Heterogeneous Networks: Challenges and Solutions in mmWave Communications

This paper reviews the challenges and protocols of mmWave communication in heterogeneous networks and proposes solutions to tackle them. It discusses the feasibility of mmWave communication in different deployment scenarios and explores the use of beamforming technology. The paper also addresses the issues of random access in mmWave cellular networks and the need for a new frame structure. Overall, it provides insights into the implementation of mmWave communications in HetSNets.

HETEROGENEOUS NETWORKS
By Name
Course
Instructor
Institution
Location
Date
Abstract - This paper is a review of five academic
papers on mmWave communication. The papers
reviewed are “10 Gb/s HetSNets with Millimeter-
Wave Communications: Access and Networking –
Challenges and Protocols,” “Random Access in
Millimeter-Wave Beamforming Cellular
Networks: Issues and Approaches,” “Low-
Latency Heterogeneous Networks with
Millimeter-Wave Communications,”
“Safeguarding 5G wireless communication
networks using physical layer security,” and
“Coverage and capacity in mmWave cellular
systems.”
Heterogeneous and small cell networks
(HetsNets) lead to an increase in efficiency of
spectrum and throughput through hierarchical
deployments. Attaining the capacity requirements of
future 5G unwired networks, the most feasible
technique that has been proposed is the mmWave
communications with unprecedented spectral
resources. Despite the better understanding of the
mmWave physical layer, there are still challenges in
implementing it in HetSNets effectively and
efficiently from the perspective of networking and
access. The 3GPP backwards-compatible frame
structure, with a basis on time-division duplex that
enhances backhaul links as well as high capacity
access. Further discusses the kinds of networking
challenges associated with the multihop characteristic
of the mmWave backhauling mesh.
In the recent past, the mmWave
communication has been given much attention thanks
to its enabling relationship with the 5G technology.
This is because there has been development in
mmWave beam forming technology as well as
campaigns in channel measurements championed by
the industry together with scholars. Thus, this article
contains an analysis on basic issues of RACH with
regards to mmWave communications and give the
possible solutions to tackle the challenges and give
future directions on the same.
In this section of the article, a review on
HetNets end-to-end latency with mmWave
communications has been given. Generally, there is a
challenge in formulating and optimizing the delay
problem with buffers in mmWave cellular
technology, owing to the fact that conventional
techniques of graph-based network optimization do
not apply when queues are put into consideration.
Thus, this paper has attempted to form an adaptive
low-latency strategy that employs the use of
cooperative networking in reducing end-to-end
latency. In the end, there is a demonstration that
Heterogeneous Networks 1

proper cooperative networking is significant in
lowering end-to-end latency.
Index terms
Beam scheduling, BS, Cell densification, HetSNets,
FDD, LOS, MeNB, mmWave
communications, MS, NLOS, OFDM, QoE, QoS,
RACH, Random access preamble, RAR, RAT, RLF,
SenBs, Spectral efficiency, TDD, UEs, Zadoff-Chu,
3GPP LTE network
INTRODUCTION
Various quality of service requirements for
wireless communication traffic load has been on the
upward trend exponentially as a result of the
widespread use of applications of mobile internet by
smart terminals. This trend is expected to continue,
requiring the new and emergent wireless network
designs to take care of these requirements. The
emerging 5G network technology is expected to
come with superior features that will enable users to
experience less of the communication challenges
associated with the current 4G and 3G technologies.
These features may include, but not limited to
features such as achieving a factor of 10 to 100
higher user data rates, and enhancing the overarching
vision of tactile internet that will need to have an
end-to-end latency which is lower than 5ms so as to
offer a response time which is ultra-fast.
Similarly, to cope with the ever-growing
mobile traffic, an enhancement on radio access
technology is key, by increasing cell density,
increasing spectral efficiency and an increased
bandwidth frequency. Cell densification has a
challenge in handling a large volume of interference
between cells due to the fact that cell deployment
usually occurs in an unplanned way, and the
interference management signaling is limited by
network non-ideal backhaul links. Thus, with respect
to these challenges, dealing with the traffic demand
will need a straightforward solution like increasing
the communication bandwidth. Owing to this, the
mmWave band has been deemed as the most
promising cellular network band.
Latency is considered to be one of the key
players in the future of mobile communication,
especially in QoS. However, low-latency is an uphill
task in 5G heterogeneous networks thanks to the
following reasons:
● The use of buffers in 5G for handling the
unexpected heavy traffic, while the queuing
delay seriously malfunctions the QoS in 5G
● Performing networking optimizations for
lesser latency is made difficult by the
diversity of HetNets architecture and/or
RATs
Deployment of mmWave communications in
HetsNets
Basically HetsNets comprises of a multiple
layer of radio access nodes in a 3GPP LTE network,
for instance, a macrocell eNB (MeNB) and several
multiple small cell eNBs (SenBS) like femto, relay
and pico eNBS. Such networks operate by having
each of their SeNB combining its backhaul data with
data obtained from the other network nodes then
sending to the MeNB. Separating SeNBs with short
distance ranges of 100-200m enables the mitigation
of severe propagation losses. Similarly, mmWave
radio communication can be used to provide
coverage within the small cells, thereby limiting the
interference level undergone on the sub 3-GHz
frequency bands applied in traditional cellphone
communication [1].
There are various deployment scenarios for
mmWave communication in HetsNet used for user
access links and backhaul as discussed briefly below.
Heterogeneous Networks 2

Figure 1. Illustration of typical deployment scenarios
with both mmWave and microwave systems.
Scenario 1 (Baseline) - a wired backhaul such as an
optical fibre is used to connect a MeNB to the SeNB
that it donates. This is a traditional approach where
either the MeNB or SeNB serve the user equipment
UEs on the microwave band. This scenario requires
that interference between the SeNBs and the MeNB
must be avoided maximally by carefully designing
interference coordination schemes.
Scenario 2 - here the UEs communicate to the MeNB
via a microwave band, while the UEs served by the
SeNBs operate on an mmWave radio. A wired
backhaul is used to connect the MeNB to the SeNBs.
Despite this scenario not requiring an interference
coordination scheme, the UES must support dual
bands in order to have a smooth handover between
the SeNBs and the MeNB.
Scenario 3 - here, the mmWave radio comes in
handy for backhaul transmission between the SeNBs
and the MeNB. Its implementation needs a single-
hop for backhaul transmission using mmWave radio.
This scenario achieves a quick deployment of SeNBs
due to the fact the UEs are not changed as the
network facilities get an upgrade. As in the case of
scenario 1, advanced interference coordination must
be performed in this scenario [2].
Scenario 4 - In scenario 4, single-hop wireless
backhaul adopts the use of mmWave communication
for the SeNBs. In addition, the UEs are served in a
small cell by the SeNBS via mmWave radio that
thanks to its tremendous bandwidth can raise
significantly the capacity of the network.
Scenario 5 - Improving network capacity can also be
achieved by increasing the reusability of
geographic spectrum, leading to a dense small cell
deployment. Dense SeNBs can be effectively
connected with the MeNb using the multihop
wireless backhaul. This scenario is where the SeNBs
cooperate amongst themselves and via an mmWave
radio communicate with their donor MeNB.
In conclusion, scenario 1 is the baseline, while
scenario 2, 3 and 4 act as a subset to scenario 5.
Frame structure
Heterogeneous Networks 3

MmWave communication bands’
propagation features are not similar to those of
microwave bands like the multipath delay and
Doppler frequency shift. Thus, the OFDM parameters
for 3GPP LTE systems’ microwave communications
cannot be applied to mmWave networks unless
modified. Contrastingly, mmWave communication
bands contain much bigger bandwidths of frequency,
meaning there is need for enlargement of bandwidth
per subcarrier so as to maintain the complexity and
size of the first Fourier transform (FFT).
Furthermore, it is compulsory to consider the
Figure 2. Illustration of the proposed frame
structure: a) modified TDD frame structure for
mmWave communications; b) mmWave subframe
configuration for multihop communications
backward compatibility with 3GPP LTE systems
when bringing on board the mmWave networks to
HetSNets [3]. Therefore, a new frame structure must
be designed for mmWave communications in order
to enhance an ultimate efficient design of system that
coexists with microwave HetSNets communications.
.
Adoption of a wireless backhaul in the
HetSNets with mmWave systems enables the
mmWave subframe configuration to support
multihop transmission. For a single transmission, an
mmWave subframe can help.
MAC and networking design challenges
A highly directional beam forming increase
the number of difficulties with regards to the design
of the network. This article only puts its focus on
networking layers and medium access control.
HetSNets has made it possible to have a multihop
transmission thanks to small cells’ dense deployment
using narrow beams possessing with large mmWave
systems bands gain [4]. The vast spectral resources of
the mmWave communication radios will enable users
to enjoy an experience that will offer unprecedented
services that are similar to wired technology. This
paper attempts to study the feasibility of the proposed
mmWave communication under various scenarios,
and further propose a new technology of 3GPP LTE,
which is a backward-compliment frame structure.
This enables this proposal to come up with
a technology that answers the challenges in mmWave
technology and at the same time attain an aggregated
cell throughput of approximately 13Gb/s, a
magnitude order that even the best current 5G design
cannot match.
MmWave Beam forming Cellular networks
Unlike in the traditional cellular network, a
mmWave cellular network uses a highly directional
beamforming at both the MS and Bs, where a huge
number of antenna elements are packed into a small
device due to the mmWave ability to have small
wavelengths of usually 1-10mm band. A transmitter
chooses a transmit beam pattern that determines the
Heterogeneous Networks 4

weights of phase shifter for steering the beam
towards a particular direction. Thereafter, the
receiver chooses a pattern for the receive beam for
receiving the signals in a particular direction. To
acquire a high beamforming gain, directions of
receive and transmit beams must have a proper
alignment with one another. When the MS is
connected, the index for the best transmit beam of the
BS is periodically fed to the BS from the MS by the
use of the channels of the uplink control and the best
MS transmit beam is reported via the channels of the
downlink control, so as to have the best pair for data
transmission using the best beam pair in uplink and
downlink [5].
Random Access in mmWave Cellular Networks:
omnidirectional antenna vs directional antenna
There exist cases where the best direction
cannot be initially known at either the MS or the BS,
an those cases are listed below:
● The instance when the MS tries initially
accessing the network
● The period when the MS is recovering from
an RLF
● When the MS is performing a procedure on
handover
In the above cases, commonly, the random
access procedure is conducted between the BS and
the MS. Owing to the fact that the best beam pair is
unknown, the MS is forced to transmit the RACH
preambles in several directions, making only a few of
the transmissions to attain a high beamforming gain
when the receive and the transmit beams approach
alignment.
Figure 3. The frame structure for random access in
mmWave beamforming cellular networks.
Critical issues in random access in mmWave
cellular networks
The use of beamforming technique is better
when used on random access preamble transmissions
by use of directional antennas. In the first random
access procedure in mmWave cellular networks, in
multiple directions the transmission of the preambles
occur at the MS and are received at the BS in
multiple directions [6]. In this step, the information
on the best transmit beam index at the BS should also
be relayed to enable the use of the best transmit beam
in the subsequent step. The second step of this
procedure involves the transmission of the RAR from
the BS with the help of the best transmit beam and is
received using the best receive beam at the MS. in
addition the best transmit-receive beam pair aid in the
performance of the third and the fourth steps.
There is likely to be a problem due to the
naive approach in the first step of preamble
transmission, as only a few of the receive and
transmit pairs can cause a high beamforming gain,
due to the misalignments in the directions of receive
and transmit beam. Thus, there should be a much
longer preamble duration for RACH than the one for
uplink data and control channels that utilize the best
pair of beam so as to achieve the target coverage.
Heterogeneous Networks 5

Figure 4. The random access procedure in mmWave
beam forming cellular
Networks.
Impacts of long duration of the random
access preamble
Impacts of four different aspects are
scrutinized in this section in order to identify the
important factors in mmWave cellular network.
RLF recovery and initial access
The random access procedure is performed
for the initial access. The MS needs to be in idle state
as long as possible in order to save power. This
implies that there has to be a short time of transition
from idle to connected state. Designing an efficient
state transition is not easy as a result of random
access experiencing long duration. Besides the initial
access, an occurrence of RLF enables the MS to re-
establish connection by use of random access. Long
random access time corresponds to a longer service
interruption time, thus heavily affecting the user
experience [7].
Handover
A large number of small cells will be seen in
mmWave cellular networks. The handover of the MS
frequently occurs as the MS moves through the cells.
This means the handover time needs to be as short as
possible to guarantee the QoE of network MSs [7, 8].
Nevertheless, since the time needed for handover
completion may be made longer by the random
access procedure, there may be a severe degradation
of the QoE f users, particularly for real time services.
Uplink-downlink configuration
TDD (time-division duplex) is more
appropriate than in FDD in mmWave systems, due to
the fact that spectrum flexibility and radio resource
efficiency are significant in wideband systems. Many
slots for the uplink time are much needed due to the
long RACH frame. Thus, its inflexible configuring
the uplink-downlink ratio in the system of TDD
based on the uplink-downlink traffic ratio.
Beam Scheduling
The number of receive digital chains in the
BS limit the number of receive beams that
simultaneously can be used in a time slot. The
smaller the number of receive digital chains, the
lower the flexibility of data beam scheduling. In
addition, the problem is escalated by the multiplexing
of different MSs’ uplink data signals in the frequency
domain instead of multiplexing in the time domain.
Heterogeneous Networks 6

Possible approaches to address the issues
To overcome the major challenge of random
access, which is the need to have long total duration
of the RACH, the following solutions are proposed.
However, one should note that, the preamble
bandwidth does not affect the RACH performance
much.
Enhanced Performance of Preamble Detection
For a certain cell coverage, improving the
preamble performance detection can reduce the
RACH duration. The performance is determined by
sequence design and the detection algorithm of the
preamble at the receiver. Designing a new preamble
sequence will not guarantee an improved
performance, because the Zadoff-Chu sequence
applied in LTE has desired properties such as
minimum cross-correlation and ideal cyclic
autocorrelation. Nevertheless, by improving the
detection algorithm, it’s possible to improve the
performance. In mmWave communication systems,
the BS receive multiple signals, making the
performance to rely on how those signals get
combined [8].
Multiple digital chains at the BS
Using multiple digital chains at the BS
reduces the required SNR, thus improving the RACH
performance. An example is simultaneously steering
in the same direction, multiple receive beams, and the
received signals become accumulated non-
coherently, leading to a performance gain.
Engaging Beam Reciprocity
The channel reciprocity for a random access
procedure can be exploited for a mmWave system
operating in a TDD mode. If the channel reciprocity
holds, the design of random access procedure can be
done for the MS to transmit a preamble only in the
best direction and acquire a high beamforming gain
[9]. It is important to note that to ensure beam
reciprocity, more so in NLOS surroundings, RF
circuits should be calibrated, since different
characteristics of RF circuitry of transmitter and
receiver may not allow channel reciprocity to hold
[10].
Cell Deployments
Owing to the fact that the path loss of an
NLOS channel is much bigger than that of a LOS
channel, one way of solving the initial challenges is
carefully deploying BSs in locations suitable for
easier formation of a LOS link between MS and BS.
For the purposes of having a large area coverage with
this constraint, however, there is need for a big
number of BSs [13]. Thus, the method of cell
planning is deemed highly significant in mmWave
cellular communication.
Such important parameters as scheduling,
initial access, uplink-downlink configuration as well
as handover have been considered when analysing
the important factors in random access channel
design. The performance gain from multi-directional
beamforming without having a prior knowledge of
the best beam pair is still achievable following
confirmation from several numerical simulations
conducted in this paper. As the study for random
access in mmWave is still in its early stages, more
research and study are expected to be conducted for
an effective mmWave cellular network development
for 5G technology [12, 14].
Low-latency communications
Future wireless communications require
low-latency to support several massive delay-
sensitive applications. For instance, the end-to-end
latency for 5G network will be expected to be in the
range of 1-5 MS, which is by far lower than the ones
for 3G and 4G LTE networks, making it rather a
challenge to attain an ultra-low latency system in
future communication technologies. With respect to
the unprecedented massive data volumes in 5G
systems, the transmitters and receivers are fitted with
large buffers [15]. However, buffered wireless
systems experience a queuing delay that significantly
affects their latency. Thus, to achieve an ultra-low
latency in buffered HetNets, it is crucial to reduce the
queuing delay.
Low-latency HetNets with Buffer
In spite of the great achievements in
lowering the end-to-end latency in HetNets with
buffers, much still need to be done. The main
challenge lies in incorporating buffers that make the
problem dissimilar to the conventional ones.
Particularly, with buffers, the end-to-end latency
depends on the queuing state at the buffer, each link’s
capacity and on the sequence of arrival [14, 16].
Therefore, the formulation of end-to-end latency for
networks with buffers cannot be ruled as simply as
traditional graph-based network optimization
challenge like a minimum cost flow challenge,
shortest path challenge or a maximum flow problem.
Without any iota of doubt, it has been noticed that
much has not been studied about problem of latency
minimization for heterogeneous networks with
buffers. Thus, this section of the paper proposes an
adaptive low-latency strategy based on cooperative
networking.
Adaptive low-latency strategy based on cooperative
networking
Adaptive low-latency strategy
In the control plane of the HetNets, the
MeNB makes decision and performs the roles of a
controller that determines scheme of networking
Heterogeneous Networks 7

based on the information collected from the UE and
the SeNBs. Consequently, the UE, the SeNBs and the
MeNB respect the decision from the MeNB so as to
enable the corresponding networking to be performed
on the data plane afterwards. This section adopts an
adaptive low-latency transmission strategy in an
heterogeneous network with mmWave system that
offers the optimal networking scheme based on the
information of acquired channel. In the adaptive
strategy there are two plane, the data plane and the
control plane [17]. The control plane performs such
functions as gathering channel information, arranging
traffic allocation, and directing the networking
scheme. The control plane only controls the control
signals.
On the other hand, data plane is used for
transmission of data, and all the operations therein
are conducted under the control signals received. The
MeNB will make a judgment, with regards to the
information in the UE, to whether the UE belongs to
the small cells. The MeNB will fulfil a downlink
transmission in the microwave band if the UE does
not fall in either of the small cells, else the MeNB
designs networking scheme, where there will be a
potential participation of the SeNBs [18]. There is an
ever increasing requirement in latency and data rates,
thus there is need to find ways of addressing this
issue and HetNets with mmWave has been deemed
able to improve significantly the capacity and
network coverage.
This paper has looked at a HetNet
comprising of two SeNBs, a UE and an
Figure 5. Diagram of adaptive low-latency strategy,
where first-in first-out (FIFO) queues are used for
the buffer-aided HetNet
MeNB, and conducted an investigation on
the strategy of low latency for the downlink
transmission to the UE from the MeNB. An adaptive
strategy has been introduced for buffered HetNets,
based on cooperative networking that greatly through
traffic allocation optimization minimizes the latency
[20]. This section has shown that proper cooperative
networking is key to minimising end-to-end latency,
thus offering an insight on management of traffic as
well as
CONCLUSION
It is optimization of network for HetNets of
the future. beyond any doubts that the application of
HetsNets with mmWave communication is essential
in realizing the future wireless 5G technology. The
vast spectral resources of the mmWave
communication radios will enable users to enjoy an
experience that will offer unprecedented services that
are similar to wired technology. This paper attempts
to study the feasibility of the proposed mmWave
communication under various scenarios, and further
propose a new technology of 3GPP LTE, which is a
backward-compliment frame structure. This enables
this proposal to come up with a technology that
answers the challenges in mmWave technology and
at the same time attain an aggregated cell throughput
of approximately 13Gb/s, a magnitude order that
even the best current 5G design cannot match.
The basic knowledge of the mmWave
technology has been further cemented by the
development of new techniques in mmWave physical
layer. To address the future of mmWave
communications, the research in HetSNets will be a
subject of motivation by tight coupling of the
characteristics unique to mmWave communications
as well as unwired heterogeneous networks. Most
particularly, a detailed study and research on
decoupling the data and control channels will be
needed, that will bring about radio resource
Heterogeneous Networks 8

management protocols, call admission control as well
as a new handover.
A presentation on an overview random
access in mmWave beamforming cellular networks
has been made. Such important parameters as
scheduling, initial access, uplink-downlink
configuration as well as handover have been
considered when analysing the important factors in
random access channel design. The performance gain
from multi-directional beamforming without having a
prior knowledge of the best beam pair is still
achievable following confirmation from several
numerical simulations conducted in this paper. As the
study for random access in mmWave is still in its
early stages, more research and study are expected to
be conducted for an effective mmWave cellular
network development for 5G technology.
This paper has looked at a HetNet
comprising of two SeNBs, a UE and an MeNB, and
conducted an investigation on the strategy of low
latency for the downlink transmission to the UE from
the MeNB. An adaptive strategy has been introduced
for buffered HetNets, based on cooperative
networking that greatly through traffic allocation
optimization minimizes the latency. This section has
shown that proper cooperative networking is key to
minimising end-to-end latency, thus offering an
insight on management of traffic as well as
optimization of network for HetNets of the future.
References
[1] “Scenarios, Requirements and KPIs for 5G
Mobile and Wireless System,” ICT-317669-
METIS/D1.1, May 2013.
[2] L. Lei et al., “Operator Controlled Device-to-
Device Communications in LTE-Advanced
Networks,” IEEE Wireless, vol. 19, no. 3, June 2012,
pp. 96–104.
[3] K. Zheng et al., “Energy-Efficient Wireless In-
Home: The Need for Interference-Controlled
Femtocells,” IEEE Wireless Commun., vol. 18, no. 6,
Dec. 2011, pp. 36–44.
[4] E. G. Larsson et al., “Massive MIMO for Next
Generation Wireless Systems,” IEEE Commun.
Mag., vol.52, no. 2, Feb. 2014, pp. 186–95.
[5] Z. Pi and F. Khan, “A Millimeter-Wave Massive
MIMO System for Next Generation Mobile
Broadband,” Proc. IEEE ASILOMAR, 2012, pp.
693–98.
[6] D. Astely et al., “LTE Release 12 and Beyond,”
IEEE Commun. Mag., vol. 51, no. 7, July 2013, pp.
154–60.
[7] S. Hur et al., “Millimeter-Wave Beamforming for
Wireless Backhaul and Access in Small Cell
Networks,” IEEE Trans. Commun., vol. 61, no. 10,
Oct. 2013, pp. 4391–03.
[8] J. Qiao et al., “MAC-Layer Concurrent
Beamforming Protocol for Indoor Millimeter
[9] Z. Pi and F. Khan, “An Introduction to
Millimeter-Wave Mobile Broadband Systems,” IEEE
Commun. Mag., vol. 49, no. 6, June 2011, pp. 101–
07.
[10] E. G. Larsson et al., “Massive MIMO for Next
Generation Wireless Systems,” IEEE Commun.
Mag., vol. 52, no. 2, Feb. 2014, pp. 186–95.
[11] W. Roh et al., “Millimeter-Wave Beamforming
as an Enabling Technology for 5G Cellular
Communications: Theoretical Feasibility and
Prototype Results,” IEEE Commun. Mag., vol. 52,
no. 2, Feb. 2014, pp. 106–13.
[12] T. S. Rappaport et al., “Millimeter Wave Mobile
Communications for 5G Cellular: It Will Work!,”
IEEE Access, vol. 1, May 2013, pp. 335–49.
[13] T. S. Rappaport et al., “Broadband Millimeter-
Wave Propagation Measurements and Models Using
Adaptive-Beam Antennas for Outdoor Urban
Cellular Communications,” IEEE Trans. Antennas
Propag., vol. 61, no. 4, Apr. 2013, pp. 1850–59.
[14] S. Sesia, I. Toufik, and M. Baker, LTE – The
UMTS Long Term Evolution: From Theory to
Practice, 2nd ed., Wiley, 2011.
[15] ITU-R M.2135-1, “Guidelines for Evaluation of
Radio Interface Technologies for IMT-Advanced,”
2009.
[16] K. Zheng et al., “10 Gb/s HetNets with
Millimeter-Wave Communications: Access and
Networking Challenges and Protocols,” IEEE
Commun. Mag., vol. 53, no. 1, Jan. 2015, pp. 222–
31.
[17] T. S. Rappaport et al., “Millimeter Wave Mobile
Communications for 5G Cellular: It Will Work!”
IEEE Access, vol. 1, 2013, pp. 335–49.
[18] M. Xiao et al., “Millimeter Wave
Communications for Future Mobile Networks,” IEEE
JSAC, vol. 35, no. 9, Sept. 2017, pp. 1909–35.
[19] S. Mumtaz et al., “Terahertz Communication for
Vehicular Networks,” IEEE Trans. Vehic. Tech., vol.
66, no. 7, 2017, pp. 5617–25.
Heterogeneous Networks 9

[20] A. J. Morgado et al., “Hybrid Resource
Allocation for Milli-meter-Wave NOMA, ” IEEE
Wireless Commun., vol. 24, no.
Heterogeneous Networks 10