White paper

Designing, Operating
and Optimizing
Unifed Heterogeneous Networks

Contents
2. Executive summary
3. The pressures of booming
mobile broadband use
4. Heterogeneous network
deployment and
expansion roadmaps
6. Heterogeneous network
optimization
6. Traffc steering
and mobility
management
8. nterference
management
10. Energy saving
11. Heterogeneous network
confguration
12. Heterogeneous network
fault management
13. The future of networks
14. Meeting the challenge
2 Unified Heterogeneous Networks
of heterogeneous networks and
explains how Nokia Siemens
Networks can help communications
service providers (CSPs) address
them. t discusses how to design
roadmaps for expansion and how to
optimize systems to manage traffc
steering and mobility, interference and
energy saving. t also explains how
to automate confguration and fault
management in order to keep OPEX
under control. t's clear that a unifed
approach to managing heterogeneous
networks is essential to achieve
the necessary fexibility and cost-
effciency while delivering the kind of
seamless broadband connectivity that
consumers increasingly demand.
Executive summary
Growing demand for low-cost mobile
broadband connectivity is driving
the development of heterogeneous
cellular networks. A range of different
radio access technologies (RATs)
and WiFi will all co-exist, and macro
cells will be complemented by a
multitude of smaller cells, such
as micro, pico and femto cells.
Such heterogeneous systems will
be signifcantly more complex to
manage than today's networks and
therefore require fully Self Organizing
Networks (SON).
This white paper gives an overview
of the key challenges inherent in the
design, operation and optimization
Figure 1. Heterogeneous networks combine a range of radio access technologies.
Liquid Radio
Wide area Medium area Hot spots Indoor
LTE TD-LTE HSPA+
Femto
GSM
LTE
Femto
HSPA+
WiFi
Wide area Medium area Hot spots Indoor
HSPA /
LTE
HSPA+ /
LTE GSM
HSPA+
Femto
GSM /
HSPA
LTE
Femto WiFi
A combination of exponentially
growing demand for mobile data
and fat-lining average revenue per
user (ARPU) makes it tough for
CSPs looking to evolve their mobile
communications networks. The
rules of the game are changing, as
mobile indoor capacity becomes
more valuable and machine-to-
machine (M2M) communications
take off and create demand for more
diverse services. At the same time,
factors such as energy effciency are
emerging as crucial KPs.
ncreasing traffc, novel services, cost
and energy considerations will all lead
to signifcantly more diverse cellular
networks in the future:
Multi-RAT (sometimes the result
of legacy infrastructure, catering
for multiple cellular standards and
fxed-mobile convergence).
Multi-layered (with macro,
micro, pico and femto cells
of different size).
Multi-vendor (where different
access points have been provided
by different vendors).
The pressures of booming
mobile broadband use
3 Unified Heterogeneous Networks
This means that the design, operation
and optimization of future networks
present novel challenges and require
new methodologies. While a lot of
hype has already been generated
around specifc topics such as
interference management or SON, the
aim of this white paper is to look at the
big picture regarding heterogeneous
systems in general. There are two
key questions:
What will drive particular
heterogeneous deployments
and how can a CSP determine
which is the right roadmap for
expansion?
What are the key challenges
inherent in operating
and optimizing complex
heterogeneous systems and how
can CSPs address them?
MuIti-RAT

The reason for multi-RAT
deployments is simple. Typically,
CSPs already have wide-area
GSM coverage and WCDMA/
HSPA in densely populated urban
areas. They're probably deploying
LTE in hotspots or (in, e.g., the
case of Germany) in rural areas in
order to exploit the digital dividend.
They may also consider re-farming
existing GSM frequency bands to
WCDMA/HSPA or LTE, so they can
update their equipment gradually
to more spectrally effcient radio
standards. t's likely that GSM,
WCDMA/HSPA and LTE will
continue to coexist and evolve in the
long term for several reasons:
GSM may be the only system
providing ubiquitous voice
coverage and may be used
for M2M.
nvestments in HSPA may not yet
be amortized.
Long-term spectrum licenses
may require CSPs to use a
particular RAT.
A large population of legacy
terminals can force CSPs to
keep legacy air interfaces
running.
The cellular standards already
mentioned will also coexist with
technologies such as WiFi. n
fact, offoading data traffc from
cellular air interfaces to WiFi is
highly attractive for CSPs from a
cost point of view, allowing them to
reduce traffc in 3G/4G networks
and use comparatively inexpensive
backhaul infrastructure. A mobile
CSP that also owns the WiFi
access infrastructure can deliver a
seamless data experience for end
users.
MuIti-Iayer
Many networks will include an
overlay of cells of different sizes
using a single RAT. For instance,
outdoor terminals may be served by
a combination of micro and macro
cells. Pico cells may provide both
outdoor and indoor coverage in
hotspots such as train stations or
shopping malls with a typical cell
radius of up to 200 meters. Femto
cells are used indoors in cells of no
more than 10-25m radius. While
pico cells are deployed by a CSP,
femto cells are typically user-
deployed in the form of private home
base stations, often denoted as
home (e)NBs or H(e)NBs. There's
also a distinction between open and
closed subscriber group (OSG/CSG)
femto cells, where CSG cells serve
a constrained set of users. Figure 2
shows a typical multi-layer network.
The trend towards multi-layer
deployments is driven by the need
to provide better indoor services and
respond to heterogeneous traffc
demands, as well as by cost and
energy effciency considerations.
4 Unified Heterogeneous Networks
Heterogeneous network
deployment and
expansion roadmaps
Liquid Radio
Indoors
Femto cell
Cell radius
> 300m
100-300m
< 200m
10-25m
Macro cell
Micro cell
Pico cell
Femto cell
Tx power
46 dBm
40 dBm
> 24 dBm
< 20 dBm
Indoors
Femto cell
Indoors
Femto cell
Cell radius
> 300m
100-300m
< 200m
10-25m
Macro cell
Micro cell
Pico cell
Femto cell
Tx power
46 dBm
40 dBm
> 24 dBm
< 20 dBm
Indoors
Femto cell
Figure 2. llustration of a typical multi-layer setup.
5 Unified Heterogeneous Networks
Relaying is an interesting option
in cases where a wired backhaul
connection between the base station
and the network is uneconomical or
technically unfeasible, and where an
in-band or out-band backhaul over-
the-air to the donor cell is preferable.
Terminals typically experience relay
cells as individual layers in the system.
How can CSPs determine the
right expansion roadmap?
An optimal network expansion
roadmap depends on various CSP-
and location-specifc parameters and
assumptions, such as:
existing legacy infrastructure in
terms of sites, base stations and
backhaul.
the availability of spectrum and
terminals for specifc RATs.
traffc demand, user mobility
and revenue forecast for a
particular area.
cost-related aspects (such as
backhaul infrastructure cost, site
rental, labor and energy).
general strategic decisions
regarding services to be provided
and the metric to be optimized
(such as ubiquitous connectivity
anytime and for anybody vs. peak
data rates for certain consumers).
Establishing an expansion roadmap
requires sophisticated performance
evaluation methodology, detailed
cost models and measurement
data. The impact of the uncertainty
inherent in parameters such as
traffc forecasts can be mitigated by
investing in fexible base stations,
where changes can be made later
on via a software upgrade.
2100 MHz
2100 MHz
Handovers
between
LTE and HSPA
2100 MHz
10x
Global Mobile Traffic Forecast
A
n
n
u
a
l

t
r
a
f
f
i
c

(
P
B
)
25,000
20,000
15,000
10,000
5,000
50,000
45,000
40,000
35,000
30,000
0
2010 2011 2012 2013 2014 2015
Mobile voice (16kbps) Mobile tablet Mobile laptop Mobile handheld
Liquid Radio
Mobile Internet Traffic (ExaByte/year)
25
20
15
10
5
0
2010 2011 2012 2013 2014 2015
WCDMA/HSPA
LTE
Mobile Laptop
Upgrade to 6-sector
Upgrade to the 2
nd
carrier
Add HSPA macro sites Add HSPA micro cells
Upgrade to 3
rd
carrier Existing macro sites Upgrade to 2
nd
carrier
Upgrade to 6-sector
Add LTE micro cells
(new or reused HSPA micro sites)
New LTE RAT
at existing HSPA sites

The take-up time of
LTE strongly depends
on spectrum and LTE
terminal availability
Figure 3. An expansion roadmap for an urban scenario with strongly increasing traffc.
Figure 3 shows an example of an
expansion roadmap.
The traffc distribution can vary
widely throughout a given network.
This, combined with the practical
deployment limitations of different
upgrade options, means that CSPs
may pursue several expansion paths
simultaneously in one network.
CSPs need an automated process
to identify which parts of the network
need which upgrade. n the long run,
many CSPs will also be managing
networks in which equipment from
different vendors is used in the same
geographical area. n this case, it is
particularly important that all network
management functions are multi-
vendor-capable.
Operating and optimizing complex
heterogeneous systems presents
several key challenges, such as
how to distribute traffc effciently
between cells, RATs and layers
while guaranteeing seamless user
mobility, how to alleviate the impact
of interference and how to adapt the
system effciently to meet changing
traffc demand.
Trafc steering and
mobiIity management
Traffc steering allows CSPs to
optimize their resources, improve
the way users experience services
and minimize power consumption
by directing the traffc to a particular
RAT or layer. Traffc steering is a
tool for reducing OPEX and limiting
or postponing CAPEX, especially in
complex heterogeneous systems.
t works hand-in-hand with mobility
management to ensure a reasonable
number of handovers and avoid radio
link failures. t also needs to consider
other factors such as the capabilities
of the terminals and network, the
delivery of services and quality
of service (QoS), the load in
different RATs and layers and
power consumption.
Trafc steering in
different standards
Both WCDMA/HSPA and LTE can
perform traffc steering in various
ways. These are typically divided
into those that apply to terminals in
the idle state and those that deal
with the connected state. n the idle
state, the terminals are responsible
for measuring their environment and
deciding which RAT or layer to camp
on. The network can infuence these
decisions by applying various means:
Basic biasing involves the
provisioning of cell reselection
hysteresis, (the signal power
thresholds used to prompt cell
reselection) and cell quality offsets
to terminals.
Hierarchical cell structure
in WCDMA/HSPA enables
prioritization for cell reselection.
Absolute priorities can be
assigned for intra- or inter-RAT
carriers to terminals.
Cell barring forbids terminals from
camping on any cells that are
included on a black list.
n a connected state, the network
is responsible for triggering a hand-
over between cells, RATs or layers.
Hand-over decisions are usually
based on terminal measurements, but
WCDMA/HSPA and LTE networks
have various options to infuence the
scope of these measurements. The
challenge is to balance the need to let
terminals take enough measurements
to ensure maximum offoad and
smooth mobility without taking so
many that the measurements drain
the terminal batteries and degrade
network performance. n general,
all the traffc steering and mobility
management concepts used for idle
and connected state terminals have
to be well-aligned. Further, mobility
robustness optimization (MRO) is
required to identify the root cause of
any radio link failures and call drops,
and to adjust traffc steering and
mobility management parameters as
a result.
Femto cells present special
challenges, since they may be
deployed in large numbers and are
typically outside the CSP's control.
A user can move or turn off a femto
base station at any time, for example.
Mobility is also diffcult in the context
of closed subscriber group (CSG)
femtos, and there may be confusion
over cell identities in very dense
deployments owing to limited sets
of scrambling codes or physical cell
identifers. While both WCDMA/HSPA
and LTE now provide various options
to indicate femto proximity, CSG
identifcation and disambiguation,
mobility in dense deployments
still calls for sophisticated
proprietary solutions.
Carrier aggregation techniques may
make it possible to schedule traffc
fexibly on multiple (co-located)
layers using simple radio resource
management rather than hand-overs.
HSPA-LTE carrier aggregation will
potentially enable a similar approach
for multi-RAT traffc steering.
Offoading as much traffc as possible
to femto or WiFi cells and balancing
the remaining load between the
remaining layers and RATs helps
CSPs to use their infrastructure
effciently and offer a homogeneous
QoS for end users. Table 1
summarizes specifc solutions for
certain scenarios.
The outIook for trafc steering
and mobiIity management
There are various ways of achieving
traffc steering and mobility
management in WCDMA/HSPA
and LTE, but they all depend on
networks performing well in two
fundamental areas:
Networks must use the available
resources effciently and adjust
biases and priorities appropriately.
They must also provide terminals
with up-to-date cell lists and
black lists.
Networks should optimize the
frequency at which terminals
perform measurements.
Clearly, both of these issues
require base stations to monitor
their neighborhood. More than this,
however, CSPs looking for end-to-
end traffc steering need a unifed
heterogeneous network management
solution that spans all RATs and
layers, as well as the backhaul
infrastructure and core network.
Future standardization work
will focus on functions such as
introducing speed-dependent black
list confgurations, enhancing the
proximity indication procedure
for CSG cells and adding cell
type indicators to the mobility
state defnition.
6 Unified Heterogeneous Networks
Heterogeneous network
optimization
7 Unified Heterogeneous Networks
Scenario / use case Recommended idIe mode soIution Recommended connected mode soIution

Macro-micro / macro-pico depIoyment Use a scenario-specifc extent of basic biasing Apply cell-specifc offsets to terminal
to perform small cell range extensions and measurements to enlarge small cell
promote macro offoad. hand-over areas.

Macro-femto depIoyment n WCDMA/UMTS: Primary scrambling code Use of proximity indications to promote
(CSG-aware terminals) (PSC) ranges connected to femtos should be hand-over to femto cells with minimal
broadcast to terminals. terminal measurement.

n LTE: Physical cell identifer (PC) ranges of
CSG cells to be broadcasted to terminals.


Macro-femto depIoyment Adjust biasing concepts and neighbor cell lists to incentivize terminals to or prevent
(Legacy WCDMA/UMTS terminals) them from measuring and accessing CSG femto cells.

Inter-RAT trafc steering between Use absolute priorities to enforce Not applicable if all LTE-capable terminals
WCDMA/UMTS and LTE offoad to LTE. are kept in LTE.
(Low LTE terminal penetration)

Inter-RAT trafc steering between Adjust absolute priorities according to load, n WCDMA/UMTS: Adjust neighbor cell lists
WCDMA/UMTS and LTE ideally use dedicated absolute priorities (NCLs) according to load.
(Medium/high LTE terminal penetration) (sent to terminals after connection release).
n LTE: Adjust black lists according to load.

WiFi ofoad End users want a seamless data service experience in which their device is always effortlessly
connected to either cellular network or WiFi access point. Device management and automated
network discovery functions allow CSPs to manage this aspect of traffc steering effciently, and
logically integrate WiFi networks into their core network.

Fast-moving terminaIs n WCDMA/UMTS: Terminals determine speed n WCDMA/UMTS: Network shall
(i.e. preventing such terminals themselves, switch to "high mobility state and estimate terminal speed based on
from using small cells) refrain from using small cells via hierarchical previous hand-overs, then provide
cell structures. neighbor cell lists (NCLs) containing
only macro cells to high-speed terminals.
n LTE: Terminals determine speed
themselves and obtain scaling factors n LTE: Network shall estimate terminal
infuencing cell reselection. speed via history information, restrict
inter-frequency measurements and
provide modifed black lists accordingly.
Table 1: Recommended traffc steering and mobility management solutions.
8 Unified Heterogeneous Networks
Interference management
nter-cell interference is already one
of the limiting factors in today's mobile
communications systems, especially
in dense, urban deployments. The
problem is even worse in the context
of multi-layer networks, as illustrated
in Figure 4.
f both the macro cell and the
smaller cell are using the same
radio resources (so-called co-
channel deployment), the following
interference problems can occur:
n the downIink, a terminal
assigned to the macro base station
may see strong interference
coming from a small cell, leading to
a so-called macro layer coverage-
hole. This problem is particularly
pronounced if the smaller cell
serves a CSG, in which case a
terminal may be very close to
a small cell but not allowed to
connect to it (case A). On the other
hand, a terminal served by a small
cell may see strong interference
from a macro cell, in particular if
a cell range extension is used to
enforce offoad (case B).
n the upIink, a terminal assigned
to the macro cell but close
to the cell-edge will typically
create strong interference to the
small cell (case C). However,
this degradation of small cell
performance is often acceptable,
since each terminal connected
to a small cell typically accesses
a much larger share of radio
resources. A more problematic
aspect is the uplink interference
that a potentially large number of
small cell terminals may generate
towards one macro cell (case D).
nterference can be attenuated or
increased if the cell border is shifted
towards the larger or smaller cell, for
example, by traffc steering or mobility
management.
How can CSPs mitigate
interference probIems?
Antenna downtilt is a common
approach to avoid interference in a
homogeneous macro deployment,
where it can be achieved fexibly
using smart antenna concepts. n a
multi-layer context, however, downtilt
optimization is rarely applicable, since
CSPs need to ensure macro umbrella
coverage and so tend to leave macro
cell downtilt untouched. n addition,
the antennas used for micro, pico and
femto cells are often mounted at a low
level, so that downtilt is not suitable.
Smart resource reuse is required
when interference cannot be avoided
by physical means. t may be better
to split interfering entities onto
orthogonal resources (divided by time
or frequency) and abandon the idea
of fully using the resources in each
cell. An optimal scheme would let
adjacent cells cooperatively decide
upon resource usage, requiring
complex signaling between cells. A
more pragmatic approach is to employ
static resource reuse concepts:
Reserve some resources for
macro-only, small-cell-only or
constrained usage - so-called
fractional frequency reuse. Escape
carrier concepts are a good
example, where dedicated carriers
are reserved for macro usage
only. LTE Release 10 includes a
feature that allows recurring time
slots to be reserved for some
layers, referred to as enhanced
inter-cell interference coordination
(eCC). eCC only works if base
stations can be synchronized and
terminals have good measurement
capabilities.
Certain layers may be allowed
to access some resources only
with a reduced transmission
power. This is called soft frequency
reuse.
Power control parameters can be
adjusted to either defne a power
offset to be used by all the elements
in a particular layer and/or to apply
power capping. This is an effective
way to trade the performance of some
layers against others, for example,
by improving macro cell-edge
performance at the price of small cell
performance.
2100 MHz
2100 MHz
Handovers
between
LTE and HSPA
2100 MHz
10x
Liquid Radio
Mobile Internet Traffic (ExaByte/year)
25
20
15
10
5
0
2010 2011 2012 2013 2014 2015
Mobile Handheld
Mobile Laptop
B: Femto terminals seeing
DL macro interference,
esp. under cell range
extension
B: Femto terminals seeing
DL macro interference,
esp. under cell range
extension
D: Many small cell
terminals creating uplink
interference to macro cell
C: Macro terminal creating
strong uplink
interference to small cell
A: Macro terminal
seeing strong
downlink interference
from small cell
Pico cell
Many pico cells
Macro cell
Figure 4. Major interference problems in multi-layer confgurations.
9 Unified Heterogeneous Networks
Optimized interference
management soIutions
A combination of smart resource
reuse and power control generally
provides the best solution.
Table 2 summarizes this best use of
this approach in different scenarios.
Figure 5 shows how a particular power
control adaptation and escape carrier
concept can trade femto for macro
performance in the downlink.
The outIook for
interference management
CSPs must take inter-layer
interference into consideration in
the operation and optimization of
heterogeneous networks. While
both WCDMA/UMTS and LTE
provide the basic functionalities
to address interference issues,
effcient interference management
requires a unifed heterogeneous
network solution that includes the
entire network setup. t must also be
well-aligned with traffc steering and
mobility management.
Several improvements are under
consideration for LTE Release 11,
including enhanced terminal receivers,
improved eCC and autonomous
carrier-based interference
management. Autonomous
component carrier selection (ACCS)
Scenario / use case Recommended resource usage Recommended power controI adaptation

Macro-femto deployment Resource reuse (co-channel), but with one Downlink: Femto power calibration based on
carrier free of femto CSG cells if possible. network listener mode (NLM), where femtos
sense the level of interference around them.

Uplink: Power capping of WCDMA/UMTS
femto terminals and customized power
control setting for LTE femto terminals.

Macro-micro or macro-pico deployment Dedicated micro or pico carrier is preferred, Downlink: No power control
but co-channel operation is also feasible. n adaptation needed.
the latter case, eCC should be used in the
context of small cell range extension. Uplink: Power capping of WCDMA/
UMTS small cell terminals and customized
power control setting for LTE small cell
terminals for co-channel deployment.
Table 2: Recommended interference management schemes for multi-layer HSPA/LTE.
2100 MHz
2100 MHz
Handovers
between
LTE and HSPA
2100 MHz
10x
Median femto user throughput (Mbps)
5
t
h

p
e
r
c
e
n
t
i
l
e

m
a
c
r
o

u
s
e
r
t
h
r
o
u
g
h
p
u
t

(
M
b
p
s
)
0.2
0.1
0
0
0.4
0.5
0.3
10 20 30 40
Mobile voice (16kbps) Mobile tablet Mobile laptop Mobile handheld
Liquid Radio
Mobile Internet Traffic (ExaByte/year)
25
20
15
10
5
0
2010 2011 2012 2013 2014 2015
M Handheld
Mobile Laptop
Escape carrier +
power control
Escape
carrier
Co-channel +
power control
Co-channel
Low femto density
High femto density
is being discussed, in which base
stations sense the use of certain
carriers in their neighborhood and
adjust their resource consumption
accordingly. Coordinated multi-
point (CoMP) techniques are also
being discussed for heterogeneous
networks. These allow network
operators to exploit interference as
useful signal energy, rather than
treating it as a burden.

Figure 5. The impact of power control adaptation and escape carrier concepts on downlink
performance in LTE macro-femto scenarios (CSG case).
10 Unified Heterogeneous Networks
Energy saving
Energy effciency is increasingly
important in terms of reducing both
CO
2
emissions and costs. Since the
base stations consume the lion's
share of energy in a typical network,
effciency is particularly crucial in
dense heterogeneous systems.
Energy saving through turning
off base stations
While replacing old base stations
with more power-effcient single-RAN
equipment is the most intuitive option,
major savings can also result from
enabling systems to turn off access
points when they're not needed. n
a homogeneous network of cells,
all but a certain pattern of cells
might be turned off, reducing the
cell density and increasing the size
of the remaining cells. n multi-layer
deployments, CSPs may switch off
the smaller layer of cells in off-peak
situations, so that the larger cells
can take over without changing the
coverage area.
While it is straightforward to
automatically turn off an access point
when it experiences zero load for
some time, deciding when to turn it
on again could be trickier. There are
several potential solutions:
Turn base stations on or off
based on a predefned schedule
generated from historical traffc
statistics.
Periodically switch on all hotspots
and then switch off those that
experience low load.
Reactivate hotspots when a certain
oT (nterference over Thermal
Noise) threshold is exceeded, since
this indicates that nearby terminals
are transmitting to an adjacent cell.
An optimized approach
to energy saving
Looking at today's multi-RAT and
emerging multi-layer networks, a
centralized operations, administration
and maintenance (OAM)-based
solution is the most promising energy
saving solution in the short and
medium term. However, this may be
problematic in multi-vendor networks
and in large networks where a
distributed approach could react faster
to changing traffc conditions. Hence,
a fully standardized solution with local
decisions taken by base stations and
exchanged between them looks more
promising in the long term.
11 Unified Heterogeneous Networks
Heterogeneous network
confguration
Effcient operation of heterogeneous
networks requires each network
element to be well defned in terms of
its particular role and location within
the network, and to be context-aware.
Clearly, this should be automated as
far as possible to control OPEX.
Auto-connectivity and
auto-commissioning
Automatic provisioning and
confguration of new network elements
reduces the time and effort involved
in getting newly installed hardware
up and running. One simple solution
is to use RFD site tags or GPS
information to automatically keep track
of equipment deployed at different
sites and then obtain the required
confguration information from the
OAM system, as shown in Figure 6.
An alternative approach for automatic
relay-node confguration is for a relay
to connect itself initially to an arbitrary
base station as if it were a terminal in
order to access the confguration
data. t then sets up the feeder
link to its donor base station and
begins relaying.
Automatic neighbor reIations
After confguration, a network
element has to obtain and update
information on its environment in
order to adjust traffc steering, mobility
management and interference
management settings. n particular,
it needs up-to-date knowledge of
neighboring cells, since missing or
newly added neighbors can result
in poor performance and call drops.
Automatic neighbor relations (ANR)
depend on terminals to detect and
report neighboring cells, providing a
very effcient and "always-on way
of keeping base stations updated,
even if radio conditions change. n the
context of heterogeneous networks,
WCDMA/HSPA and LTE use different
techniques to let terminals discover
different layers and RATs. The amount
of information must be balanced
carefully against the effort of making
such measurements.
Optimized ANR soIutions
While terminal-based environment
discovery is certainly a very fexible
and self-adapting mechanism in a
changing network topology, it may be
several years to establish a suffcient
number of terminals capable of
supporting ANR among the terminal
population. Solutions based on
network management systems are
therefore still an attractive option. n
this case neighbor cells' identifer and
related P addresses are provided by
the OAM system. Hybrid architectures
can exploit both the information
collected by terminals, which is
specifcally important in the context
of user-deployed femto cells, and
information from the OAM system.
2100 MHz
2100 MHz
Handovers
between
LTE and HSPA
2100 MHz
10x
Global Mobile Traffic Forecast
A
n
n
u
a
l

t
r
a
f
f
i
c

(
P
B
)
25,000
20,000
15,000
10,000
5,000
50,000
45,000
40,000
35,000
30,000
0
2010 2011 2012 2013 2014 2015
Mobile voice (16kbps) Mobile tablet Mobile laptop Mobile handheld
Liquid Radio
Mobile Internet Traffic (ExaByte/year)
25
20
15
10
5
0
2010 2011 2012 2013 2014 2015
Mobile Handheld
Mobile Laptop
Explosion of
traffc volume
Network element
HW-ID
HW-ID
Config Data
Location / site info
OAM system
Database
Location
Site info
(obtained through RFID)
or
Location / site info
Config Data
HW-ID
Figure 6. Automatic site identifcation and hardware-to-site-mapping.
12 Unified Heterogeneous Networks
A lot of man hours are already
invested in fault detection, diagnosis
and compensation in today's
networks. This effort will become
prohibitive in heterogeneous
networks, to a point where the high
number of network elements cannot
be managed in a conventional way.
Automatic cell degradation detection,
diagnosis and healing functionality will
be crucial for CSPs looking to control
their OPEX. While the existence of
multiple layers and RATs offers some
redundancy to alleviate the need
for automatic healing, the existence
of user-deployed femto nodes will
require novel fault management.
Faults generally fall into one of
two categories:
Obvious faults in hardware and
software are easy to detect, since
the equipment typically initiates an
alarm.
Other faults are more diffcult
to detect, such as RF failures
(antenna direction and connectivity
issues, power amplifer
degradation and so on), scheduling
problems or persistent hand-over
failures because of the wrong
parameter settings. "Sleeping
cells are especially problematic.
These cells are not functioning at
all and accept no traffc, but create
no alarm.
The frst step in fault management
is to differentiate between normal
and abnormal system behavior. Any
"symptoms of abnormal behavior are
linked to a set of potential causes,
in most cases based on empirical
knowledge supplied to the system
by human operators. n order to
locate specifc areas affected by a
problem - such as interference around
a new, user-placed base station, or
hand-over problems in a certain area
- CSPs should consider terminal-
centric measurement data, including
timestamps and location information.
This enables them to focus on the
problem from the terminal perspective.
What happens when
a faiIure is detected?
Often, the most intuitive response to a
failure will be to reset the affected cell
and wait for it to start working again.
f this does not resolve the problem,
further action may be taken:
f a pico or femto cell fails, the
macro layer of the same RAT or a
different RAT may compensate. f
the problem persists, the neighbor
relationships of adjacent cells will
have to be adapted.
f a macro cell fails, smaller cells
may initially compensate for the
reduced coverage, while in the
long term so-called cell outage
compensation (COC) may be used.
This adjusts the transmit power or
downtilt of adjacent macro cells to
re-establish coverage.
f the behavior of a small cell
degrades the performance of a
macro cell, the long-term solution
might be to turn the small cell off.
The outIook for fauIt management
n summary, the key to fault
management in heterogeneous
networks is to correlate the detection
and confguration management
information from across the different
layers and RATs, diagnose the
problem and decide on a solution.
Often the quality of the decision is
more important than the quality of the
actual solution itself. A well-designed
and aligned fault management
function can signifcantly reduce the
extent of human troubleshooting
as well as the time required until a
degradation is rectifed, leading to
potentially vast OPEX savings.
Heterogeneous network
fault management
13 Unified Heterogeneous Networks
The future
of networks
The arrival of coordinated multi-point
(CoMP) techniques will add a new
dimension to heterogeneous networks.
CoMP techniques may be based
on baseband pooling and hotelling,
in which the signal processing for
multiple cells is performed in a central
location, or even virtualized in a
cloud RAN.
New cognitive concepts will add yet
another dimension of fexible and
optimized spectrum utilization to the
game. These include ideas such
as femto frequency and/or network
sharing, in which multiple CSPs make
use of the same femto access points,
share parts of their core network or
share the same spectrum bands in a
fexible way.
The introduction of smart sensing,
geo-location and database information
sharing will deliver more agile use
of resources and increased overall
spectrum utilization, both of which will
help to drive down the total cost
of ownership.
All these developments will increase
the complexity of operating and
optimizing future cellular networks
even further. This makes the need
for truly unifed heterogeneous
network solutions even greater.
14 Unified Heterogeneous Networks
Mobile CSPs will undoubtedly be
investing in heterogeneous networks
over the next few years, complete
with multiple RATs, multiple layers
and multiple vendors. Traffc growth,
ubiquitous demand for high-quality
services, cost and energy effciency
considerations are all driving them
that way. While this general trend
is undeniable, the precise choice of
technologies and the corresponding
expansion roadmap that will best suit
each CSP's situation are less obvious.
Dense heterogeneous systems
clearly present novel challenges,
thanks to the high number of base
stations (many of which may be user-
deployed) and the resulting complex
interdependencies. CSPs will meet
these challenges with a co-ordinated
set of modular solutions:
Plug'n'play. t is essential for OPEX
reduction that all base stations are
self-confgurable.
Neighborhood-aware base
stations. Network elements must
measure and share information
about their neighborhood. This is
best done through a hybrid solution
based on a network management
system in conjunction with ANR.
Smart and robust end-to-end traffc
steering. Networks must rely on a
range of context-aware, effcient
and aligned traffc steering and
mobility management strategies.
Minimization of terminal-based
carrier and cell measurements.
Such measurements must be
minimized to prevent them from
sucking the life out of batteries and
impairing performance.
Smart resource reuse and
advanced power control are
essential to alleviate interference.
Smart energy management adapts
the active cell density to traffc
demand.
Automatic cell degradation
detection and a network
management system enable an
effective response to failures.
Nokia Siemens Networks supports
CSPs as they wrestle with the
increasing complexities of evolving
networks. We build smart and unifed
heterogeneous networks that enable
all network RATs and layers to be
viewed as a logically unifed network
with automated management via
the award winning Nokia Siemens
Networks SON Suite, seamless
interworking and an uncompromising
quality of experience for end users -
even in a multi-vendor environment.
n other words, we provide a unifed
approach that enables CSPs to
serve the growing demand for mobile
data while keeping costs frmly
under control.
2100 MHz
2100 MHz
Handovers
between
LTE and HSPA
2100 MHz
10x
Global Mobile Traffic Forecast
A
n
n
u
a
l

t
r
a
f
f
i
c

(
P
B
)
25,000
20,000
15,000
10,000
5,000
50,000
45,000
40,000
35,000
30,000
0
2010 2011 2012 2013 2014 2015
Mobile voice (16kbps) Mobile tablet Mobile laptop Mobile handheld
Liquid Radio
Mobile Internet Traffic (ExaByte/year)
25
20
15
10
5
0
2010 2011 2012 2013 2014 2015
Mobile Handheld
Unified Heterogeneous Network Approach
Mobile Laptop
WCDMA/HSPA
LTE
Upgrade to 6-sector
Upgrade to the 2nd carrier
Add HSPA macro sites Add HSPA micro cells
Upgrade to 3rd carrier Existing macro sites Upgrade to 2nd carrier
Upgrade to 6-sector
Add LTE micro cells
(new or reused HSPA micro sites)
New LTE RAT
at existing HSPA sites

Deployment
Configuration
Auto-Connectivity/
Commissioning
Automatic Neighbour
Relations
Optimization
Traffic Steering &
Mobility Management
Interference
Management
Energy Saving
Fault Management
Plug n play
Smart E2E traffic steering
Automatic cell
degradation detection
Minimization of
UE-based measurement
Smart power control
and resource reuse
Smart energy
management
Neighbourhood-
aware base stations
Figure 7. Unifed heterogeneous network key building blocks.
Meeting the challenge
15 Unified Heterogeneous Networks
Abbreviations
ACCS Autonomous Component Carrier Selection
ANR Automatic Neighbor Relations
ARPU Average Revenue Per User
CAPEX Capital Expenditure
COC Cell Outage Compensation
CoMP Coordinated Multi-Point
CSG Closed Subscriber Group
CSP Communications Service Provider
eCC Enhanced nter-Cell nterference Coordination
(e)NB Enhanced Node B
GSM Global System for Mobile Communications
H(e)NB Home Enhanced Node B
HSPA High Speed Packet Access
LTE Long-Term Evolution
M2M Machine-to-Machine
MRO Mobility Robustness Optimization
NCL Neighbor Cell List
NLM Network Listener Mode
OAM Operations, Administration and Maintenance
OPEX Operating Expenditure
PC Physical Cell dentifer
PSC Primary Scrambling Code
QoS Quality of Service
RAT Radio Access Technology
SON Self-Organizing Network
UMTS Universal Mobile Telecommunications System
WCDMA Wideband Code Division Multiple Access
Nokia Siemens Networks Corporation
P.O. Box.1
F-020022 NOKA SEMENS NETWORKS
Finland
Visiting address
Karaportti 3, ESPOO, Finland
Switchboard +358 71 400 4000
Product code: C401-00727-WP-201107-1-EN
Copyright 2011 Nokia Siemens Networks.
AII rights reserved.
A license is hereby granted to download and print a
copy of this document for personal use only. No other
license to any other intellectual property rights is granted
herein. Unless expressly permitted herein, reproduction,
transfer, distribution or storage of part or all of the
contents in any form without the prior written permission
of Nokia Siemens Networks is prohibited.
The content of this document is provided "AS S,
without warranties of any kind with regards its accuracy
or reliability, and specifcally excluding all implied
warranties, for example of merchantability, ftness for
purpose, title and non-infringement. n no event shall
Nokia Siemens Networks be liable for any special,
indirect or consequential damages, or any damages
whatsoever resulting form loss of use, data or profts,
arising out of or in connection with the use of the
document. Nokia Siemens Networks reserves the right
to revise the document or withdraw it at any time without
prior notice.
Nokia is a registered trademark of Nokia Corporation,
Siemens is a registered trademark of Siemens AG.
The wave logo is a trademark of Nokia Siemens
Networks Oy. Other company and product names
mentioned in this document may be trademarks of
their respective owners, and they are mentioned for
identifcation purposes only.
www.nokiasiemensnetworks.com