
paths for a signal to take between two ends. Herein, massive MIMO with advanced beamforming
is an ideal solution of this issue. Much better news, because the carrier frequencies are so high
for these bands, the size of the massive MIMO antenna arrays is highly reduced to be feasible in
practice. Secondly, over the past couple of Sections, we have been discussing that 5G new radio
should simultaneously support a massive number of connected devices, i.e., mMTC. Thus, the 5G
new radio requires excessive available system capacity at the same time, while the standard
orthogonal user multiplexing techniques, e.g., OFDMA, are capacity-limited by the number of
capacity can simply be increased by factors of 10; why, simply because base-stations become
able to schedule users on different spatial directions or orientations; however, on the same time
and frequency radio resources, instead of them being scheduled separately in time or frequency.
That is simply one more added degree of freedom to maximize the system capacity. Accordingly,
48
the associated capacity gains of the massive MIMO technology can be further enhanced more by
Furthermore, in theory, massive MIMO communications work out in both the frequency and time
division duplexing modes (FDD and TDD). In FDD, the channel state information from mobile
phones to base-stations is mandatory since both the downlink and uplink bands are on different
spectrum allocations. However, with TDD, the spectrum is unpaired, where both uplink and
downlink bands are on the same spectrum. Hence, base-stations can figuratively estimate the
downlink channel state information from the corresponding uplink channels, –so called as TDD
channel reciprocity, thus, entirely removing the requirement of the channel station information
feedback overhead from users to transmitters and further enabling the massive MIMO
communications. This in fact aligns very well with the fact that the early deployments of the 5G
new radio are envisioned over the TDD 3.5 GHz bands, due to regulatory reasons. Thus, massive
MIMO can be well integrated with the 5G new radio deployments since Day-1.
Well, in that sequence, we will go through further complications of the massive MIMO
communications and what can it actually do? while keeping the same low level of details we have
always maintained.
First of all, there may a confusion between the term precoding and beamforming, so let’s settle
it down first.
49
–signal precoding at transmitter is a way to direct the signal energy of this user towards its
strongest channel paths, known from its latest channel state information feedback. So, precoding
is only concerned about a single user at a time. However, beamforming has mainly the same
function as the precoding process while also considering the other concurrent transmissions of
other users that may interfere with the intended user. In a simple sense, it maximizes the desired
energy towards the best channel paths of each user while minimizing its interfering energy with
–with precoding only, you may have a maximum energy towards a desired transmission, but it
can be the case also that such transmission is impacted by severe interference as well. Hence,
capacity may be degraded though. Massive MIMO significantly enhances both the precoding and
beamforming operations due to the antenna array gain and the more identified channel paths
So, beamforming is defined as a signal processing capability with which the input signal is phase
and gain altered on each antenna element across the entire array, in order to have a maximum
energy transmission towards intended users and theoretically zero energy towards other users,
–in practice, such absolute zero is not feasible; instead, it can be -20 to -60 dB less energy.
However, such accurate beamforming comes with the cost of the processing complexity.
Additionally, with former 4G radios, and actually all former cellular systems, antennas are so called
as passive elements. This means that they do not process signals and just air transmit what has
been input. Accordingly, the RF units behind the antennas have to do the full job of precoding and
50
beamforming. Then, through cables, –high quality ethernet cables, Fiber or others, they signal
these well-constructed RF signals to antennas for air transmissions. That could be reasonable if
the number of antennas is fairly small. With massive MIMO, we simply need way too many of these
high-quality cables.
As an alternative way, an active antenna design has been introduced with the 5G new radio. This
implies that antennas now can process input signals, combine them, and finally transmit them
over air, –simply, antennas have become active elements. Well then, each individual antenna shall
have an RF unit right behind the antenna on the physical tower. This way, no cables are even
needed.
–in the standard setup, an RF unit can serve more than one physical antenna to reduce cost, such
that there will be a difference between the numbers of the mounted antennas and RF units,
respectively. In the antenna design theories, herein in this case, the RF units are called as the
antenna ports while the antennas themselves are so called as the physical antenna elements.
–remember that, the number of physical antennas determines how directive is the antenna array.
Simply put, more antenna leads to much more beamforming directivity in the space. Although,
the number of the RF units behind antennas defines the beamforming capability in sense of how
many beams or signals the antenna array can process simultaneously. And finally, the
configuration of the antenna array, either planar, linear, circular, etc, defines the dimensionality
of the beamforming. This means, for example, a linear array, configured only in the horizonal
51
directional, can only beamform the horizontal space. Although, for instance, a planar array with
antenna elements being setup horizontally and vertically, can beamform the horizontal and
vertical directions. With this case, in theory, the base-stations are able to truly point the
transmission directions nearly towards the exact locations of the intended users. This is simply
Well then what about the 5G numbers for advanced beamforming? The 5G new radio, starting
from 3GPP release-15, could indeed support up to 64 RF units while it was limited to 32 with
LTE/4G release-14. Furthermore, as being debated before, one significant differentiation aspect
for the 5G new radio is its support for control channels’ beamforming. With the 4G radios,
advanced beamforming was only applicable to data channels; however, to reach an ultimate
system reliability with the 5G new radio, this has been changed.
Also, without going into too deep details, as you may probably know, the MIMO capacity is always
limited by the lower number of the mounted antennas at both the transmitter and receiver,
respectively. Clearly, this is an issue with the 4G radios, where users can have a maximum of 4
antennas. With the 5G new radios, there is no a standard limitation on the mobile end capability.
One final remark to conclude this Section, over higher carrier frequencies, i.e., 5G FR2, antenna
radiation patterns become much more directive, especially with more and more mounted
antennas. Hence, the 5G transmissions over these spectra would be very narrow and very
directive. In other mean, with the former 4G radios, basically base-stations transmit in all
52
directions. Then using precoding, they attempt on best-effort basis to focus the energy towards
the best channel paths of the intended user(s). That is also one reason why we tend to physically
down-tilt the antenna panels, in order for them not to overshoot other surrounding base-stations
as well. Although, with the 5G, this is not a problem anymore because transmissions simply have
become beam-based and being directive with a fairly good precision to the exact locations of the
users.
This very same transmission behavior introduces a lot of other issues though with the 5G new
radio. Now, the 5G transmissions, including both control and data channels, will be sensed by
users only towards special directions because they are way too narrow.
separated users, the question now is: as the transmission beam/direction is very narrow in the
space, so, in which direction should the base-station transmit and beamform such information?
Answer is basically to make these transmissions on user centric basis, instead of being single
network-centric transmissions. Thus, with the massive MIMO communications over the FR2,
every 5G transmission and reception shall be directed towards a single mobile phone, and that
– you are probably thinking this may be too much overhead to send both data and control
information separately towards each users’ direction. Yes, you are ultimately right here. That is
why the 5G new radio borrows a similar concept from the former 4G communications, which is
53
the channel state information reference signals (CSI-RS), –no worries if you do not follow, check
below example.
–in both LTE and 5G radios, there are two control signals so called as the primary and secondary
synchronization signals (PSS and SSS). Basically, users utilize them to get to know the base-
station identification and to have themselves synchronized, i.e., their RF clocks, with the
transmission periodicity of the intended base-station. With the 4G technology, these signals are
being sent with a predefined time periodicity and over specific frequency resources. In the
spatial/direction domain, these are transmitted basically over the entire sector coverage because
these signals are not beamformed. Thus, users just have to monitor and scan the operating
bandwidth for some time until they could lock on the PSS and SSS signals accordingly. With the
5G new radio, instead, the sector coverage space is divided into several divisions or parts, where
at first, base-stations transmit the PSS and SSS signals towards the first division, i.e., using a
beamformed transmission towards that direction. Sometime later, it repeats the same process
but towards the second coverage direction and etc until it spans the entire sector space. Upon
that event, base-stations start over from beginning again. Thus, 5G users, regardless of their
locations, will have a chance to lock on a PSS/SSS beamformed transmission that is closest
possible to their geo-locations. Accordingly, such beam sweeping in time needs to be done quite
fast, so not to make users to wait for long times before they can get a sufficient PSS/SSS
beamformed transmission.
54
During connected mode, the 5G new radio shall instruct users accordingly on which directions
they should receive or transmit. Thus, one would expect that some beam failure mechanisms are
–you see the difference, users in idle mode try to catch the beam that is closest possible to their
locations, using their own sensing capabilities; why, because base-stations do not yet identify
these users. Later, during user connected modes, and for a more efficient and controlled
beamforming operation, base-stations formally instruct the active users of the beams that they
–this entire procedure lies below the umbrella of the beam scheduling and monitoring of the 5G
new radio standardization, which includes techniques for beam measurements, beam scheduling,
Bandwidth parts technology is being newly introduced with the 5G new radio. Although, lets first
define what the problem that such technology is addressing here? As of now, we have clarified
the standardized 5G new radio spectrum and potential bandwidth allocations over FR1 and FR2.
Recall, a maximum bandwidth allocation of 100 MHz and 400 MHz is supported for 5G FR1 and
FR2, respectively. This is simply way too much, compared to the former 4G where the maximum
supported bandwidth is 20 MHz. Though, what is the problem coming from? Simply, from the
devices, i.e., the mobile phones. During special events, cell phones need to monitor the entire
55
communication bandwidth, for several different reasons, such as frequency measurements,
frequency scanning, and initial wake up operation. This sets a clear restriction on the required
complexity of the radio chain of the cell phones in order for them to scan a very wide bandwidth
as fast as possible. As a result, 5G cell phones shall get very complex and costly as well.
A simple solution; though, not simple at all, could be feasible by converting the entire large
bandwidth chunk into several smaller bandwidth pipelines or parts, where from the mobile phone
perspective, these smaller bandwidth parts would be understood as the entire system bandwidth.
Thus, users need only to monitor and scan these smaller bandwidth allocations.
Accordingly, based the 5G new radio definitions in 3GPP specs of Rel-15, the system carrier
components now have common physical resource blocks (CRB), unlikely with 4G PRBs, which can
be assigned to any configured bandwidth part. Thus, a bandwidth part actually contains the
system PRBs. A bandwidth part is then defined by a starting PRB index and a number of
successive PRBs. It can be utilized as downlink or uplink and users may have been configured with
However most importantly, bandwidth parts follow a modular design. This means that each
bandwidth part can have its own numerology, –do not you remember what the numerology is?
please revise the ‘5G flexible frame structure’ Section, its own bandwidth, its own transmission
56
and multiplexing users with different requirements over different bandwidth parts, that in fact
adapt their own configurations in time to the associated user targets and radio states.
–bandwidth parts technology is still an ongoing research towards the 5G new radio though. Its
baseline network and user signaling have been set as of now. Simply, how the network and users
will be coordinated on a bandwidth part agreement. Though, for keeping the detail level at
Since the time of the 4G systems, the frequency and time division duplexing (FDD and TDD)
–an FDD transmission denotes that the downlink and uplink bands are deployed on separate
spectra at the same time. Accordingly, users can transmit in the uplink while base-stations are
transmitting in the downlink as well. However, with TDD, both uplink and downlink bands are
unpaired. This implies there is a single band which can be used either for downlink or uplink
transmission at a time. Usually, networks alternate between the downlink and uplink transmission
many downlink transmission opportunities within a radio frame, i.e., every 10 ms, than the
corresponding uplink opportunities, when they have much offered downlink traffic than the
uplink one.
57
–thus, it in principal implies that the system adjusts its transmission direction based on traffic
availability. Assume the hypothetical scenario of a single base-station with a single user, which is
receiving in the downlink direction at all time. With FDD, the downlink band can be fully allocated
to this user; however, the paired uplink band shall stay idle without any transmissions. This is a
significant loss of the resource utilization. With TDD on the other hand, the base-station shall
decide to allocate the single TDD band to downlink transmissions for a longer time period until
Accordingly, FDD clearly requires double the bandwidth of the TDD. But, it also achieves doubled
capacity. Then, what are the key deciding factors with which a network operator shall go for TDD
or FDD system?
Simply, the need and availability. For some telecom operators, available bands may not be
allocated over the FDD standardized spectrum, thus, they may not have a choice except to utilize
the TDD technology. Furthermore, when spectrum and operational expenses are primarily the
main concern, TDD is then always the best option, –operators basically pay for a single band
instead. Moreover, as being discussed over the ‘massive MIMO’ Section, TDD makes such massive
MIMO technology feasible in practice up to this very moment. With FDD, the channel state
information, as being pointed out before, is a critical obstruction to implement the massive MIMO
antenna arrays while this problem is not as severe, or basically not there at all, with the TDD
technology.
58
–in fact, some operators may also choose TDD because of its transmission flexibility, especially
to the sporadic traffic patterns. Referred to previous example, when sporadic traffic is applicable,
e.g., 5G URLLC-alike traffic, TDD offers moving the system capacity between the downlink and
uplink directions dynamically in time based on the available traffic pattern. This is very convenient
to the 5G industrial cases where devices such as robots are expected to send updates during
specific times. Thus, the network configures itself with sufficient uplink transmission
opportunities during the times these end devices should be transmitting their updates in the
Although, TDD was not that flexible since beginning. Coming up with the 4G radio technology,
TDD was so called as a static TDD. A static TDD denotes that every radio frame periodicity, i.e.,
10 ms in time, there are specific radio subframes that are allocated to downlink and others to
uplink. This is predefined from the network planning and configuration phase and is unified across
all base-stations. This is a critical limitation for two main reasons: first, such predefined frame
structure of fixed downlink and uplink transmission subframes may not match the available
traffic pattern and load at a given time unit. Second, not all adjacent base-stations may have a
similar traffic patterns at all times, so to have a unified frame structure for all base-stations is
clearly suboptimal.
Thus, one stage later, with the 4G-Pro standards, 3GPP has introduced some dynamicity to the
TDD operation. Several radio frame configurations with different numbers of the downlink and
uplink transmission subframes within every radio frame are introduced, where base-stations can
59
dynamically in time select the one frame that best meets its traffic demand. For example, a base-
station with much higher downlink available traffic than uplink, selects a radio frame with higher
Although, the minimum frame update periodicity was actually the frame duration itself, i.e., 10
ms. This is considered way too large update delay since with the 5G new radio, URLLC-alike traffic
type is expected to fluctuate per 1 ms basis. Thus, such TDD frame adaptation periodicity should
Thus, in 5G standards, in fact, such frame periodicity can be based on the mini-slot basis, that can
be way less than 1 ms. In a simpler sense, with the 5G new radio in fully flexible TDD systems,
traffic demands.
Being said that, although, there is one critical issue against such TDD flexibility. The issue occurs
when neighboring base-stations may have different directional transmission directions. For
instance, at an arbitrary time instant, a base-station with downlink subframe while adjacent base-
station adopts an uplink subframe. Now, we inflict an additional type of interference. That is the
cross-link interference, where the downlink transmission, with the larger power from the
downlink base-station, interferes with the lower power uplink transmission at the adjacent uplink
base-station.
60
Such cross-link interference issue is one of the most critical and yet open research problems, and
on top of the 3GPP agenda during the ongoing release-16 standardization, and potentially
5G Multi Connectivity
Multi-connectivity refers to the situation when mobile phones are connected to, i.e., receive and
transmit on, resources from different base-stations at the same time. These could be over radio
–well, the multi-connectivity was also there with the former LTE technology. With LTE, the data
planes of the user transmissions could be shared or aggregated between different LTE/ 4G base-
stations. In sense, users could either have their data transmissions duplicated from two adjacent
base-stations for extra link robustness or data aggregation for boosting the achievable
throughput. One important note here, the control links of users are still within a single
–thus, one clear difference herein is that 5G multi-connectivity include the control plane as well.
Thus, 5G users could have their data and control links, established from different base-stations.
Hopefully motivated enough to conclude that towards the 5G new radio targets, multi-
connectivity has become of a significant importance, especially during the early 5G deployments.
First, what benefits does multi-connectivity bring to the 5G new radio? In a simple way, a lot. For
instance, it boosts the system reliability and user throughput. However, most importantly, multi-
61
connectivity shall make it very smooth to deploy the early 5G new radio implementations. At first,
it would be way too expensive to deploy a standalone 5G radio and core architecture. Thus,
operators are expected to have tight interconnections between existing 4G structures and the
newly introduced 5G systems. For example, users establish radio connections with the 5G new
radio and core connectivity with the former 4G evolved packet core structure. Herein, there are
coverage due to the lower spectrum, as well as 5G connections of beyond 60 GHz, to boost the
data rates.
Inter-RAT multi-connectivity: basically, users establish and maintain connections with 5G and
4G radios as well.
– although, multi-connectivity is not all-green scheme here. It has a fundamental and pricy
limitation though. Multi-connectivity could indeed boost the user throughput and link reliability.
However, it may highly degrade the system spectral efficiency due to the poorly utilized radio
resources. Thus, 3GPP is considering several creative solutions to fast switch users back and forth
among the single and multi-connectivity modes. That is, the network shall decide which users and
62
Furthermore, 3GPP has standardized several configurations for the 5G multi-connectivity
–herein, I mean that users could actually divide their protocol stack into several slices, and have
–Over the upcoming book notes (Part-2) of the detailed 5G service-based core architecture, we
shall go in much more depth into the 5G dual connectivity, and the associated functional split.
As has been pointed out over the ‘5G Massive MIMO and Beamforming’ Section, transmissions
over the 5G new radio are always beamformed. This denotes that base-stations and users
communicate to each other’s only in certain directions at certain times. The benefits of such
behavior are quite significant in fact. For example, a great reduction of the inter-user interference
has become feasible then, since users may not even see other users’ concurrent transmissions.
To be mentioned also, the antenna array gain boosts the user achievable capacity and network
coverage as well. However, this imposes a simple question herein: how the common control
channels are being transmitted in the downlink direction? These channels need to be acquired by
–well, the common channels of LTE are always transmitted towards all directions at a time. More
63
periodicity within specific radio frequency resources. Although, on the spatial domain, they can
Thus, to truly put some hands-on knowledge here, lets recall the PSS and SSS signaling we have
gone briefly through over former Sections. These signals are quite essential for respective users
to camp on a base-station. Next, users tend to perform the radio random access procedure to
get their control and data connections established, from the radio and core sides, respectively.
Thus, such signals simply need to be sensed within every short time period from all users, which
With the 5G new radio beamformed access, in a simple way, the PSS and SSS are beamformed
into specific spatial directions at a time. A single time unit later, they are sent again on the same
frequency resources but towards an adjacent spatial direction instead (with simple angle
rotation), and so until the sector area is covered by beams. This is simply so called as beam
sweeping, since we simply sweep or slide the beamforming coverage in time. In this regard, 3GPP
has set up several definitions. An SS block (SSB) basically implies a single beam direction at a time.
It is entitled as block information because such beam indeed holds a block of information
elements such as the SSS, PSS, and main information broadcast (MIB). This means that a user in
a certain location just needs to lock on a single beam (the closest one), because every beam
modulates the entire information needed by this user end. Then, an SS burst is defined to include
all SSBs within a certain SS transmission time, –no worries, check the below example.
64
–say the SSS and PSS information need to be transmitted every 20 ms, and for a continuous
transmission duration of 4 ms across specific radio frequency resources. Thus, let’s assume we
have four SSBs, with four beams into four different beamformed directions. At the first eligible
ms, SSB 1 into beam 1 is transmitted. At the second eligible ms, SSB 2 into beam 2 is then
transmitted and so on. Accordingly, SSB 1, 2, 3 and 4 are just called the SS burst transmission at
As the name suggests, a reference signal term denotes a pre-defined and pre-known signal to
both transmitter and receiver. Such signal is transmitted for the sake of the receiver to estimate
how much signal damage (basically amplitude distortion) can be inflicted due to the wireless
channel in between. Thus, it can compensate such damage later when actual data packets are
being received. Herein, a major change has been enforced on the reference signal design within
the 5G new radio systems. However, to make it clear and simple, lets first recall how reference
Well, generically, the 4G radio standardized the cell-common reference signals. Simply put, these
are cell specific rather than user specific. Accordingly, they are being periodically sent in time over
specific frequency sources, regardless of the connected user capacity, i.e., how many users are
connected at the mean time. However, starting from LTE release-10, 3GPP has introduced the
user-centric reference signals, i.e., user specific. This means that the radio interface genuinely
transmits reference signals towards a specific active user and on a set of time-frequency
65
resources that are only accessible by this user. Basically, such type of the reference signals is
Now, here is a question: why are these cell common reference signals not efficient enough to be
utilized with the 5G new radio? Clearly, these signals will be transmitted at all times. Just imagine
the case that a very low number of users are connected to an arbitrary 4G radio interface. These
users may require much lower reference signal transmission overhead than what is actually being
transmitted, in order to estimate their wireless channels. This is accordingly a significant resource
loss then. One more reason, as we know by now, 5G transmissions are beamformed. Thus, users
may not be separated on the resource domain but rather on the spatial domain.
–this denotes with the 5G new radio, base-stations may then schedule different users at the same
Thus, it becomes fairly important to have the user channel estimates also based on the
transmitted beams, instead of only being based on the time and frequency resources.
Being said that, the 5G new radio only considers user specific reference signals for both control
and data channels. One clear advantage here is that, unlike the 4G radio, when the number of
connected users is quite low, the 5G new radio does only transmit reference signals towards the
frequency resources occupied by these users, with much lower control overhead compared the
LTE case, with which reference signals spanning all system time and frequency resources will be
transmitted, even if some of these will not be used; basically because there are not enough users.
66
–for that reason, it has been always said that the 5G new radio should be flexible enough to send
So, let’s get more specific here, 5G new radio uses the same 4G reference signals as well as
introducing a new reference signal design for phase tracking. Accordingly, the 5G new radio
Demodulation Reference Signal (DMRS): from its name, DMRS helps the receiver to de-
modulate the received signal, by suppressing the negative effects of the wireless channel in
between the transmitter and receiver, i.e., channel distortion and power decay effects. Hence, a
generic transmitter shall always append the DMRS symbols within its to-be-transmitted data
payload for the other end to be able listen to DMRS symbols, estimate channel fading, and finally
decode the useful data payload. If DMRS symbols are not present in the transmitted payload,
such payload may not get successfully decoded at its intended receiver; thus, DMRS are quite
essential. To be mentioned also, DMRS is a dual reference signal, meaning that it can be used
Sounding Reference Signal (SRS): once more, as the name implies, such reference signal is to
sound the wireless channel. Well, you are probably confused, and you are rightfully right. Indeed,
SRS does almost exactly the same job as the DMRS, by just helping the receiver to estimate the
inflicted channel distortion. The only difference is that DMRS are only transmitted when actual
and useful data payload is about to be transmitted as well; however, SRS can be either
67
transmitted periodically or based on an event occurrence, i.e., event-triggered, from a user end
to its e serving base-station. Also, you probably noticed, the SRC is only transmitted in the uplink
Channel State Information Reference Signal (CSI-RS): it is the analogous reference signal to the
SRS; but, in the downlink direction. The CSI-RS has a vital importance for MIMO communications
as simply users need to tell their serving base-stations about their downlink channel estimates,
–without an accurate CSI-RS estimation and CSI feedback (as briefly discussed before), advanced
So far, the former reference signals have been there also with the 4G standards. However, the
5G new radio introduces a new reference signal design. But, before going through the last
reference signal type of the 5G new radio, lets recall something about how a wireless channel can
distort our wireless signals? Theoretically, a channel introduces an amplitude and phase error or
distortion to our useful transmitted data. So, the receiver shall perceive a completely different
version of the transmitted data symbols and then, may not be able to decode them correctly.
This channel distortion is time and frequency variant, –that is a reason why the radio interface
Apparently, the DMRS, SRS, and CSI-RS help the receiver end to estimate the amplitude distortion
of the received symbols, hence, receiver is able to compensate for it and correctly decode the
68
data payload. A logical question pops up here: where is the phase compensation? Simply, with the
4G radio, phase compensation was not an issue. The majority of the 4G networks are deployed
over the sub 6 GHz spectrum, where practically it was proven that the phase noise or error of the
wireless channels over this spectrum is minor. Although, with the 5G new radio, new spectrum
ranges from 20 to 60 GHz can be integrated with the 5G radio standards. Herein, we got the issue
back again and for that reason, the 5G new radio defines a new reference signal.
Phase Tracking Reference Signal (PTRS): probably now, you already know what is it about? PTRS
is a reference signal to simply track the phase error or distortion of the wireless channel for the
receiver to compensate for it prior to decoding. PTRS signals are transmitted from one OFDM
symbol to another. And yes, that is quite much overhead indeed. The reason for that over the
higher ranges of FR2, the channel phase distortion is quite aggressive such that a phase error of
complete few degrees can be inflicted from an OFDM symbol just to the next one.
–before I finally close this Section, I would like to stress out that there are further pieces of the
cake here. The 5G new radio in fact integrates several other vital system blocks that critically
impact the overall performance such as, for instance, the dynamic user scheduling, the in-
resource control channels, and many more. These will be covered in depth over upcoming book
notes.
69
Beyond 3GPP 5G Release-15
Release-15 has been a big milestone in mid-2018 for 3GPP towards achieving the 5G promised
the end of 2019. As usual, release-16 has several work items, tackling several issues. Herein, we
Further enhancements to E-UTRAN (4G): this work item specifically investigates the integration
of the 5G new radio functional split, explained earlier; however, into the 4G-Pro standards. This is
basically to make it easier when the 5G new radio protocol stack is split between 5G and 4G radios.
Further enhancements to narrow band internet of things (NB-IoT): NB-IoT applications are
commercialized starting from LTE release-13, while transmissions over narrower subcarrier bands
were the main shot there. Further enhancements include terminated user transmissions, i.e.,
users can early decide to terminate transmissions upon completion, for an ultimate saving of the
Non-terrestrial communications for 5G new radio: this work item studies the initial
requirements towards the full integration of the satellite communication into the 5G new radio.
This includes, but not limited to, roaming and data offloading from 5G new radio to satellite, and
Towards the promised 5G system targets, the following work items are ongoing as well:
70
MIMO enhancements for 5G new radio, including the MU-MIMO, downlink and uplink
71
Concluding Remarks
By this, we conclude all discussions presented in these book notes. As being evident,
practice, respectively. This year, we are also expecting early Beta commercial lunches
of the 5G new radio across the globe. Creative and breaking use cases and novel
applications shall continue development and further integration into our cellular
technology out of these standardization activities. Hopefully, these book notes were
sufficient and simple-to-digest enough to give you a glimpse of how the 5G new radio
research activities are currently ongoing, as well as to the main concept creation of the
Over the upcoming book note series, the 5G core system architecture associated with
For further feedback on the content of these notes, please feel free to get in contact
72
References to 5G 3GPP Documents
73
TS 38.508-2, 5GS; User Equipment (UE) conformance specification; Part 2: Common Implementation
Conformance Statement (ICS) proforma
TS 38509, 5GS; Special conformance testing functions for User Equipment (UE)
TS 38.521-1, NR; User Equipment (UE) conformance specification; Radio transmission and reception; Part
1: Range 1 standalone
TS 38.521-2, NR; User Equipment (UE) conformance specification; Radio transmission and reception; Part
2: Range 2 standalone
TS 38.521-3, NR; User Equipment (UE) conformance specification; Radio transmission and reception; Part
3: Range 1 and Range 2 Interworking operation with other radios
TS 38.521-4, NR; User Equipment (UE) conformance specification; Radio transmission and reception; Part
4: Performance
TS 38522, NR; User Equipment (UE) conformance specification; Applicability of radio transmission, radio
reception and radio resource management test cases
TS 38.523-1, 5GS; User Equipment (UE) conformance specification; Part 1: Protocol
TS 38.523-2, 5GS; User Equipment (UE) conformance specification; Part 2: Applicability of protocol test
cases
TS 38.523-3, 5GS; User Equipment (UE) conformance specification; Part 3: Protocol Test Suites
TS 38533, NR; User Equipment (UE) conformance specification; Radio Resource Management (RRM)
TS 38806, Study of separation of NR Control Plane (CP) and User Plane (UP) for split option 2
TS 38810, NR; Study on test methods
TS 38811, Study on New Radio (NR) to support non-terrestrial networks
TS 38812, Study on Non-Orthogonal Multiple Access (NOMA) for NR
TS 38813, New frequency range for NR (3.3-4.2 GHz)
TS 38814, New frequency range for NR (4.4-5.0 GHz)
TS 38815, New frequency range for NR (24.25-29.5 GHz)
TS 38816, Study on Central Unit (CU) - Distributed Unit (DU) lower layer split for NR
TS 38.817-01, General aspects for User Equipment (UE) Radio Frequency (RF) for NR
TS 38.817-02, General aspects for Base Station (BS) Radio Frequency (RF) for NR
TS 38818, General aspects for Radio Resource Management (RRM) and demodulation for NR
TS 38874, NR; Study on integrated access and backhaul
TS 38889, Study on NR-based access to unlicensed spectrum
TS 38900, Study on channel model for frequency spectrum above 6 GHz
TS 38901, Study on channel model for frequencies from 0.5 to 100 GHz
TS 38903, NR; Derivation of test tolerances and measurement uncertainty for User Equipment (UE)
conformance test cases
TS 38905, NR; Derivation of test points for radio transmission and reception User Equipment (UE)
conformance test cases
TS 38912, Study on New Radio (NR) access technology
TS 38913, Study on scenarios and requirements for next generation access technologies
74