Planning Non Line-of-Sight Wireless Backhaul Networks

Planning Non-Line-of-Sight
Wireless Backhaul Networks

WHITEPAPER

Table of Contents
Introduction .................................................................................................................................................. 3
The Benefits of Small Cell Base Stations ....................................................................................................... 3
Overview of Cost of Backhaul Systems ......................................................................................................... 5
NLOS Wireless Backhaul Deployment Scenario ............................................................................................ 6
System Gain and Path Loss ........................................................................................................................... 7
Capacity, Coverage & Link Reliability ............................................................................................................ 8
Network Level Simulation & Planning .......................................................................................................... 8
Tools for NLOS Small Cell Backhaul Planning ................................................................................................ 9
Digital Geographical Data ......................................................................................................................... 9
Path-loss Matrices................................................................................................................................... 10
Comparing NLOS to LOS System Availability in Urban Areas...................................................................... 10
Conclusion ................................................................................................................................................... 13

Planning NLOS Wireless Backhaul Networks 2

Introduction
Small Cell base stations (SCBS) are widely recognized as a solution to the insatiable demand for mobile
data traffic which has been about doubling every year. Mobile traffic is expected to reach 10.8 exabytes
per month in 2016 from 0.6 exabytes per month in 20111. The macro-cellular architecture where base
station antennas are mounted on building rooftops has served the purpose of providing mobile voice
service very well. However, the advent of mobile data services has exposed weaknesses in the macro-
cellular architecture specifically on supporting acceptable levels of service and performance for non-
evenly distributed traffic. Small cell base stations are used to solve the ‘capacity hotspot’ problem where
there is a high level of subscribers concentrated in one location competing for limited service resources
resulting in poor performance to all users. Small cells are also used to solve the ‘coverage hole’ problem
where poor signal quality results in poor or no service.

Adoption of small cells and their deployment in large numbers are subject to resolving a few challenges,
of which one of the most prominent is backhaul. Current backhaul solutions such as optical fiber or line-
of-sight (LOS) microwave systems used for macro-cells don’t scale for small cells. Aside from technical
limitations of some technologies, the fundamental challenge is that of cost: small cell backhaul has to be
very cost effective for the entire business case for small cell deployments to be widely adopted.

This white paper provides an overview of the planning and design of Non-line-of-sight (NLOS) wireless
backhaul systems. NLOS systems can be deployed anywhere very quickly and as a result provide a very
cost effective solution both in terms of capital and operational expenditures than current wireless and
wireline solutions that are significantly more expensive.

The Benefits of Small Cell Base Stations
Small cell base stations are deployed low above ground (e.g. 3-5 m) and in close proximity to the mobile
subscriber which results in superior performance over macro cells that are deployed high above ground
and cover relatively large areas (typically around 300-500 m radius in urban areas). Capacity is injected
into the network by adding SCBS in key locations where a gathering of mobile users drive traffic demand
in a small area. This serves the additional purpose of off-loading traffic from the large macro cell
covering the area.

Another use case for small cell base stations is to enhance cell edge performance. Macro cells exhibit
poor performance at the cell edge: LTE was designed to provide about 0.04 – 0.06 bps/Hz/user at cell
edge in the downlink, and about half that in the uplink. In comparison, this is 2-3 times better than 3G
cell-edge spectral efficiency. Hence, small cells will be deployed in greater numbers at the cell edge to
improve user performance especially as a cell edge user may use a disproportionate amount of base
station resources. Table 1 shows the distribution of performance in a typical 4G wireless macro-cell in
urban area. The maximum cell range operates at the lowest modulation coding scheme which provides
the highest resiliency and robustness to errors. The area of a cell that operates on this modulation rate

1
Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update, 2011–2016.


is 40% of the total cell size. Hence, if we assume uniform subscriber distribution, a very large percentage
of users can only be served at the lowest performance level.

Table 1 Performance distribution in a wireless cell.
Spectral Efficiency Cell Radius Area
Modulation % Area
(b/s/Hz) (m) (Sq. Km)
QPSK1/2 1 610 0.39 40%
16QAM1/2 2 472 0.31 33%
64QAM2/3 4 319 0.26 27%
Total 0.97 100%

Small cell base stations offer considerable advantage to inject capacity in a wireless network. To
illustrate this, Figure 1 shows a simulation whereby 2 and 4 small base stations with 1 W output power
deployed in the coverage area of a macro-cell with 250 meter cell radius. The total downlink capacity
increases by 100% and 173% over that of a single sector with 2 and 4 small base stations, respectively2.

Figure 1 Aggregate downlink sector throughput for 0/2/4 small cell base
stations per macro-cell sector.

Small cell base stations are not the only solution to address the explosive demand for capacity. WiFi
offload is another example. When deployed outdoors, WiFi offload nodes will also require backhaul.
NLOS backhaul solutions can be used in this case in the same way as they are used to backhaul small cell
base stations.

2
The results are part of a more comprehensive study of small cell performance by Yuhan Zhou and Professor Wei
Yu at the Department of Electrical Engineering, University of Toronto and sponsored in part by BLiNQ Networks.


Overview of Cost of Backhaul Systems
Small cell base stations are a real solution to the capacity challenge, but there are a number of
challenges that must be overcome to allow their deployment in large numbers. Some of these
challenges are technical (e.g. interference management and handover), but most are related to cost of
deployment. The cost of backhaul figures prominently in this as current backhaul solutions, whether
wireline technologies including optical fiber or LOS wireless used in macro cell backhaul, do not scale for
small cell base stations. While a comprehensive look at the total cost of ownership for different
backhaul solutions is not within the scope of this white paper, we will touch on some of the main costs
related to planning and deployment as we will expand later in this paper on these two topics.

The cost of small cell wireless backhaul includes the cost of equipment, spectrum and recurring
expenses such as power and lease costs for antennas. It also includes the cost of planning, design and
deployment. These latter costs vary significantly between the different solutions and make a big
difference in the ability of the network operator to roll out a large number of small cells rapidly. For
example, a line-of-sight microwave system requires, as its name implies, a direct line-of-sight between
the two nodes, which is difficult to achieve in urban areas because of buildings, trees and other
structures. Moreover, LOS systems require two installation crews be present at the two ends for
alignment purpose. The alignment process itself is lengthy and leads to additional financial burdens: for
example, blocking a traffic lane during the installation process for a period exceeding a certain threshold
(e.g. would require certain municipal licenses and municipal worker’s presence on site, e.g. police or
traffic coordinator). For these reasons, a new cost effective and scalable backhaul solution is required
for small cells.

Non-line-of-sight (NLOS) wireless backhaul provides such a solution for cost-effective scalable small cell
deployments. NLOS typically uses an OFDM physical layer (Orthogonal Frequency Division Multiplexing)
that divides a wide-band frequency channel into many narrow-band sub-carriers. This result in high
tolerance to multipath fading and the consequent wireless channel impairments in a manner not
possible to achieve in LOS systems whose physical layer are based on a single wide-band carrier. NLOS
solutions operate in low-cost licensed band TDD (Time Division Duplex) spectrum in the sub 6 GHz
bands. The propagation conditions in such bands as 2.3, 2.6 and 3.5 GHz are very favorable to outdoor
deployments in urban areas providing a fair trade-off between coverage range and capacity of the
backhaul link. NLOS systems by definition allow operators to deploy systems anywhere a small cell is
required. This is a great advantage since traffic can be highly localized which limits the options of
locating a LOS small cell. Moreover, deploying NLOS system requires a single technician and one truck
roll only to place a remote backhaul module on a pole or building side wall and orient it towards a best
serving hub module.

The design of NLOS backhaul network is different from that of LOS microwave since detailed path
analysis is no longer required. Rather, planning tools are used to provide an estimate of the
performance at a location where small cell is desired. In this regard, NLOS backhaul is similarly planned
to fixed wireless access service systems.


NLOS Wireless Backhaul Deployment Scenario
NLOS wireless backhaul network consists of a hub module (HM) and a remote backhaul module (RBM).
The hub module connects to several RBMs in a point-to-multipoint configuration. The hub module is
typically situated at a fiber point-of-presence or where high-capacity LOS microwave system is available
to connect the hub to the core network. An existing macro-cell can be such a site. Typical height for a
hub module in an urban area ranges between 25-40 meters above ground.

The hub modules are typically connected to a wide horizontal beam width antenna (65 or 90 degrees).
Therefore, they have a relatively wide coverage and can connect to multiple RBMs. Multiple hubs can be
located on the same site in a configuration similar to the macro-cellular network. For example, three or
four hubs can be located on a building rooftop to serve the entire surroundings of the site. Figure 2
shows conceptually how the hub module is mounted on a building rooftop where fiber or LOS
microwave backhaul to the core network is present. Multiple remote backhaul modules are connected
wirelessly to the hub module in a point-to-multipoint configuration while they connect to the small cell
base station over a Cat5 Ethernet cable.

Figure 2 NLOS wireless backhaul hub modules are mounted on building rooftop. Multiple
remote backhaul modules connect to the hub module in point-to-multipoint configuration.

The remote backhaul module is co-located with the small cell base station on a pole (e.g. light pole,
utility pole, etc.), building sidewall, or some other structure. Small cell base stations, and henceforth the
RBMs, are typically located at 3-8 m in height. Therefore, the remote backhaul module is located below
the roofline and typically, there would not be a clear line-of-sight to the hub module.

RBMs feature have a directional antenna which is pointed towards the best serving hub module. The
directional antennas on RBMs serve to increase the system gain and to reduce interference by directing
transmitted electromagnetic energy into the direction of the serving hub module. They also receive less
interference from hub modules not located within their antenna boresight.

For the simulations presented in this white paper BLiNQ’s X100 NLOS wireless backhaul system with 83
Mbps peak capacity in 10 MHz channel is used. This is achieved with 256QAM 7/8 modulation in MIMO
(multiple input multiple output) operation with two transmit and receive antennas in spatial
multiplexing mode where two code words are transmitted simultaneously, one on each antenna. Lower


capacity is achieved depending on the signal quality (as measured by carrier-to-noise plus interference
ratio; CINR) and corresponding modulation coding scheme (MCS) selection.

System Gain and Path Loss
As in any wireless system, the system gain provides the allowable path loss between the transmitter and
receiver. The system gain is a function of a number of parameters that include:

• Output power
• Transmit and receive antenna gain
• Receiver sensitivity (which is a function of channel bandwidth, noise figure of the receiver, and
the signal-to-noise ratio at the desired modulation scheme)
• Specific features of the transmitter and receiver such as transmission and reception diversity.

The allowable path loss is the maximum propagation attenuation that a link can encounter without loss
of communication. The path loss is inversely related to distance r from the transmitter (∝ ‫ି ݎ‬௡ ) where n
is the power decay factor (or path loss exponent). To find the maximum distance a receiver can be
located away from the transmitter, the path loss is reduced by a ‘fade margin’ that accounts for the
fading expected in the wireless channel.

Non-line-of-sight wireless backhaul solutions operate at too low a frequency (compared to classical
broadband backhaul) to be impacted by rain, snow, or other natural conditions which impact the
performance of LOS systems. However, in turn, NLOS systems are subject to shadow and small scale
fading. Shadow fading is the result of the interaction with large objects such as buildings in surrounding
environmental clutter that could be vastly different at two different locations having the same
transmitter-receiver separation. Small scale fading is caused by the addition of multipath components
(with different phases) and thus varies for small changes in location of the RBM. For a given RBM
location, reflections off objects in motion will cause the small scale fade to vary with time. Hence, a
proper fade margin must be accounted for in path loss calculations.

Shadow fading is referred to as log-normal shadowing and is fully characterized by a mean and a
standard deviation which typically is about 9 dB for urban environment. For a given contour around a
transmitter, shadow fading can be calculated for a certain probability of receiving a signal at that
contour. The fraction of total area with connectivity (signal above threshold) is derived from the contour
probability.

The amount of small scale fading depends on the wireless propagation channel which is impacted by the
deployment scenario (height of antennas), surrounding environment (e.g. dense urban area) as well as
by the type and parameters of the antenna system (e.g. beamwidth). Small scale fade margin is also
related to the desired percentage of link availability. In small cell wireless backhaul, it’s typically
acceptable to have 99.9% link availability. We factored 4 dB of standard deviation for small scale fading
in a typical NLOS small cell backhaul deployment scenario.


Table 2 shows an example of fade margin for path loss exponent of 4 and composite standard deviation
of 9.85 dB estimated for the above deployment scenario in a dense urban area. Therefore, 32 dB of fade
margin results in 99.9% probability of having connectivity at the contour around a transmitter which
translates into 99.99% of total area within this contour of having connectivity.

Table 2 Fade margin for path loss exponent of 4 and standard deviation of 9.85.
Fade Margin (dB) Contour Reliability Area Reliability
14.2 85% 97.4%
16.2 90% 98.35%
25.4 99% 99.87%
32.4 99.9% 99.99%

Capacity, Coverage & Link Reliability
The system gain, fade margin and resulting path loss define the link performance from a capacity, link
reliability and coverage perspective. Figure 3 shows that for a maximum system gain of 136 dB it is
possible to achieve about 300 m coverage radius with 32 dB of fade margin which corresponds to the
chosen link availability. The model used is that for an urban area at 3.5 GHz. Increasing the fade margin
increases the link availability and reduces the maximum allowable path loss and consequently the
distance between the transmitter and receiver. Selecting a modulation coding scheme with lower CINR
requirements increases the system gain while it reduces the link capacity. However, this results in either
larger distance, greater reliability or a combination of both.

Figure 3 Trade-off between path loss and fade margin for a preset system gain.

Network Level Simulation & Planning
Having discussed the design of NLOS wireless backhaul conceptually, we like to outline the use of RF
network planning tools in the design process. Network level simulations are necessary in the perspective


of a large-scale introduction of small-cells. For the mobile user service coverage, the predictions are
helpful for the characterization of the inter-cell interference, the evaluation of interference mitigation
techniques, the evaluation of achievable capacity offloading and power consumption. For the wireless
backhaul design, the simulations are paramount to predict the performance of the backhaul network
especially in urban areas under NLOS conditions. Network level simulations also provide relevant inputs
towards innovative business models for the equipment and the system.

SIRADEL extended usual radio-planning simulation tool capability and network design methodologies to
assess small-cell deployment coverage and performance in addition to the NLOS backhaul component.
An accurate description of a real environment (high-resolution geographical map data) and site-specific
path-loss models allow for the prediction of NLOS areas, in-street canyoning effects and realistic
outdoor-to-indoor propagation. In the remainder of the paper, we consider only the NLOS urban
propagation loss estimation for the backhaul link (hub module and a remote backhaul module.) We do
not use the simulation capabilities for links between the small cells and the mobile users. Before we
present results for the backhaul link simulations, we discuss the tools available for NLOS small cell
backhaul planning which center on two main elements: a. high resolution geographical data and b. path
loss matrices.

Tools for NLOS Small Cell Backhaul Planning

Digital Geographical Data
RF coverage tools make use of digital descriptions of the geographical environment. This information
allows us to place the transmitters at desired locations on a digital map, and to visualize the areas in
which the signal reception can be observed. The geographical data also contains crucial information to
predict the radio-wave propagation.

Geographical map data is composed of a set of distinct layers; each layer contains the description of one
terrain characteristic such as geographical obstacle type (also called clutter or land usage), clutter height
above ground, and ground altitude (or terrain elevation).

A pixel matrix (aka raster) layer is a collection of geo-localized rectangular pixel matrices; the value at
each pixel may represent at one precise location a clutter type in a DLU (Digital Land Usage), a clutter
height in a DHM (Digital Height Model), or an altitude in a DTM (Digital Terrain Model) or DEM (Digital
Elevation Model).

A raster matrix with a resolution lower than 10 meters is generally referred to as HR (high resolution)
matrix: it provides clear and precise contours of every single building, row of trees or small urban
woods.

A clutter type may be represented by closed horizontal polygons that are vector data; sometimes the
vector layer contains only building data but actually all clutter types can be represented (buildings,
vegetation, bridges, and water). A 3D vector layer contains attributes that gives the clutter height
associated with each of the above-mentioned vector polygons.


Figure 4 Vector data showing buildings, vegetation, and bridges in Paris, France.

Path-loss Matrices
The coverage simulations are obtained by predicting path-loss matrices, realized from the ray-based
“Volcano” model developed by SIRADEL. This model generates multiple propagation paths in complex
environments in reduced computation time. The method is based on separating the calculation of
vertical propagation effects from lateral effects in a two-step process. In the first step, the lateral
interactions of the radio wave with building façades (represented with 1m precision in a 3D vector
database) give rise to multiple reflected and diffracted horizontal “rays.” In the second step, a 3D ray
trajectory is obtained by constructing an “unfolded profile,” which contains all terrain obstacles found in
the ray’s vertical propagation plane (ground, building or vegetation). This information is extracted from
the 3D vector database completed by a high-resolution terrain altitude grid (5 meters resolution in
horizontal plane and 1 meter vertical precision). Losses are calculated from diffraction off the obstacles.

The Volcano tool gathers many heuristic parameters obtained from many years of engineering
experience, thus the ray-based model is understood as pre-calibrated. Nevertheless the present work
has benefited from a specific calibration, since a large amount of measurements were available in the
study environment at the studied frequency.

Comparing NLOS to LOS System Availability in Urban Areas
One of the main purposes of using a RF coverage tool for NLOS small cell backhaul is to determine the
availability of service in a specific location which is a key to reducing small cell deployment timeline and
cost. The design process begins by identifying the location and height of the hub modules which are
typically located at a fiber point-of-presence of where LOS microwave is available to connect the hub to
the core network. Next, we completed simulations to obtain the coverage achieved from the hub
locations over a relatively wide area for different small cell heights. The assumption is that the remote
backhaul module would be placed at the same height as the small cell. Results of the simulations would
provide a measure of expected backhaul performance over the area under simulations. The process is
shown in Figure 5.


Network Topology
Fixed macro network
- Hub modules located on sites

Backhaul Simulation
Coverage simulations Service
- Hub modules on existing site availability
- Different small cell heights

Analysis
Design optimization of NLOS backhaul system
- For different small cell heights
- For different backhaul system link budgets

Figure 5 Simulation Synopsis.

Simulations were performed for a central area of Paris where small cell base stations were located at 5,
7, 10, 15 and 20m. The hub modules were placed at locations of existing macro-base stations with
heights between 30m and 40m above ground. The average building height is 25m. The simulation
parameters are given in Table 3.

Table 3 Hub module & location properties

Sectorization 4 Sectors
Antenna Horizontal aperture: 47°, Vertical Aperture: 7°
Antenna down-tilt Depends on Macro BS height, SCBS height and MacroBS cell radius
Antenna height 3 meters above a dominant building rooftop
Antenna locations At the edge of rooftop
Antenna gain 17 dBi
Tx Power 24 dBm
Tx EIRP 41 dBm


Figure 6 presents availability of service for the different small cell base stations heights (i.e. remote
backhaul module height). Each map presents results for both LOS (28 GHz) and NLOS (3.5 GHz) backhaul
systems for comparative purposes. It can be seen that the greater the height of the small cell, the
greater the area over which a certain modulation scheme is achieved which results from lower path loss
between the hub and remote backhaul modules.

a) Small cell height = 5m b) Small cell height = 7m

c) Small cell height = 10m d) Small cell height = 15m

e) Small cell height = 20m
Figure 6 Availability of service for LOS and NLOS backhaul for different Small cell
antenna heights: (a) 5m, (b) 7m, (c) 10m, (d) 15m, (e) 20m


Table 4 summarizes the availability of service for a NLOS system operating at 3.5 GHz for 256QAM 7/8
and 5/8 with 95% and 50% availability probability and LOS system at 28 GHz. It is important to state
that this availability figure is that of “area” availability which is the probability that the signal will be of
sufficient quality to establish a link at that location according to the corresponding modulation scheme.
As expected, a greater probability for LOS service is possible for the higher small cell base station. But at
typical SCBS height of 5 m, the probability of LOS achieving connectivity is only 5%. By comparison, NLOS
system would be available in 95% of the area for 256QAM 5/8 modulation (which corresponds to 61
Mbps in a 10 MHz channel).

Table 4 Example of availability of service in dense urban area.
3.5 GHz 3.5 GHz
SCBS
256 QAM 5/8 256 QAM 7/8 LOS
Height
50% 95% 50% 95%
5m 87.1% 43.1% 45.9% 14.9% 5.0%
7m 90.0% 46.7% 49.7% 16.7% 6.2%
10 m 91.1% 52.0% 55.3% 19.8% 9.9%
15 m 93.2% 65.1% 68.5% 26.6% 22.2%
20 m 96.8% 79.9% 82.8% 37.4% 51.6%

Conclusion
A cost effective backhaul solution is required to enable the business case for large scale small cell
deployments. NLOS wireless backhaul is the solution that includes the low cost of equipment, the low
cost spectrum which can be licensed on a block basis thereby resulting in diminishing marginal cost per
added link, and the low cost of planning, installation and deployment. In this paper, we presented a
solution by BLiNQ Networks that is optimized for NLOS small cell wireless backhaul. It operates in the
sub 6 GHz licensed bands and allows operators to roll out small cell base stations with a single technician
and truck roll for the deployment of the backhaul solution. The system can be deployed anywhere on
light infrastructure assets in point-to-multipoint configuration which reduces total cost of ownership.
The paper illustrates how NLOS wireless systems can be planned using SIRADEL’s high resolution
geographic digital data and Volcano ray-based model which is optimized for accuracy and speed to
calculate path loss. Simulations show that NLOS service in dense urban area such as in the city of Paris is
available with very high probability even at relatively low heights (e.g. 43% of locations with 95%
confidence at 5 m height for 256QAM 5/8) while building rooflines make connectivity for line-of-sight
solutions challenging such that only 5% of locations would have a direct LOS at SCBS height of 5 m.


BLiNQ Networks Inc. SIRADEL S.A.S.
400 March Road, Suite 240 Wireless Expertise and Research Center
Ottawa, ON K2K 3H4 Canada 1920 Yonge Street, Suite 200
Main: 613-599-3388 Toronto, ON M4S 3E2 Canada
info@blinqnetworks.com Tel: +1 416 572 7618
www.blinqnetworks.com info@siradel.com
www.siradel.com

SIRADEL is a high-tech company created in 1994 and based in France, China (Hong-Kong) and Canada (Toronto).
The portfolio of the company is composed of:
- 3D GIS data and RF measurements
- Advanced RF Tools (Volcano, VolcanoLab)
- Management and Technology Consulting
SIRADEL serves more than 250 Customers in about 50 countries. SIRADEL's solution brings more reliable and
realistic assessments of wireless network and wireless equipment performances. The key benefits for SIRADEL’s
partners are to significantly reduce the cost of radio infrastructures. The profile of its Customers is diverse:
wireless carriers, radio access equipment manufacturer, regulation bodies, utilities, consultants.

BLiNQ Networks was founded in June 2010 after the acquisition of intellectual property and wireless assets from
Nortel Networks. BLiNQ is a pioneer of wireless backhaul solutions that fundamentally change the way mobile
operators deliver mobile broadband services in urban areas. BLiNQ uses cost-effective sub-6 GHz spectrum and
unique and patent-pending Managed Adaptive Resource Allocation (MARA) technology to provide network-level
intelligence, self-organizing network capabilities, and eliminate interference challenges to maximize spectral
efficiency. BLiNQ is headquartered in Plano, TX with research and development facilities in Ottawa, Canada. For
more information, please visit www.blinqnetworks.com.

The information presented herein is to the best of our knowledge true and accurate and is subject to change without notice. No
warranty or guarantee expressed or implied is made regarding the performance or suitability of any product. All product or
service names are the property of their respective owners. © BLiNQ Networks Inc. & SIRADEL S.A.S. 2012. All Rights Reserved.


Planning Non Line-of-Sight Wireless Backhaul Networks

More Related Content

What's hot (19)

Similar to Planning Non Line-of-Sight Wireless Backhaul Networks (20)

More from Frank Rayal (7)

Recently uploaded (20)

Planning Non Line-of-Sight Wireless Backhaul Networks