SANSA-645047 D2.3 D2.3 Definition of reference scenarios, overall system architectures, research challenges, requirements and KPIs Grant Agreement nº: 645047 Project Acronym: SANSA Project Title: Shared Access Terrestrial-Satellite Backhaul Network enabled by Smart Antennas Contractual delivery date: 01/02/2016 Actual delivery date: 01/02/2016 Contributing WP WP2 Dissemination level: Public Editors: AVA Contributors: CTTC, TAS, ULUX, AIT, OTE Abstract: This deliverable contains the outcomes of Task 2.3 and Task 2.4. It will specify the scenarios and network architectures covered in SANSA project as well as the technological requirements and KPIs.
131
Embed
D2.3 Definition of reference scenarios, overall …...D2.3: Definition of reference scenarios, overall system architectures, research challenges, requirements and KPIs Date: 01/02/2016
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
SANSA-645047
D2.3
D2.3
Definition of reference scenarios, overall system architectures, research challenges, requirements and
Abstract: This deliverable contains the outcomes of Task 2.3 and Task 2.4. It will specify the scenarios and
network architectures covered in SANSA project as well as the technological requirements and
KPIs.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 2 of 131
SANSA-645047
D2.3
Document History
Version Date Editor Modification 0.0 22/06/2015 AVA Document creation and Table of Contents
0.1 09/09/2015 AVA Initial input from AVA regarding the overall scenario
selection strategy, the rural scenarios, specific
network elements and some examples of KPIs as
well as initial input regarding end to end system
architecture from CTTC
0.2 29/09/2015 AVA CTTC input regarding urban scenario and research
challenges, OTE input on the requirements and the
research challenges, AVA input on research
challenges and system architecture requirements.
KPI inputs by TASE and OTE
0.3 -0.5 From
29/09/2015
to
18/01/2016
AVA Various inputs from AVA, ULUX, CTTC, OTE, TAS &
OTE.
0.6 18/01/2016 AVA Document Ready for QA
0.7 25/01/2016 AVA QA by AIT
0.8 29/01/2016 AVA QA by AVA
0.9 29/01/2016 AVA QA documents addressed
1.0 29/01/2016 AVA Final document ready for submission
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 3 of 131
SANSA-645047
D2.3
Contributors
Name Company Contributions include
Georgios Ziaragkas AVA All document
Georgia Poziopoulou AVA All document
Simon Watts AVA Quality Assurance
George Agapiou OTE Chapters 3, 5
Ioanna Papafili OTE Chapter 6
Pantelis Argyrokastritis OTE Chapters 3, 5
Panagiotis Matzoros OTE Chapters 3, 5
Xavier Artiga Campos CTTC Document Review
Miguel Ángel Vázquez CTTC Chapters 2, 3 and 6
Jose Núñez Martinez CTTC Chapter 4, 5
Musbah Shaat CTTC Chapters 2, 3
Symeon Chatzinotas ULUX Chapter 3, 5, 6
Dimitrios Christopoulos ULUX Chapter 3, 5, 6
Shree Krishna Sharma ULUX Chapter 3, 5, 6
Christos Tsinos ULUX Chapter 3, 5, 6
Isaac Moreno TASE Chapters 4, 6
Dimitris Ntaikos AIT Chapters 2, 3, 4
Kostas Voulgaris AIT Quality Assurance
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 4 of 131
SANSA-645047
D2.3
Table of Contents
List of Tables ......................................................................................................................................... 11
List of Acronyms .................................................................................................................................... 12
Table A-2: SINR values of the different sub-scenarios ........................................................................ 125
Table A-3: Interference power levels .................................................................................................. 126
Table A-4: Parameters for the 18 GHz terrestrial link budget ............................................................ 126
Table A-5: Terrestrial link budget ....................................................................................................... 129
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 12 of 131
SANSA-645047
D2.3
List of Acronyms
ADC Analog-to-Digital Converter
AP Application Protocol
ASIC Application Specific Integrated Circuit
BATS Broadband Access via integrated Terrestrial and Satellite systems
BN Backhaul Node
BS Base Station
BSS Business Support System
CAPEX Capital Expenditure
CDN Content Delivery Network
CS Compressive Sensing
CSI Channel State Information
CZ Cognitive Zone
DL Downlink
DoA Direction of Arrival
DSCP Differentiated Services Code Point
DVB-S Digital Video Broadcasting – Satellite
EC European Commission
EIRP Effective Isotropic Radiated Power
eNB eNodeB
EPC Evolved Packet Core
ETSI European Telecommunications Standards Institute
EU European Union
FDD Frequency Division Duplex
FPGA Field-Programmable Gate Array
FSS Fixed Satellite Service
FWD Forward (channel)
GEO Geostationary Orbit
GTP-U GPRS Tunnelling Protocol User Plane
HDFSS High-Density applications in the Fixed-Satellite Service
HeNB Home eNodeB
HNM Hybrid Network Manager
HSS Home Subscriber Server
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 13 of 131
SANSA-645047
D2.3
HTS High Throughput Satellite
IA Interference Alignment
iBN Intelligent Backhaul Node
IPsec IP Security
KPI Key Performance Indicator
LAA License Assisted Access
LoS Line of Sight
LSA License Spectrum Access
LSAS Large Scale Antenna Systems
LTE Long Term Evolution
MBBS Macrocell Backhaul Base Station
MBS Mobile Base Station
MCN Mobile Core Network
MME Mobile Management Entity
MNO Mobile Network Operator
MODCOD Modulation Coding
NG HTS Next Generation High Throughput Satellite
NMS Network Management System
O&M Operation & Maintenance
OFDMA Orthogonal Frequency Division Multiple Access
OPEX Operational Expenditure
OSS Operations Support System
PDN-GW Packet Data Network Gateway
PLR Packet Loss Ratio
PtMP Point-to-MultiPoint
PtP Point-to-Point
QoS Quality of Service
R&D Research & Development
RAN Radio Access Network
REM Radio Environment Map
RF Radio Frequency
RTN Return (channel)
SA Smart Antenna
S-GW Serving Gateway
SINR Signal to Interference plus Noise Ratio
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 14 of 131
SANSA-645047
D2.3
SMS Short Messaging Service
SNR Signal to Noise Ratio
SoTA State of the Art
TCO Total Cost of Ownership
TDD Time Division Duplex
TDMA Time Division Multiple Access
TNA Transport Network Architecture
UE User Equipment
UL Uplink
UWB Ultra Wide Band
VSAT Very Small Aperture Terminal
Wi-Fi Wireless Fidelity
WP Work Package
WSS Wideband Spectrum Sensing
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 15 of 131
SANSA-645047
D2.3
Executive Summary
This is Deliverable D2.3 of the SANSA “Shared Access Terrestrial-Satellite Backhaul Network enabled
by Smart Antennas” project which documents the outcome of Task 2.3 “Scenario selection and
network architecture definition” and Task 2.4 “Definition of Key Performance Indicators. The
objective of this deliverable is to define the appropriate use cases, with its possible assumptions and
constraints, and to define and form the overall system architecture in an end-to-end fashion. It also
aims at looking at the technical challenges to be addressed when designing our proposed solutions
and how they are improving the current state-of-the-art. It is also defining the Key Performance
Indicators to be used while evaluating, comparing and selecting among different technological
solutions. The material documented in D2.3 will allow the SANSA consortium to define the reference
scenarios, the overall system architectures, the research challenges/requirements and the Key
Performance Indicators.
In this report after identifying the use cases, the most relevant scenarios are being selected and the
topologies to be used to further research the interference landscape, network architectures and the
design of the key enabling components are provided. End to end system architecture is also
provided along with the KPIs that will be used to evaluate the novelties and the performance
enhancement that SANSA introduces. Last but not least, all research challenges associated with
SANSA are presented.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 16 of 131
SANSA-645047
D2.3
1. Introduction
The objective of this document is to define the use cases and scenarios, as well as the end-to-end
system architecture and the end-to-end KPIs that will be used to assess the SANSA solution. Lastly,
the technical challenges that will be studied in the following WPs will be specified.
This deliverable is the output of the work done under Tasks 2.3 and 2.4 of WP2. WP2 aims to define
the scenarios and network architectures that will be used in SANSA as well as the Key Performance
Indicators (KPIs) for evaluating the proposed SANSA solution. Figure 1-1 illustrates how Tasks 2.3
and 2.4 fit within the WP2 work plan.
Figure 1-1: WP2 work plan
This remainder of this deliverable is divided into six chapters that are organised as follows:
Chapter 2 discusses the SANSA use cases and explains why they are relevant to the project.
These are radio link failure and/or congestion, new node deployment, CDN integration in the
network and remote cell connectivity which can refer either to a standalone cell or a moving
base station.
Chapter 3 presents the 5-axis scenario definition strategy and fully defines the SANSA scenarios.
The strategy followed takes into consideration the type of deployment, the CDN integration, the
type of spectrum sharing and the terrestrial and satellite link characteristics to create the 6
scenarios which will be used from the following WPs to further investigate the SANSA solution.
Additionally, this chapter summarises the common general scenario characteristics such as the
air interfaces and spectrum licenses used, the strategy for selecting the nodes that will be
equipped with a satellite terminal and traffic forecasting at the macrocell and CDN level. Lastly,
the interference landscape for the SANSA network is presented along with benchmarking
topologies for the rural, urban and moving base station scenarios with a focus on the
interference analysis. These benchmark topologies provide the basis of the interference
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 17 of 131
SANSA-645047
D2.3
mitigation techniques that will be studied in WP3.
Chapter 4 defines the SANSA end-to-end architecture including the transport and key enabling
components architecture.
Chapter 5 reviews the objectives of the project and sets appropriate end-to-end KPIs which will
be used to evaluate the outcomes of the project. For every KPI a definition, the method of
evaluation, the target value and a motivation with clear links to the project objectives is
provided.
Chapter 6 identifies the research challenges that will drive the work in subsequent WPs. The
main thematic categories presented are challenges associated with the integration of the
terrestrial and satellite backhaul nodes, the key enabling components, the spectrum sharing
and the integrated backhaul service delivery.
Chapter 7 draws conclusions from the analysis of the preceding chapters.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 18 of 131
SANSA-645047
D2.3
2. SANSA use cases
2.1 Scope SANSA project was stimulated by a set of existing problems in the current mobile backhauling
networks. In order to solve these problems and design an enhanced network with improved
backhauling capabilities, a set of objectives has been proposed. These objectives are challenging
current network designs and try to solve the problems they are encountering. They are also relevant
and according to the targets set by 5G. Within the frame of the solutions design, a number of use
cases have been identified in order to properly define the problems and create the appropriate
framework, the scenarios, in which a number of problems are addressed in a structured manner. In
order to quantify our scenarios and to be able to simulate problems and test solutions, instances of a
part of the network under a specific scenario will be used as the reference topology.
As already mentioned in the previous paragraph and throughout the document the terms use case,
scenario, topology will be used. At this point it is very important to clearly define the meaning of
each of the terms in order to avoid confusions and present clearly the way scenario definition is
structured.
Use case is the term used to describe a problem created or a challenge that has to be addressed
within a network. In SANSA the use cases are resiliency, offloading, new node development, CDN
integration and remote cell connectivity. It is worth mentioning here that the periodicity in which
events occur is really important to define the scenarios properly.
Scenario is a specific situation, where one or more of the use cases have to be addressed. The main
separation is into three main categories rural, urban and moving base station. Within each of the
main categories there are sub-scenarios, which are direct results of changes in the scenario
parameters (e.g. the use of CDN or not in a rural environment is a variable that can change the rural
scenario).
Topology is the term used to define an instantiation of a part of a network. It will be used to
demonstrate the new features and capabilities that SANSA will offer. Interference analysis, hybrid
network management and optimization, satellite integration, smart antenna capabilities, energy
efficiency and re-configurability will be demonstrated on benchmark topologies but the results can
be extrapolated and used to different network topologies as SANSA is not aiming at providing a
network oriented solution, but a solution that can be adopted by any network and any operator.
Subsequent WPs will work taking into consideration as reference the scenarios and the topologies as
they will be defined within this deliverable.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 19 of 131
SANSA-645047
D2.3
2.2 SANSA selected use cases
2.2.1 Radio link failure One of the main SANSA characteristics is the ability to offer resilience to cases of link failure. The
satellite connectivity and the use of the smart antennas with the help of the Hybrid Network
Manager add flexibility to backhauling networks by either providing an alternative route, through
satellite, where satellite terminals are available to the user or through an alternative terrestrial link
when another node is available with the use of a smart antenna with steerable beams.
In Figure 2-1 some examples of link failure events and how SANSA system provides resilience to the
affected links are presented:
Event 1: Link A-B fails. Traffic from node A cannot access the core.
Event 2: Link D-G fails. Traffic should follow the D-E-F-G path to access the core. This creates
congestion to the rest of the nodes.
Event 3: Node G fails. No access to the core.
Event 4: Link F-I fails. Traffic from nodes H, I and J cannot access the core.
Event 5: Link H-I fails. Node H cannot access the core.
The above mentioned events are causing network failures. SANSA network is designed in a manner
to offer resilience for link failures. The system response as illustrated in Figure 2-2 to each of the
events is as follows:
Event 1: Link A-B fails. The satellite terminal on node A redirects traffic to the core network
through the satellite network.
Event 2: Link D-G fails. This link is part of a ring topology that reroutes traffic through the
alternative way. The novelty with SANSA is that the traffic that could create moderate or
heavy congestion to the system can be partially offloaded at node E, with the help of the
iBN, according to the traffic classification rules that have already been set.
Event 3: Node G fails. Node E acts as a backup aggregator, routing the traffic through the
satellite network to the core or directly to the internet, depending on the selected service
model.
Event 4: Link F-I fails. Again the VSAT terminal at node F is used as a backup connection to
the core or to the internet, routing traffic from all affected nodes (H, I, J).
Event 5: Link H-I fails. Node H cannot access the core but the steerable smart antenna
directs its beam from I to J and establishes a new terrestrial link. The increased traffic can be
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 20 of 131
SANSA-645047
D2.3
offloaded through the satellite connection at node I.
The events description and responses are presented in Table 2-1.
Table 2-1: Link failure events description and responses
Event Failure Description Response
1 Link A-B Traffic from node A cannot
access the core.
The satellite terminal on node A redirects
traffic to the core network through the
satellite network.
2 Link D-G Traffic should follow the D-E-
F-G path to access the core.
This creates congestion to the
rest of the nodes.
This link is part of a ring topology that
reroutes traffic through the alternative way.
The novelty with SANSA is that the traffic
that could create moderate or heavy
congestion to the system can be partially
offloaded at node E, with the help of the
iBN, according to the traffic classification
rules that have already been set.
3 Node G No access to the core. Node E acts as a backup aggregator, routing
the traffic through the satellite network to
the core or directly to the internet,
depending on the selected service model.
4 Link F-I Traffic from nodes H, I and J
cannot access the core.
Again the VSAT terminal at node F is used
as a backup connection to the core or to the
internet, routing traffic from all affected
nodes (H, I, J).
5 Link H-I Node H cannot access the
core.
Node H cannot access the core but the
steerable smart antenna directs its beam
from I to J and establishes a new terrestrial
link. The increased traffic can be offloaded
through the satellite connection at node I.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 21 of 131
SANSA-645047
D2.3
A B C
D
E
F
G
H
IJ
Mobile Core Network/EPC
Mobile Core Network/EPC
Gateway antenna
Satellite backbone fibre
network
Satellite backbone fibre
network
Ka band GEO satellite
TeleportTeleport
BEFORE LINK FAILURES
Satellite hub
Mobile Core Network/EPC
Mobile Core Network/EPC
Figure 2-1: Radio link topology before failures
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 22 of 131
SANSA-645047
D2.3
SA
A B C
D
E
F
G
H
IJ
Mobile Core Network/EPC
Mobile Core Network/EPC
Gateway antenna
Satellite backbone fibre
network
Satellite backbone fibre
network
Ka band GEO satellite
TeleportTeleport
Link Failures
New Links
Event 1
Event 2
Event 3
Event 4
Event 5
EVENTS AND SOLUTIONSSatellite
hub
Mobile Core Network/EPC
Mobile Core Network/EPC
Figure 2-2: Radio link failures events and solutions
2.2.2 Radio link congestion One of the SANSA use cases is to provide offloading capability via satellite to the backhaul network.
This use case is illustrated in the Figure 2-3 below.
Figure 2-3a presents the network topology before the events of congestion. Based on the topology
all the nodes are routing their traffic appropriately so that they reach Node G that is connected to
the EPC with optical fibre. The arrows indicate the flow of traffic within the backhaul network.
In Figure 2-3b four congestion events occur in the network:
Event 1: Heavy congestion on the link B-C which affects the traffic coming from Nodes A and
B.
Event 2: Moderate congestion on the link D-G which affects the traffic coming from Nodes A,
B, C and E.
Event 3: Heavy congestion on the link F-G which affects the traffic coming from Nodes I, J
and H.
Event 4: Moderate congestion on the link I-F which affects the traffic coming from Nodes J
and H.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 23 of 131
SANSA-645047
D2.3
The congestion events trigger the HNM to adapt the topology to optimise the overall performance.
Figure 2-3c illustrates how the topology changes to tackle the congestion events:
Event 1: Node A activates the satellite link for backhauling instead of forwarding the traffic
to Node B which helps decongest the link B-C.
Event 2: In Node E part of the traffic is offloaded thought the satellite link resulting in a
reduction of the total traffic arriving at Node D. Subsequently the link D-G is decongested.
Event 3: The SA in Node F creates a second link with Node E (F-E) so that the traffic from
Node F is split between the links F-G and F-E. The traffic arriving at Node E is then
backhauled through the satellite link of the node.
Event 4: As the link I-F is starting to become congested, Node I offloads part of the traffic
through the satellite link.
The events description and responses are presented in
Table 2-2.
Table 2-2: Congestions events description and responses
Event Congestion events Response
1 Heavy congestion on the link B-C which
affects the traffic coming from Nodes A and
B.
Node A activates the satellite link for
backhauling instead of forwarding the
traffic to Node B which helps decongest the
link B-C.
2 Moderate congestion on the link D-G which
affects the traffic coming from Nodes A, B, C
and E.
In Node E part of the traffic is offloaded
thought the satellite link resulting in a
reduction of the total traffic arriving at
Node D. Subsequently the link D-G is
decongested.
3 Heavy congestion on the link F-G which
affects the traffic coming from Nodes I, J and
H.
The SA in Node F creates a second link with
Node E (F-E) so that the traffic from Node F
is split between the links F-G and F-E. The
traffic arriving at Node E is then backhauled
through the satellite link of the node.
4 Moderate congestion on the link I-F which
affects the traffic coming from Nodes J and
As the link I-F is starting to become
congested, Node I offloads part of the
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 24 of 131
SANSA-645047
D2.3
H. traffic through the satellite link.
A B C
D
E
F
G
H
IJ
Mobile Core Network/EPC
Mobile Core Network/EPC
a) BEFORE CONGESTION
b) CONGESTION EVENTS
A B C
D
E
F
G
H
IJ
Event 1
Event 2
Event 3Event 4
Mobile Core Network/EPC
Mobile Core Network/EPC
c) CONGESTION RESOLUTION
A B C
D
E
F
G
H
IJ
Satellite backbone fibre
network
Satellite backbone fibre
networkMobile Core Network/EPC
Mobile Core Network/EPC
Mobile Core Network/EPC
Mobile Core Network/EPC
New links
Not congested links
Moderately congested links
Heavily congested links
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 25 of 131
SANSA-645047
D2.3
Figure 2-3: Congestion use case for the rural scenario
2.2.3 New node deployment/incorporation The deployment and incorporation of new nodes is a major issue for mobile operators. One of the
SANSA targeted contributions is to provide operators with hybrid nodes that are low cost, easily
deployable and energy efficient.
The work on cost effective solutions regarding the antennas along with the use of already existing
technology as far as terrestrial equipment is concerned will have a positive impact on the overall
hybrid node cost.
SANSA offers flexibility as a result of the hybrid nature of the nodes that incorporates and reassure
seamless connectivity through satellite or terrestrial paths. In practice this will result in a solution
that is not heavily dependent on the already existing infrastructure as satellite connection ensures a
high capacity solution is available in areas with poor backbone network infrastructure and areas
where is difficult geographically to establish a microwave link. The operators do not have to wait for
new backbone infrastructure to be developed before they can deploy their new nodes.
As far as energy consumption is concerned, new nodes will have minimized power consumption and
the smart network reconfiguration and power mode central management will make sure that there
is no excessive power consumption with positive impact on both OPEX costs and deployment in
areas where power supply is limited (e.g. use of photovoltaic panels to supply with power the hybrid
node).
2.2.4 CDN integration As far as Content Delivery Networks are concerned, within the scenarios, the integration of these
networks can be supported either by a sole satellite connection or through combined satellite and
terrestrial links.
2.2.4.1 Satellite fed cache CDN In this case all the content sent to the edge CDN is fed only through satellite. The content is
delivered to the cache, which is located at the eNB, through the satellite connection as can be seen
in Figure 2-4.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 26 of 131
SANSA-645047
D2.3
Figure 2-4: Satellite fed CDN cache
2.2.4.2 Cache fed through satellite and terrestrial link In this case the CDN edge network is fed both through the satellite and the terrestrial connection
available at the eNB. The content is delivered to the cache either through the satellite or through the
terrestrial connection depending on the network conditions as can be seen in Figure 2-5.
Figure 2-5: CDN cache fed through satellite and terrestrial connection
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 27 of 131
SANSA-645047
D2.3
2.2.5 Remote cell connectivity
2.2.5.1 Standalone cell One of the main problems that current mobile networks encounter is the difficulty in deploying
network infrastructure in areas that are isolated (e.g. islands), or areas that lack backbone
infrastructure and the geographical location does not allow the establishment of microwave links.
Standalone cells can bypass the lack of infrastructure by sending data through a satellite connection.
A less common use-case is the one wherein the moving platform facilitates the backhauling of a
remote cell site. For example, consider a complex of islands where traffic is backhauled through a
multi-hop chain of terrestrial microwave links. Assume that a cruise ship equipped with an iBN
approaches the coast of one of these islands. In that case, it can provide satellite backhauling to the
corresponding remote cell, as it is shown in Figure 2-8.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 29 of 131
SANSA-645047
D2.3
Figure 2-8: Moving platform facilitating the backhauling of a remote cell
2.3 Relation of Event periodicity to the Selected Use-cases The main characteristic of the SASNA network is re-configurability in terms of topology and
connectivity between links as it takes advantage of having the smart antennas along with the
satellite connected backhaul nodes.
The radio link failures as well as the congestion events are occurring the same way and with the
same frequency in both urban and rural deployment. As an example, the urban deployment involves
medium-to-high dense network in the normal working environment, and any additional event
normally add more traffic to the network and may cause network congestion. The events can be
classified into the following categories according to their occurring frequency. Without loss of
generality, the description will be based on the congestion link use case; link failure can be
interrupted similarly.
2.3.1 Periodic-occurring events These events include both the temporal and/or spatial periodicity. In terms of congestion related
with time, mainly the residence areas will have more traffic at night or non-working days, while the
industrial/offices zones and universities usually have high traffic demand during working hours.
Having part of the backhauling nodes with the satellite connection ability will prevent possible
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 30 of 131
SANSA-645047
D2.3
network congestion. An example of this case is explained in Figure 2-9 where the university
backhauling node A is connected to the Evolved Packet Core (EPC) via terrestrial networks. In case of
low traffic like non-working hours, the backhaul link can support the active nodes with no problem.
As the traffic increases in the working hours, the link to the EPC might be congested –the same in
case of link failure-, the backhauling node has two alternatives; a) to reach the EPC through the
backhauling node C as ACD or b) to get benefit from the satellite backhauling. The decision will
be part of the work to be done in other WPs depending on the type of the traffic, the load on the
node C, etc.
Figure 2-9: Periodic occurring events
2.3.2 Semi-periodic events
These types of events consider the case where high peak of traffic is generated due to regular
celebration, meeting or occasions. One expressive example is the sport stadium where the local
team plays home matches. In such a case, the stadium coverage would benefit from having one or
several hybrid backhauling nodes that are able to tackle the increment in the traffic and over some
additional bandwidth temporal need as depicted in Figure 2-10.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 31 of 131
SANSA-645047
D2.3
Figure 2-10: Semi-periodic occurring events
2.3.3 Rarely repeated events This category of events counts for the meetings or festivals with no prior special network node
planning. These types of events differ from the previous ones in that it’s possible that there is no
hybrid backhauling nodes placed in the event site like the stadium or the university examples. In this
case the network should react rapidly where the nearest hybrid backhauling node can off-load the
generated traffic. Example of this type of events is depicted in Figure 2-11 where the link CD is
congested due to the high traffic demand and the alternative CA link is used where the node A is
the nearest hybrid backhauling node.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 32 of 131
SANSA-645047
D2.3
Figure 2-11: Rarely repeated events
2.3.4 Seasonality Another aspect of the periodicity of the events is their seasonality. Summer and winter holiday
destinations face traffic fluctuations throughout the year that depend on the season. Summer
resorts are encountering high traffic demands during summer months whereas winter destinations,
like ski resorts are having the same problem during winter months.
SANSA through a mechanism similar to the one presented in 2.3.1 Periodic-occurring events will
offload traffic during peak months.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 33 of 131
SANSA-645047
D2.3
3. SANSA Scenario Definition
3.1 Overall strategy
There is a five axis presentation of the various elements that could jointly provide a well-defined
scenario as illustrated in Figure 3-1.
Figure 3-1: SANSA scenarios characterization
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 34 of 131
SANSA-645047
D2.3
The Ka-Band shared spectrum between terrestrial and satellite links. This could be the downlink or
both uplink and downlink. The main case in the scenario definition is the DL only spectrum. This is
because there is no need to extend bandwidth in the UL direction, due to FWD:RTN ratio increasing
(i.e. in excess of 6:1). Shared spectrum in the uplink also requires a change in regulations. The
scenarios, where interference in both uplink and downlink exists, are provided mainly for research
and interference mitigation techniques enhancement reasons and may be used in the future in case
of highly increased traffic demands to support the deployment of satellite terminals in the 27.8285 –
28.4445 GHz and in the 28.9485 – 29.4525 GHZ bands.
The satellite carrier bandwidth is also a very important factor regarding the scenario definition
process. The main three carrier types that will be examined through SANSA are the state of the art
(SOTA) carriers (54 MHz downlink and 9 MHz uplink carrier bandwidth). Since SANSA is an R&D
project aiming at providing backhaul solutions within the 2020 timeframe, the beyond state off the
art concerning carrier bandwidth should be taken into account. The other two references regarding
carrier are BATS, with a carrier bandwidth size of 421 MHz for the downlink and 21.7 MHz for the
uplink [64] and UWB with a carrier size of 230 MHz for the downlink and 9 MHz for the uplink 0.
Another important distinction between several scenarios is the type of deployment. Urban
scenarios are the scenarios with high node density, including small cells, high traffic requirements
but easy access to high speed optical fibre networks as well. On the other hand, rural scenarios are
for less populated areas with microwave link connections mainly and not easy access to fixed high
speed broadband networks. Another type of deployment is mobile platform deployment for trains,
ships, airplanes etc.
The CDN design is an emerging need for modern mobile networks as this type of traffic has an
increasing trend and will be the dominant type of traffic within the years to follow as already
mentioned in the state of art deliverable [2]. The existence of a CDN could be examined here as well
as how CDN cache should be fed, either through a terrestrial or a satellite link. We should also be
able to prepare nodes to facilitate the installation of such caching systems in case this will be
required in the future.
The last important factor in the scenario definition process is the type of the terrestrial links.
Antenna characteristics, channel characteristics, carrier characteristics and every link characteristic
should be clearly defined for each selected scenario.
In the next figure (Figure 3-2) there is a compact representation of the scenario selection strategy. A
matrix that takes into consideration all the above mentioned elements that can lead to a well-
defined scenario and places them in it.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 35 of 131
SANSA-645047
D2.3
Figure 3-2: Scenario selection strategy matrix example
3.2 Strategy for designing and implementing a terrestrial backhaul system
The strategy that is followed for the design, configuration and implementation of a terrestrial
backhaul link usually follows specific steps for designing the links; For the design, the implementer
considers: the broadband rate that needs to be supported, the resilience of the link so that the
connection between the nodes is not lost, flexibility in terms of resource handling and service
provisioning and compatibility between old and new installed equipment.
Topology examples (PtP, PtMP, backhaul of small cells)
The PtP and PtMP topologies have been provided in D2.1. In addition there are topologies which are
followed from small cells which are considered to be very interesting from the point of view of 5G
technologies.
Small cell deployment follows the Mobile Broadband (MBB) network densification for confronting
the data capacity crunch in the hot-spot areas. The small cell backhaul (and fronthaul) ecosystem is
an emerging technological area consisting by a large portfolio of interconnection media, including of
course wireless technology.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 36 of 131
SANSA-645047
D2.3
A typical use case is small cell network, which is deployed inside the urban street canyon, to be
connected to a nearby macro-cell site and from there entire traffic to be routed back to the core
network by utilizing the same transport infrastructure. In such an implementation the capacity
requirements of the macro-cell backhaul solution shall further increase.
3.3 SANSA selected scenarios Among the numerous scenarios, 6 scenarios in total, 2 for rural, 3 for urban and a moving base
station scenario have been selected to demonstrate the novelties and improvements SANSA
introduces to backhauling technologies with the use of satellite terrestrial hybrid networks as well as
with the use of new components. In Table 3-1 you can find a summarized presentation of the
selected scenarios according to the scenario selection strategy presented in the previous paragraph.
Table 3-1: Selected scenarios on the scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery Networks Type of deployment Terrestrial
links
Scenario
ID
DL
only
UL
+
DL
SoTa NG
HTS
UWB No
CDN
Sat-only
Multicast
Sat +
Terr
Multicast
Rural Urban Moving
base
station
18
GHz
28
GHz
Rural 1 + + + +
Rural 2 + + + + + + +
Urban 1 + + + + + +
Urban 2 + + + + + +
Urban 3 + + + + +
Moving
base
station
+ + + + +
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 37 of 131
SANSA-645047
D2.3
3.3.1 Rural case Rural areas all over Europe lack adequate broadband infrastructure. Fixed lines are usually slow and
mobile communications cannot provide the required speeds to support, hence improvements to
broadband connectivity are needed. The digital divide is present in mobile communications and the
forecast future trends show an increase in traffic demands in these areas too, making it very difficult
to cover future needs as they are forecasted by various reports [17]. The EC’s will to bridge that gap
and allow further growth and development in rural areas is supported by SANSA project and the
corresponding rural SANSA scenarios are built towards this direction.
Resilience, high speed satellite connections for traffic off loading and CDNs that can bring delivery of
specific interest closer to the user are some of the rural scenarios characteristics. The objective of
these scenarios is to further support backhauling networks, bringing as a result faster and more
reliable mobile networks. To achieve these goals will need extensive use of satellite connections as
well as use of innovative key enabling components such as smart antennas, intelligent backhaul
nodes and a hybrid network manager. Moreover, an extensive research on the interference
landscape and spectrum sharing is performed in order to help the designed system maximize
spectrum use.
3.3.1.1 Rural Scenario #1 This is the baseline to be examined throughout SANSA project. Instantiations of this could be used as
relevant rural scenarios for WPs to follow. The basic underlying idea of this scenario is the
coexistence in the same band in the downlink, the use of Next Generation HTS as a carrier
bandwidth reference, CDN cache fed by satellite multicast and terrestrial links at 18 GHz.
Table 3-2: Rural scenario #1 on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery Networks Type of deployment Terrestrial
links
Scenario
ID
DL
only
UL
+
DL
SoTa NG
HTS
UWB No
CDN
Sat-only
Multicast
Sat +
Terr
Multicast
Rural Urban Moving
base
station
18
GHz
28
GHz
Rural 1 + + + + +
3.3.1.2 Rural Scenario #2 R&D scenario UL & DL spectrum sharing should be taken into account. Interference mitigation
techniques to support the coexistence of terrestrial and satellite links for both uplink and downlink
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 38 of 131
SANSA-645047
D2.3
are needed. This is a useful scenario for future reference in case all allocated resources are used (not
in the near future) since channels of some tens of MHz have to use a band of 2 GHz. Any
interference problems in the present or in the near future would be a result of bad network planning
and resource allocation.
Table 3-3: Rural scenario #2 on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery Networks Type of deployment Terrestrial
links
Scenario
ID
DL
only
UL
+
DL
SoTa NG
HTS
UWB No
CDN
Sat-only
Multicast
Sat +
Terr
Multicast
Rural Urban Moving
base
station
18
GHz
28
GHz
Rural 2
(R&D) + + + + + + +
3.3.2 Urban Case In the urban scenario, connectivity is required at any place and at any time by humans in dense
urban environments, considering both the traffic between UEs and the Internet and direct
information exchanged between UEs. In terms of services run by the users we can find, besides
classical services such as web browsing, file download, email, social networks, there will be a strong
increase in high definition video streaming and video sharing. The particular challenge lies in the fact
that users expect the same quality of experience at their workplace, when enjoying leisure activities
such as shopping, or while moving on foot or in a vehicle. Furthermore, a particular aspect arising in
urban environments is that users tend to gather and move in “dynamic crowds”. A moderate
increase in terms of traffic demands can be considered cases like people waiting at a traffic light, or
at a bus stop. An intensive increase of traffic demands can be considered the case of a stadium
where thousands of people gather to watch a sport event.
To tackle these use cases in the urban scenario, network densification is important because of high
traffic volumes and their unpredictable traffic demand fluctuations. With large and unexpected
traffic demand fluctuations, the only solution in tree and ring topologies is to add higher capacity
links. In turn, as the increase of traffic demands can happen anywhere in the network, all the links in
the network would potentially require the addition of higher capacity links. On the other hand,
meshed solutions allow traffic to be load balanced over the topology to mitigate congestion, or using
alternative paths in case of link failure. A subset of all the backhaul nodes comprising the urban
scenario will include satellite backhaul link to reach the EPC. All backhaul nodes will include
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 39 of 131
SANSA-645047
D2.3
terrestrial links to reach the EPC. The UEs will share the backhaul links for backhauling. The SANSA
system will bridge the gap for the coexistence of terrestrial and satellite backhaul technologies,
reducing the costs derived from running these two independent wireless backhaul solutions.
For next WPs, we defined three different urban scenarios having in mind the current architecture
and the expected future ones. It is worth mentioning that in order to reflect the impact of the SANSA
it’s always recommended to compare the proposed scenarios with a reference scenario that
contains only fixed terrestrial antennas and with no satellites. Additionally, for all the considered
urban scenarios, medium-to-high density of base stations is assumed with small cell deployments
(e.g., mesh-like deployments). The scenarios are classified based on the timeline of the deployment.
The description of these scenarios is tackled in the next subsections.
3.3.2.1 Urban Scenario #1 This is the basic urban scenario that can be referred to as short-term urban one. In this scenario, the
spectrum is shared in the downlink transmission only where the actual SoTa is used as a reference
for the satellite carrier bandwidth. This scenario considers the terrestrial link at both 18 and 28 GHZ
with no CDN. With this, interference mitigation techniques will only take place at the terrestrial -
satellite downlink transmission.
Table 3-4: Urban scenario #1 on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery Networks Type of deployment Terrestrial
links
Scenario
ID
DL
only
UL
+
DL
SoTa NG
HTS
UWB No
CDN
Sat-only
Multicast
Sat +
Terr
Multicast
Rural Urban Moving
base
station
18
GHz
28
GHz
Urban 1 + + + + + +
3.3.2.2 Urban Scenario #2 This scenario reflects the medium-term deployment and is different from the previous one in both
the satellite carrier bandwidth and the level of the content delivery in the network. Having the same
spectrum sharing in the downlink only, this scenario differs in the sense that it uses the next
generation HTS as satellite carrier bandwidth reference. Satellite multicast is the proposed
technique for content delivery. The satellite user data demands for this scenario are higher than in
the previous one, as CDN will be deployed. Since the frequency band granularity of the satellite
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 40 of 131
SANSA-645047
D2.3
standard will be increased, the frequency allocation capabilities for the hybrid systems will be
increased.
Table 3-5: Urban scenario #2 on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery
Networks
Type of deployment Terrestrial
links
Scenario ID DL only UL
+
DL
SoTa NG
HTS
UW
B
No
CDN
Sat-
only
Multic
ast
Sat +
Terr
Multic
ast
Rur
al
Urba
n
Movin
g base
statio
n
18
GH
z
28
GHz
Urban 2 + + + + + +
3.3.2.3 Urban Scenario #3 This scenario will additionally consider the UL and DL spectrum sharing and therefore, use of
interference mitigation techniques will be mandatory. Moreover, as CDN in the terrestrial segment is
considered, multicast or multi-group multicast techniques shall be deployed.
Table 3-6: Urban scenario #3 on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery
Networks
Type of deployment Terrestrial
links
Scenario ID DL only UL
+
DL
SoTa NG
HTS
UW
B
No
CDN
Sat-
only
Multic
ast
Sat +
Terr
Multic
ast
Rur
al
Urba
n
Movin
g base
statio
n
18
GH
z
28
GHz
Urban 3 + + + + + +
3.3.3 Moving base station scenario Typically, in cruise ships one or more base stations / access points are installed to provide radio
access to the passengers. Traffic is backhauled with the help of a satellite link. Such a setup can
benefit from the capabilities of the SANSA architecture.
In this scenario, we consider a cruise ship equipped with an iBN, according to the SANSA concept, so
that both a terrestrial and a satellite link are available for mobile backhauling.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 41 of 131
SANSA-645047
D2.3
As it is shown in the following table, we assume spectrum sharing only in the DL and utilization of
the SoTA satellite carrier bandwidth.
Table 3-7: Moving base station scenario on scenario selection strategy matrix
Spectrum
sharing
Satellite Carrier
Bandwidth
Content Delivery
Networks
Type of deployment Terrestrial
links
Scenario ID DL
only
UL
+
DL
SoTA NG
HTS
UW
B
No
CDN
Sat-
only
Multic
ast
Sat +
Terr
Multic
ast
Rur
al
Urba
n
Movin
g base
statio
n
18
GH
z
28
GHz
Moving base
station + + + + +
The scenario is divided in two sub-scenarios, and each one of them is mapped to two use cases.
Cruise Ship Sub-Scenario #1: Anchored Ship
This scenario refers to an anchored ship. In this case, we can take advantage of the terrestrial
backhauling infrastructure. We distinguish between two use cases:
Use case #1: Only the terrestrial backhauling link is used, while the satellite link serves as
backup solution in case of a terrestrial link’s failure.
Use case #2: Both the terrestrial and the satellite backhauling links are used at the same
time, where the latter link is utilised for offloading the former link and enhancing the
backhauling capacity of the system.
Cruise Ship Sub-Scenario #2: Ship
This scenario refers to a ship. When the ship starts sailing, then the iBN should switch to the
satellite-only mode fast enough, such that no service disruption is encountered. Once the ship starts
its cruise, we distinguish between two use cases:
Use case #1: This use case refers to a cruise ship that is mostly located close to the shore. If
the ship is moving along the coastal line, then from time to time both backhauling options
will be available (e.g., when passing near an island). Hence, the iBN should (a) switch from
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 42 of 131
SANSA-645047
D2.3
the satellite-only mode to terrestrial-satellite mode or vice versa when the ship approaches
or leaving the shore, respectively; and (b) should make the optimum use of the backhauling
resources when both satellite and terrestrial backhauling options are available (e.g., use
both backhauling links to enhance capacity or use only the terrestrial link if the quality of the
satellite link is not sufficiently high).
Use case #2: This use case refers to a cruise ship that is mostly out in the open sea. While
the ship is at the middle of the ocean, then backhauling will be constantly provided through
the satellite link. When, on the other hand, the ship approaches the shore, then again both
backhauling options will become available as in use-case #1 of this scenario.
These use cases are illustrated in Figure 3-3.
Figure 3-3: Moving base station scenario
3.4 Scenarios General Characteristics
3.4.1 Frequency bands In the context of SANSA, D2.1 [1] studied the regulatory situation of the Ka-band in Europe in order
to assess the feasibility of deploying terminals for High Density (HD) applications in the Fixed
Satellite Service (FSS) accessing the full shared and exclusive civil Ka-band for backhauling traffic
from the cellular mobile networks.
Figure 3-4 illustrates the maximum bandwidth that could be available for the SANSA satellite
segment considering that HDFSS could be deployed in both exclusive and shared parts of the
spectrum.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 43 of 131
SANSA-645047
D2.3
Figure 3-4: Maximum available bandwidth for uncoordinated FSS stations based on CEPT implementations [1]
3.4.2 Spectrum sharing between different MNO Currently in the access mobile systems there are several schemes which are used for the spectrum
sharing. These can be summarized as:
License spectrum access (LSA): This is the mechanism where license spectrum is used among more
than one operator. The idea is to utilize unused spectrum of one operator by another MNO in case
that the main operator is underutilizing the spectrum. As an example, if one operator is using a
Frequency Division Duplex (FDD) system where the UL spectrum is underutilized, then it can be
borrowed by small cells to DL Time Division Duplex (TDD) traffic.
License assisted access (LAA): This is a mechanism invented by Ericsson and now 3GPP tries to
standardize it where LTE is used in unlicensed Wi-Fi bands in order to harmonically transfer traffic
from LTE into Wi-Fi technology.
Currently there is spectrum sharing at the access networks in some European countries in
Scandinavia. The backhaul networks usually are independent between different MNOs and no
sharing occurs between their networks mostly because of the competition that does not let easily
the operators to interoperate their networks. However, for the case where satellite and mobile
operators have to share their common spectrum in order to offer better services is very feasible and
advantageous to happen.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 44 of 131
SANSA-645047
D2.3
3.4.3 Standards
3.4.3.1 Terrestrial air interface Standards
The standards which are utilized for the PtP links are described mostly in the family of standards ETSI
EN 302 217, while the ones that refer to the PtMP links are described in the family EN 302 326-2.
These links have been very important to be used in the frequency bands from 28-29 GHz since their
use can alleviate constraints such as LOS conditions, capacity performance and also low cost of
ownership (TCO). The spectrum efficiency (bits per Hz) and the spectrum licensing fees in mmWave
backhauling are much lower than the ones used at macro frequencies. Currently most of the
backhaul systems are using FDD technologies and PtP modelling in order to be more robust to
frequency interference; however systems using TDD modelling are increasing. These are more
spectrum efficient and this trend will be accelerated in the future as capacity demands are
increasing. PtMP is used usually in urban areas where traffic demands are high.
Antennas
As the technology of the access network is moving into the arena of 5G, it is seen that large scale
antenna systems are used more often [67]. The LSAS scheme is the one where the base stations
have a large number of antennas attached to them and have been considered to provide system
energy efficiency and spectrum efficiency in future networks. Antenna arrays, where each antenna is
small (mm), and are directional at both TX and RX is an important consideration for the next 5G
MNs.
In addition, if beamforming is used then the number of transceivers can be reduced. This can lead to
lower TCO for the operator, which in turn is an additional factor that is considered for limiting the
power consumption and interference.
PHY interface
Time division duplexing (TDD) is a duplexing scheme that needs to be considered since it allocates
the bandwidth dynamically in time and uses lower cost radio equipment due to the no requirement
on filters.
3.4.3.2 Satellite air interface Satellite air interface standards have been presented in D2.2 [2] as part of the SoTA review of mobile
backhauling enabled by satellites. The main standards that have been studied are DVB-S, DVB-S2 and
DVB-S2X for the forward channel and DVB-RCS and DVB-RCS2 for the return channel.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 45 of 131
SANSA-645047
D2.3
Based on the state of the art hardware review for satellite backhaul presented in D2.2 [2], almost all
remote satellite routers receive DVB-S2 signals from the satellite gateway to the VSAT but utilize
proprietary air interfaces for the return link. In the context of SANSA the forward channel air
interface with the satellite will use DVB-S2 based on its widespread adoption but considering its
evolution, DVB-S2X which provides additional MODCODs, smaller roll-off factors and higher
modulation schemes. For the RTN channel we will adopt DVB-RCS2 to maintain compatibility with
ETSI standards.
3.4.4 Regulation In deliverable D2.1 [1] the different types of Licensing Procedures were presented; these are
presented here for completeness:
1. Individual licensing per link: This is the conventional link-by-link coordination, usually made
under administration’s responsibility. This is currently assumed to be the most efficient
method of spectrum usage for point-to-point links. Traditionally, backhaul links have been
registered on a “per link” license. Per link licensing has represented up to 65.5% of license
models in operation. Due to their directive antennas and narrow beam widths, per link
licensing is very common in PtP deployments, particularly in bands from 18 GHz to 42 GHz.
2. Block Spectrum: This is a common method for licensing point-to-multi point networks. The
user of the license is usually free to use the block at best to deploy its network. In some
cases, there might even be no limitation to the wireless communications methods used in
the block (e.g. PtP and/or PtMP, terrestrial and/or satellite, etc.). The assignment of the
block can be made through licensing (renewable, but not permanent) or through public
auction (permanent). Block spectrum assignment has been gaining traction for frequencies
from 28 GHz and above representing up to 20.7% of license models in operation. Per block
licenses are increasingly being used for PtMP links.
3. Lightly licensed spectrum: “A light licensing regime is a combination of license-exempt use
and protection of users of spectrum. This model has a first-come first-served feature where
the user notifies the regulator with the position and characteristics of the stations. The
database of installed stations containing appropriate technical parameters is publicly
available and should thus be consulted before installing new stations. A mechanism remains
necessary to enable a new entrant to challenge whether a station already recorded is really
used or not. New entrants should be able to find an agreement with existing users in case
interference criteria are exceeded”.
4. Shared License: This method is similar to lightly licensing, with a primary and secondary user
of the spectrum in a particular band, location and time.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 46 of 131
SANSA-645047
D2.3
5. Unlicensed spectrum: This method provides the most flexible and cheap usage but does not
guaranty any protection to interference.
In the SANSA context we will look both at the individual license per link and block spectrum
assignment as they are the most prevalent types of licensing nowadays. WP3 will explore the
interference mitigation techniques that can be employed for the different interference landscapes
formed by the two types of licensing.
3.5 Nodes distribution and generated traffic
3.5.1 Satellite user terminals deployment A great challenge regarding SANSA network is the deployment of satellite terminals within the
SANSA network. A deployment decision is usually based on a series of factors; these usually are
technological / performance factors, cost and regulatory or special problems that do not allow the
deployment of satellite terminals at a backhaul node. These factors are scenario based and the
deployment would not be the same in a rural, in an urban and in a moving base station scenario.
Ideally every node should have a satellite terminal, but this is not the most cost effective solution. In
a general sense the network should be a compromise between the ideal technical solutions, the
most cost effective solution and should always take into consideration factors that are not directly
related to technological or cost constraints (e.g. regulatory environment).
When it comes to the SANSA reference topologies design, it is very important to make a clear
distinction between the scenarios. It is very important to note that we consider after the regulatory
study conducted in D2.1 [1] that we can design a network that has no other restrictions apart from
technological and cost limitations. For the rural topologies, the main issues as we have already
discussed them in the use case definition are resilience, off-loading, CDN integration and standalone
cells. This means that satellite terminals will have to be placed at nodes where:
The backhaul node is the centre of a star network topology, so that the satellite link will help
with offloading and resilience where the star node is directly connected to the core network.
There is no different way to connect to the core network; this is the case of standalone cells.
Satellite terminals where Edge CDN caches are going to be placed. This will have to be
designed and implemented with the help of the CDN operator.
Every node that in case of link failure will lead to an access network problem, this can
happen to nodes that cannot establish a different terrestrial link with the use of the
steerable beam of a smart antenna.
In the case of urban deployments, the usual case is that nodes are connected to the core through a
superfast connection (fibre optic). This means that satellite links will be mainly used to support CDNs
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 47 of 131
SANSA-645047
D2.3
and offloading (especially for special events like concerts, football matches etc.). They can also
provide connectivity to urban areas where fibre networks have not been deployed yet and are not
expected to be deployed in the short term.
In the case of the moving base station scenario, satellite terminals will be needed in every node as
those platforms are expected to spend long periods of time having no access to terrestrial networks.
3.5.2 Macrocell deployments This section provides a traffic analysis modelling for current 4G/LTE networks and also a projected
analysis for 5G networks by assuming that heterogeneous networks will be used especially small
cells that are capable of generating a large amount of traffic. The analysis is performed for both
urban and rural cases although the first one is by far the most interesting and the one that generates
most traffic and revenue for the operator.
3.5.2.1 Urban scenario We can make an analysis of traffic by utilizing the characteristic physical parameters provided in the
document 36.214 [68]. Tables in section 7.1 propose basic data rates for current LTE systems and
also for the projected 5G systems. As it is seen in Figure 3-5 and Figure 3-6 we have considered
urban and rural scenario for base stations having 10 MHz bandwidth and two level backhaul
aggregation for the urban case while one level for the rural case. For current systems we can assume
that there are 3 cells per base station (BS) and one macrocell backhaul BS (MBBS) per 3 cells.
Table 3-8: Urban users’ traffic profile
Users/km2 Peak traffic/user Cell radius Cells/BS
4000 10Mbps 2km 3
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 48 of 131
SANSA-645047
D2.3
Figure 3-5: Traffic analysis for urban scenario
3.5.2.2 Rural scenario The same parameters as above also apply here, but each base station has one cell.
Table 3-9: Rural users’ traffic profile
Users/km2 Peak traffic/user Cell radius Cells/BS
50 3 Mbps 5 km 1
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 49 of 131
SANSA-645047
D2.3
Figure 3-6: Traffic analysis for rural scenario
3.5.2.3 Projected scenario for a 5G network In Table 3-10 there is a 5G network profile projected traffic by Nokia networks [66] and in Figure 3-7
there is a 5G network consisting of a number of small cells.
Table 3-10: Projected scenario for a 5G network (networks.nokia.com)
Users/km2 Peak traffic/user Small Cell radius Cell edge traffic
4000 1 Gbps 150 m 100 Mbps
Figure 3-7: 5G network consisting of a number of small cells
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 50 of 131
SANSA-645047
D2.3
3.5.3 Satellite and terrestrial content delivery traffic profile In order to study this problem, a traffic model based on the multimedia popularity is defined in this
section. The popularity is measured based on the number of requests and can be described though a
probability function. A widely-used abstraction for this function is the Zipf law, which is given by [3]:
In more detail, if we ordered the files from most to least popular at a given point in time, then the
relationship governing the frequency at which the file of rank i will appear, is given by the equation
presented above. Consequently, the probability of a request occurring for file i is inversely
proportional to its rank, with a shaping parameter α. A detailed analysis on how to choose the exact
value of α is described in [4]. The following figure (Figure 3-8) provides an example of the Zipf law:
Figure 3-8: Zipf law
3.6 Interference landscape In this section we consider all the possible sources of interference at the backhaul node level and
assess whether they are relevant to the SANSA solution. It is important to note that in the context of
the project we refer to co-channel interference i.e. the interference caused by two or more
transmitters using the same frequency.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 51 of 131
SANSA-645047
D2.3
Considering the SANSA network as a system we distinguish interference in intra- and inter-system
interference. Intra-system interference is caused by transmitters that belong to the SANSA system,
whereas inter-system interference is caused to the SANSA nodes from external transmitters using
the same frequencies. The possible sources of interference are illustrated in Figure 3-9.
3.6.1 Intra-system interference The intra-system interference can be either between the satellite and terrestrial links or only
between the terrestrial links.
3.6.1.1 Interference between satellite and terrestrial links Based on the spectrum sharing schemes between the satellite and terrestrial links that have been
identified for the different scenarios in Table 3-1, the possible causes of interference are:
From eNodeB (eNB) to neighbouring satellite terminal (DL): This is the main case of interference relevant to all the defined scenarios where spectrum sharing is implemented between the satellite DL and the terrestrial links. WP3 will examine techniques to protect the satellite terminal from the interference of an adjacent eNB operating at the same frequency.
From eNB to neighboring satellite terminal (UL): This case of interference is valid if there is shared spectrum between the satellite uplink and the terrestrial links which is not possible under the current regulatory environment.
From satellite terminal to neighbouring eNodeB (UL): This case of interference is valid if there is shared spectrum between the satellite uplink and the terrestrial links which is not possible under the current regulatory environment. There are only two long term scenarios (Rural 2 and Urban 3) that are using this case of spectrum sharing and are scenarios that are not supported by the current regulatory environment but will the coexistence and deployment of terrestrial and satellite links in Ka band and specifically in 27.8285 GHz – 28.4445 GHz and 28.9485 GHz – 29.4525 GHz.
From satellite terminal to neighbouring eNB (DL): In this case there is no interference from the satellite terminal to the eNB as the satellite terminal is only receiving at the shared frequency and not transmitting.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 52 of 131
SANSA-645047
D2.3
Figure 3-9: Interference landscape in the SANSA system
3.6.1.2 Terrestrial to terrestrial link Apart from the interference mitigation techniques that enable spectrum sharing between the
satellite and terrestrial links, it is important to investigate solutions for interference between
terrestrial links in the SANSA network. The main technique used for this will be appropriate
frequency reuse schemes depending on available spectrum resources.
3.6.2 Intersystem interference Inter-system interference can be caused by terrestrial and/or satellite links operating at the same
frequencies from different operators. WP3 plans to briefly study this scenario for the case of block
spectrum assignment and per link license.
3.7 Benchmarking topologies
3.7.1 Rural topologies In this report, we choose a typical topology from the Finnish 28 GHz database obtained by the
University of Luxembourg. Note that we don’t have any claim regarding if the chosen topology
represents a backhaul network necessarily, however it provides a good example to perform some
interference and availability analysis. As shall be shown later, such an analysis provides a good
benchmark for the underlying requirements in SANSA reference system.
In the next subsection, first we present the chosen topology located in Helsinki. Note that even
though Helsinki is a big city, in general it consists of small height buildings. Moreover, the suburban
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 53 of 131
SANSA-645047
D2.3
area of Helsinki examined here presents quite sparse building population which is a typical
characteristic of a rural environment.
3.7.1.1 Topology Example: Helsinki The selected topology is depicted in Figure 3-10. As we can see this topology consists of a number of
interconnected star topologies. This topology based on the actual data in the database is composed
of 28 links and 15 actual locations (nodes). In Table 3-11, we may see the underlying parameters
defining each link as derived from the Finnish database. Further, Table 3-12, presents the location of
each node. It should be noted that all depicted links are bidirectional.
Figure 3-10: Example backhaul topology obtained from the Finnish 28 GHz database, Helsinki.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 54 of 131
SANSA-645047
D2.3
Table 3-11: Detailed information of each link
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 55 of 131
SANSA-645047
D2.3
Table 3-12: The location of nodes
Table 3-13: Connectivity matrix
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 56 of 131
SANSA-645047
D2.3
3.7.1.2 Benchmark SINR distribution
In this section, we employ the ITU-R 452-16 interference modelling, including the free space loss as
well as the diffraction loss based on the Bullington model to derive the SINR of each receiver based
on the coordinated frequency plan in Table 3-11. This result which will be considered as the
benchmark SINR distribution is presented in Figure 3-11. We can see that all the receivers
experience SINR > 42dB, while a significant number of them experience SINR > 60dB. This is to be
expected since the benchmark topology is the outcome of careful network planning through link
registration by the national regulator.
Figure 3-11: Benchmark SINR distribution of the coordinated frequency plan in Table 3-11
3.7.1.3 SINR and interference analysis: Aggressive Frequency Reuse
In this section, we will move towards the concept of shared access promoted by SANSA, and analyse
the performance of each link, when all employ the same frequency plan. It should be noted that this
is a worst case scenario and less aggressive frequency reuse could be used in practice. We further
would like to estimate the number of required nulls to be produced by each SANSA smart antennas
to tackle the strong interferers. Here, we define the strong interferers based on the ITU-R
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 57 of 131
SANSA-645047
D2.3
recommendations [5]. An interferer is considered to be harmful if the level of received interference
increases the noise floor by 10%.
Figure 3-12 depicts the SINR distribution of the links when full frequency reuse is employed. We can
note that in this case the lower value of SINR is reduced to around 22dB from 42dB in the
benchmark model. Further, none of the links experience SINR > 48dB. This is explained by the
increased internal interference among the SANSA links. It should be noted that further degradation
might be experienced if inter-system interference from external links is taken into account.
To evaluate the number of strong interferers in each link and thus the required number of nulls in
each SANSA designed smart antenna, we can look at Figure 3-13, where the distribution of the
number of required nulls is presented. Based on this figure, we can deduce that each node needs to
be able to produce 7 nulls in average. It is expected that if less aggressive frequency is used in
combination with carrier allocation optimization, a smaller number of nulls will be required. Thus,
this number can be considered as an upper bound requirement in SANSA smart antenna techniques.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 58 of 131
SANSA-645047
D2.3
Figure 3-12: SINR distribution of the links with aggressive frequency reuse
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 59 of 131
SANSA-645047
D2.3
Figure 3-13: Distribution of the required number of nulls with aggressive frequency reuse
3.7.1.4 Terrestrial to Satellite User Terminal
In this section the impact of employing satellite links for backhauling is studied for the example
topology of Section 3.7.1.1. To that end we replace the terrestrial terminals of nodes 1, 8 and 10 of
the existing topology with satellite ones and study the interference from the remaining terrestrial
nodes. The provided analysis can be considered as a worst case scenario since we assume full
frequency reuse and the satellite terminals are placed on the nodes with the largest number of links.
It is assumed that the satellite terminals on these three nodes are pointing to the HYLAS 2 and 31.7𝑜
East. Based on the node’s position, the satellite antennas should have elevation and azimuth 21.7𝑜
and 176𝑜 respectively in order to point in this satellite group. The satellite link characteristics that
were used are the ones of Section 1.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 60 of 131
SANSA-645047
D2.3
Let us study now the interference generated to the satellite terminals by the terrestrial ones. To that
end, in Figure 3-14 we plot the aggregated interference that each one of the satellite terminals is
experiencing. As we can see, for the specific topology and the elevation angles of the satellite
terminals, the interference levels are very low so that there is no need in placing nulls in their
direction. The latter is also verified by Figure 3-15 where we plot the SINR of each one of the satellite
nodes along with the SNR. As it is shown, each node experiences interference that decreases its SINR
less than the 10% of the SNR floor, so based on the ITU-R recommendations the interference can be
considered as non- harmful [5]. Note that in cases where the satellite noses are experiencing
harmful interference from the terrestrial ones, the latter may apply transmit beamforming
techniques in order to null out the interference in the satellite ones.
Figure 3-14: Aggregate Interference on the satellite terminals due to terrestrial transmissions
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 61 of 131
SANSA-645047
D2.3
Figure 3-15: SINR of the satellite terminals
3.7.1.5 Satellite User Terminal to Terrestrial
Let us now move to the uplink scenario where we are interested for the interference to the
terrestrial terminals generated by the satellite ones. It is assumed that the satellite terminals have
the same azimuth and elevation parameters to the ones of Section 3.7.1.4. The satellite link
characteristics that were used are given in Table 3-14.
In Figure 3-16, we plot the distribution of the number of nulls required per terrestrial link due to the
transmission of the satellite terminals. The calculations are based again on the ITU-R
recommendations [5]. As it is shown, there is requirement for at most one null in only 9 links. The
latter result is very promising since the required number of nulls can be easily handled by the
antenna infrastructures of the typical backhaul nodes by the application of standard receive
beamforming strategies.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 62 of 131
SANSA-645047
D2.3
Table 3-14: Simulation parameters for uplink
Figure 3-16: Distribution of the number of nulls per link during uplink
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 63 of 131
SANSA-645047
D2.3
3.7.2 Urban topologies: The Vienna case In contrast to the rural case, it is expected that in urban scenarios a denser link deployment is
performed due to the extremely high forecasted traffic increase. As a result, a high number of
interfering will be present at each receiver. In addition, as a consequence of the population
densification and its corresponding massive cell (pico, small and macro) deployment, the distances
between nodes will be 1-2 Km maximum, leading to a substantial increase of the interference power
level.
The scope of this section is to provide a scenario example for the urban case. With this definition,
the main interfering cases are described so as an example urban topology.
3.7.2.1 Methodology For the description of the wireless backhaul topology we have used the software iQ-linkXG [6] whose
free version allows to study up to 22 microwave links. This software has available a numerous set of
radio transceivers, antennas and other additional features. After the nodes are deployed within a
map, iQ-linkXG is able to compute the interfering signals, link budgets and terrain profile taking into
account buildings.
This software only has available links at 28 GHz so that, numerical results at 18 GHz can be
extrapolated considering a proportional antenna size in order to compensate the path loss.
Moreover, we will assume full frequency reuse among the links in order to obtain a large bandwidth.
Unfortunately, the satellite link is not straight forward to be included in the software. Consequently,
we evaluate the interference power levels considering an ad-hoc example which will indicate the
major sources of interference.
3.7.2.2 Topology
The topology can be observed in Figure 3-17.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 64 of 131
SANSA-645047
D2.3
Figure 3-17: Network Topology for urban deployments
This topology has been obtained considering the city of Vienna. In Figure 3-18, on can be observe
the topology overlaid on Google Earth data.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 65 of 131
SANSA-645047
D2.3
Figure 3-18: Network Topology over Google Earth picture.
The deployment has been done considering the specific characteristics of this city and the future
deployment of small cells.
First, Vienna has a tower in the city surroundings that could be equipped with several microwave
links (Figure 3-19). We will consider that this tower has direct connection with the EPC.
Figure 3-19: Tower details
Since Vienna city has a football stadium, we consider that it will be equipped with two small cells
whose backhaul is connected through the microwave wireless backhaul (Figure 3-20).
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 66 of 131
SANSA-645047
D2.3
Figure 3-20: Stadium details
Finally, different nodes have been considered depending on the area population (city centre,
business centre).
3.7.2.3 Interference analysis
Using the interference analysis tool in iQ-link, we identify the major sources of interference. These
examples are described in the following subsections.
3.7.2.3.1 Methodology
For quantizing the interference power levels we have considered that all links (both terrestrial and
satellite) operate at the same frequency band, in the 28 GHz. As discussed previously, the results are
valid for both the 18 GHz and 28 GHz one as the antenna and power can be scaled accordingly in
order to meet the SNR requirements.
For the sake of simplicity we have assumed the same antenna pattern for both links (terrestrial and
satellite). This is described in Figure 3-21.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 67 of 131
SANSA-645047
D2.3
Figure 3-21: Antenna discrimination angle
Remarkably, the discrimination is the same for both the elevation and azimuth angles.
For the 28 GHz case, we considered the Alcatel-Lucent 9400 AWY model [7]. The EIRP for the
terrestrial node is assumed to be 50 dBm (in 30 MHz). Moreover, we consider that all
communications take place in the same frequency bin. For the evaluation of the interference in the
18 GHz band, we have considered ETHEFLEX transceiver [8] operating with a bandwidth of 30 MHz
and a target bit rate of 138.7 Mbps. This leads to a required SNR of 27 dBs.
3.7.2.3.2 Case 1: Co-located microwave links
This case considers the interference received in the link node11-tower from other transmitters
located at the tower. Figure 3-22 summarizes the interference power levels received at the
transceiver detecting the signals from node11.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 68 of 131
SANSA-645047
D2.3
Figure 3-22: Interference analysis Case 1
It can be observed in the previous plot that the resulting C/I are dramatically low. This case shall be
considered by WP3 in order to provide a solution for nodes receiving and transmitting information in
the same band. As a first approach, the different links could be allocated in separate frequency bins.
3.7.2.3.3 Case 2: Terrestrial Links
This case considers the interference received from terrestrial links located at different nodes. The
following plot depicts the histogram of the SINR values of the described topology (Figure 3-23).
Figure 3-23: Interference analysis Case 2
The values centred in -40 dB range correspond to the co-located links described in the previous
subsection. The ones that appear in the right on the diagram correspond to general links which
suffer from certain interference power levels. Indeed, even though certain links operate in a very
high SNR, there are others whose SINR is close to 0 dBs.
-60 -40 -20 0 20 40 600
1
2
3
4
5
6
7
8
9
SINR [dBs]
Lin
ks
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 69 of 131
SANSA-645047
D2.3
Consequently, the interference mitigation techniques are mandatory in order to achieve high SINR
values so that the future traffic demands are properly supported.
3.7.2.3.4 Case 3: Satellite user terminal to terrestrial node This case assumes spectrum sharing in the 28 GHz band for a satellite user terminal and a terrestrial
node. The aim of this analysis is where the terrestrial node receiver shall perform any kind of
interference mitigation.
Table 3-15: Interference data for Case 3 interference analysis
Satellite return link values Terrestrial link values
Frequency 28 GHz Frequency 28 GHz
EIRP 76 dBm EIRP 53 dBm
Bandwidth 1.28 MHz Bandwidth 30 MHz
FSL in 1Km 117.5 dB FSL in 1 Km 117.5 dB
RSL at the terrestrial
node receiver
26.5 dBm RSL at the terrestrial
node receiver
-30.5 dBm
Sensitivity 64 QAM at
10E-6 BER
-74 dBm
Minimum SINR for 64
QAM
27 dB
As presented in the Table 3-15, we will assume an example of a 1 Km terrestrial link. The satellite
values are obtained from Avanti input in Section 2. Bearing in mind that the link requires to operate
at 138.7 Mbps, the minimum SINR is 27 dB. Considering this operative point, the maximum received
interference can be -62 dBm. It has been considered that both the satellite user terminal and the
terrestrial node are located in 1Km distance (see Figure 3-24).
For achieving the maximum value of interference power level, the antenna discrimination of both
the terrestrial receiver and the satellite user terminal shall be of 88.5 dB. This value indicates that
interference mitigation techniques are mandatory for this interference scenario.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 70 of 131
SANSA-645047
D2.3
Figure 3-24: Case 3 interference at 28 GHz
3.7.2.3.5 Case 4: Terrestrial node to satellite
This section considers the case of interference from terrestrial node towards the satellite received
signal at the 28 GHz band. The following table describes the considered values.
Table 3-16: Interference data for case 4 interference analysis
Satellite return link values Terrestrial link values
Frequency 28 GHz Frequency 28 GHz
EIRP 76 dBm EIRP 13.7 dBm
Bandwidth 1.28 MHz Bandwidth 1.28 MHz
FSL in 37000 Km 212.7 dB FSL in 37000 Km 212.7 dB
RSL at the satellite -136.7 dBm RSL at the satellite -199 dBm
Required C/N for 8PSK
6/7 DVB-RCS BER 10E-
6
10.7 dB
It is important to remark that it is required to recompute the EIRP for the terrestrial link as we are
considering a lower bandwidth (1.24 MHz) dictated by the satellite link. As it can be observed, even
though the terrestrial link points directly to the satellite, the received signal levels are extremely
unbalanced (62.3 dB). Due to that, for the usual case, this interference will not likely occur.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 71 of 131
SANSA-645047
D2.3
Figure 3-25: Case 4 interference at 28 GHz
3.7.2.3.6 Case 5: Terrestrial node to satellite user terminal
This section studies the case where a terrestrial node interferes with the reception of the satellite
signal by a satellite user terminal. The band considered is 18 GHz. The next table (Table 3-17) details
the parameters used.
Table 3-17: Interference data for case 5 interference analysis
Satellite forward link values Terrestrial link values
Frequency 18 GHz Frequency 18 GHz
EIRP 84.33 dBm EIRP 53 dBm
Bandwidth 54 MHz Bandwidth 30 MHz
FSL in 37000 Km 212.7 dB FSL in 1 Km 117.5 dB
RSL at the satellite
user terminal
-98.3 dBm RSL at the satellite
user terminal
-64.5 dBm
Required C/N for 8PSK
6/7 DVB-S2 BER 10E-6
6.6 dB
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 72 of 131
SANSA-645047
D2.3
Considering again that the terrestrial link and the satellite user terminal are located within a distance
of 1 Km. From the above presented table it is evident that interference mitigation techniques are
mandatory considering the different RSL. Indeed, the maximum received interference power level
for operating at 86.9 Mbps is -119.04 dBm. As a result, the terrestrial node shall implement an
interference mitigation technique in order to provide a discrimination angle of 92.5 dB.
Figure 3-26: Case 5 interference at 18 GHz
3.7.2.3.7 Case 6: Satellite to terrestrial node
This section describes the interference values between the satellite and the terrestrial node. The 18
GHz band is considered. The following table describes the considered parameters.
Table 3-18: Interference data for Case 6 interference analysis
Satellite forward link values Terrestrial link values
Frequency 18 GHz Frequency 18 GHz
EIRP 84.33 dBm EIRP 53 dBm
Bandwidth 54 MHz Bandwidth 30 MHz
FSL in 37000 Km 212.7 dB FSL in 1 Km 117.5 dB
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 73 of 131
SANSA-645047
D2.3
RSL at the terrestrial
node
-98.3 dBm RSL at the terrestrial
node
-64.5 dBm
Required SINR for 64
QAM BER 10E-6
27 dB
As described in Case 3, the maximum interference power level at the terrestrial node is -62 dBm. For
this case, it can be observed from the table that even though the terrestrial link points to the
satellite, the satellite received signal is below the threshold. As a result, this case can be considered
as a secondary in terms of importance for the interference mitigation techniques development.
Figure 3-27: Case 6 interference at 18 GHz
3.7.3 Moving Base Station Maritime mobile communications present unique challenges regarding the application of mobile
backhauling. In SANA, we envision an advanced system wherein an iBN is installed on the ship to
provide both terrestrial and satellite mobile backhauling capabilities, as it is shown in Figure 3-28.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 74 of 131
SANSA-645047
D2.3
Figure 3-28: Cruise ship topology
The differences of this scenario with the urban and rural scenarios are the following:
When the ship is at open sea, there is no interference from neighbouring terrestrial links.
Regardless whether the ship is anchored or sailing, we have to take into account fading
caused by the movement of the ship or / and the reflection of the transmitted signal from
the sea surface.
3.7.3.1 Fading due to Reflection of the Signal from the Sea Surface
The received signal consists of the sum of a direct component and reflected waves. Multipath
reflections from the sea surface are composed of a specular (coherent) component and a diffuse
(incoherent) component. The former component is phase coherent with the direct wave, whereas
the phase of the latter component varies randomly with the motion of the sea waves. The coherent
component is produced by specular reflection from the sea, while the incoherent component is
generated by reflections from multiple statistically independent points on the sea surface. The
magnitude of these components depends on the roughness of the sea surface and the elevation
angle. Their effect is more prominent for low elevation angles. For calm sea surface, the specular
component dominates the radio propagation. This is in contrast with the typical situation in land
mobile channel modelling, where the specular ground reflection is commonly ignored. However, its
impact vanishes rapidly as the roughness of the sea surface increases. The diffuse component is not
affected that much by the roughness of the sea surface; however, in general its magnitude increases
as the roughness of the surface increases.
Having that in mind, we model the received signal as:
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 75 of 131
SANSA-645047
D2.3
,r t s t g t d t n t
where s t , g t , d t and n t
where , , and are the direct, specular, diffuse and additive Gaussian noise components,
respectively. In complex baseband representation we have:
2
2
2
Re
Re ,
Re
c
c G
c k
j πf t φ
j πf t φ
j πf t φk
k
s t Se
g t Ge
d t D e
where S , G , kD and φ , Gφ and kφ
where , , and , and are the magnitude and phase of the direct, specular and diffused components,
respectively, and cf is the carrier frequency.
These signals can also be expressed in terms of their in-phase and quadrature components as:
cos2 sin2
cos2 sin2
cos2 sin2
cos2 sin2
i c q c
i c q c
i c q c
i c q c
r t r t πf t r t πf t
s t s t πf t s t πf t
g t g t πf t g t πf t
d t d t πf t d t πf t
where the subscripts i and q denote the in-phase and quadrature signals, respectively. When the
number of diffuse components is large, then id and qd and are modelled as Gaussian random
variables, as a consequence of the central limit theorem. Typically, it is assumed that they have zero
mean and common variance dσ .
The in-phase and quadrature components of the received signal can be expressed as:
i i i i
q q q q
r t s t g t d t
r t s t g t d t
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 76 of 131
SANSA-645047
D2.3
or, equivalently:
cos cos
sin sin
i G i
q G q
r t S φ G φ d t
r t S φ G φ d t
The envelope and the phase of the received signal are:
1 22 2
1tan
i q
i
q
R t r r
rψ t
r
In the special case where 0S G , the probability density function (PDF) of the envelope follows
the Rayleigh distribution:
2
222
d
R
σ
d
Rp R e
σ
In the other extreme case where 0S , it follows a Ricean distribution:
2 2
2202 2
,d
R S
σ
d d
R RSp R e I
σ σ
where 0I
is the modified Bessel function of first kind and zero order. The Ricean factor K depends on the
elevation angle.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 77 of 131
SANSA-645047
D2.3
In most cases, the sea surface is considered to be rough enough to allow us to ignore the specular
component. Hence, typically the channel is modelled as a Ricean fading channel.
Unfortunately, there are very few published works on the characterization and measurements of
multipath fading caused due to sea surface reflections. Even worse, these studies consider
frequencies 10 GHz. For example, the methodology presented in [9] and the corresponding results
regarding the fade depth are only valid in the frequency range 0.8 – 8GHz. Similarly, the model and
measurements regarding the fade depth and fade duration for both calm and rough sea surfaces
described in the [10] refer to the frequency band of 1.5GHz.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 78 of 131
SANSA-645047
D2.3
4. Definition of end-to-end SANSA Architecture
4.1 Scope This section discusses the system architecture of the SANSA solution. The first subsection provides
an overall description of the end-to-end system architecture including access, transport, and core
network. This architecture aims at covering all the use cases and scenarios defined in previous
chapters. The second subsection details the transport network architecture (TNA), which is the main
focus of the SANSA project. The architecture of two key enabling SANSA components, the intelligent
Backhaul Node and Hybrid Network Manager is also presented here. Lastly, the architecture for the
moving base station scenario is highlighted at the end of the section.
4.2 SANSA End-to-end System Architecture The SANSA end-to-end system architecture encompasses the LTE-based Radio Access Network
(RAN), the transport network (where the research impact resides), and the Evolved Packet Core
(EPC), also referred to as core network.
The SANSA Access Architecture encompasses the mobile user equipment (UEs) and base stations
(BSs), which can be either macrocells (eNodeBs/eNBs) or small cells (Home eNodeBs/HeNBs). It is
important to note that both macrocells and small cells can be embedded in intelligent Backhaul
Nodes (iBNs) and Backhaul nodes (BNs), as shown in Figure 4-1.
Since SANSA focuses on the transport network, all the 3GPP signalling procedures in the EPC (see
Figure 4-1) and the RAN are adopted without modifications. A detailed description explanation of
the main building blocks and interfaces can be found in [11] . In summary, the main Evolved Packet
Core (EPC) building blocks are:
The Packet Data Network Gateway (PDN-GW) is the entity in charge of connecting the UE to
the external network. The P-GW supports the establishment of data bearers with the Serving
Gateway presented below such as the assignment of IP addresses for the UEs.
The Serving Gateway (S-GW) is the entity in charge of forwarding user plane packets
through the EPC. In particular, it receives tunneled packet from the UE and re-tunnels them
to the PDN-GW.
The Mobile Management Entity (MME) is the entity in charge of handling UE connectivity
and mobility. In particular, amongst other functions, the MME keeps track of the location of
the UEs.
The Home Subscriber Server (HSS) is the functional entity storing subscription data for UEs
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 79 of 131
SANSA-645047
D2.3
(e.g., user profile, access restrictions for roaming etc).
These entities exchange control plane procedures with the (H)eNBs by means of the S1-MME
interface. In particular, the S1-AP [12] (Application Protocol) provides the necessary control message
signalling between the (H)eNBs and the MME with bearer establishment and mobility management
being some of the network functions it performs.
Regarding user plane traffic, they are tunnelled through various functional entities in the EPC by
means of GPRS Tunnelling Protocol User Plane tunnels (GTP-U) [13]. The S1-U interface provides
user plane tunnelling between the (H)eNodeBs and the S-GW. The GTP-U protocol tunnels user data
between (H)eNodeBs and the S-GW, and between the S-GW and the PDN-GW. The goal of the GTP-U
protocol is to encapsulate IP traffic in flow specific tunnels to provide QoS differentiation. The S5
interface provides user plane tunnelling between the S-GW and the PDN-GW.
It is in the Transport Network where SANSA introduces its main research novelties. In the next
section, we describe the main entities optimized and introduced by the SANSA network.
Figure 4-1: SANSA System Architecture
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 80 of 131
SANSA-645047
D2.3
4.3 SANSA Transport Architecture The SANSA system focuses on evolving the Transport Network Architecture (TNA). The TNA
describes the network in charge of transporting data between the UEs and the EPC. The SANSA TNA
combines both satellite and terrestrial transport architectures. In this sense, the SANSA transport
network architecture is composed by the following key elements.
The iBN extends the internal architecture of traditional BNs by introducing new functional blocks and
interfaces for the proper management of backhaul satellite and terrestrial resources. Amongst other
functions, the iBN will embed routing, traffic classification, and energy management functions. The
iBN will operate on short to medium timescales and is reconfigurable by the HNM. It will encompass
interfaces to other iBNs, and to the EPC either directly (with a radio link) or through other iBNs.
Finally, any iBN may include a direct connection to the EPC through the satellite network. Note that
the mobile network layer information (e.g. traffic flow from a UE) traverses the iBNs encrypted. We
assume that the iBN is a trusted component by the UEs and EPC, which has enough processing
capabilities, can decrypt mobile network layer information (e.g. traffic class information) tunnelled
through the S1 interfaces to conduct certain functions such as traffic classification and routing of
traffic flows.
The Backhaul Node (BN) is a legacy entity embedding the H(eNB) in charge of carrying transport
traffic to the EPC. It neither presents routing, traffic classification, and energy management
functions. A special case of BN is that of the Mobile Base Station (MBS). The MBS is a BN that
includes mobility capabilities (e.g. a BS in a train). It is an optional element in the SANSA scenarios.
The Satellite (SAT) is a component enhanced by SANSA due to its smooth integration in the
reconfigurable terrestrial transport network. The SAT will encompass an interface to the EPC, and an
interface to iBNs. The interface between the iBN and the SAT allows the system to access the
satellite link status and use the data for traffic classification, routing, topology reconfiguration, and
interference management between the satellite and terrestrial links. Information such as satellite
carrier frequency, channel bandwidth, available data rate and link availability is constantly
monitored by the HNM.
The HNM is a new entity introduced by SANSA which includes functionalities to manage not only
satellite but also terrestrial backhaul resources. Based on global network information view based on
its monitoring capabilities, the HNM is in charge of configuring the topology formed between the
iBN nodes and their connection and configuration of the satellite resources. In this context, it can
configure backhaul resources embedded in terrestrial iBNs, MBSs, and satellite resources. It
operates on long and medium timescales.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 81 of 131
SANSA-645047
D2.3
4.4 Intelligent Backhaul Node (iBN) Architecture The iBN is the component that implements the SANSA ground (or terrestrial) transport network. The
iBNs, which are distributed throughout the SANSA meshed network, distribute and forward the user
traffic, and perform network decisions on a short (e.g. routing) and medium (e.g. energy efficiency)
time-scale basis. The iBNs integrate the following main components:
(H)eNB: This generally refers to an LTE base station, or a low-power LTE base station (e.g.
small cell). It is connected to the EPC through a backhaul network.
Routing: This function includes the routing algorithm and is in charge of distributing the
traffic among the different terrestrial and satellite modem interfaces.
Traffic Classification: The Traffic Classification function is in charge of determining the
mapping of traffic flows to the hybrid backhaul resources used to transport them.
Energy Efficiency: This function is in charge of controlling access and backhaul energy
consumption, therefore reducing operator’s OPEX while satisfying traffic demands.
Modems: An iBN includes several terrestrial and/or a satellite modem.
Antennas: According to the type of modem terrestrial smart antennas and/or a satellite
antenna provides the air interface of the iBN.
Beamforming network: This element mixes the outputs of the modems before passing it to
the smart antenna.
A representation of the different components in the iBN can be shown in Figure 4-2:
Figure 4-2: Representation of the iBN components
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 82 of 131
SANSA-645047
D2.3
4.5 Hybrid Network Manager (HNM) Architecture The HNM is the central element of the SANSA network. This manager aggregates all the network
resources and performs long and medium time-scale configuration changes. According to Figure 4-3,
it includes the following main functions:
Configuration management: This function is in charge of reconfiguring the iBNs in the
network. Also, it is used for distribution of network topology to the nodes. Configurable iBN
items are for instance the terrestrial modems and antennas.
Events management: This component monitors the network nodes and determines the
state of SANSA network.
Topology management: This module performs topology calculations to restore the hybrid
network upon node congestion or failure events. As input, it receives new network states,
and produces new topologies, forwarding them to the configuration management function.
The HNM external components are:
Radio Environment Map: This component calculates interference levels and performs the
carrier allocation. It generates the data that can be later used by the topology management
to calculate effective network throughput.
Satellite Ground Segment: This component is composed by different tools for managing the
satellite network, such as the NMS, the satellite Hub, the OSS, and the BSS for SP customers.
Figure 4-3: Representation of the HNM functions
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 83 of 131
SANSA-645047
D2.3
4.6 Moving Base Station Architecture As Moving Base Station (MBS) we refer to radio access and SANSA-complied mobile backhauling
infrastructure installed on a moving platform, such as a cruise ship. By SANSA-complied mobile
backhauling infrastructure we mean an iBN which enables both satellite and terrestrial mobile
backhauling.
The radio access segment is comprised of either Wi-Fi hotspots or LTE (H)eNBs (e.g. small cell). These
are connected through optical fibres to the iBN. Often, a concentrator is placed between the radio
access nodes and the iBN. Its role is to gather the traffic from the radio access nodes and transfer it
to the iBN. The radio access network may have its own network management system (NMS) for
operation and maintenance (O&M) purposes. Such a setup is shown in Figure 4-4.
Figure 4-4: Moving base station architecture
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 84 of 131
SANSA-645047
D2.3
5. Definition of Key Performance Indicators
5.1 Objectives The objective of this section is to define the end-to-end key performance indicators (KPIs) that are
going to be used in order to evaluate the performance of the SANSA backhaul network solution.
The KPIs defined for SANSA refer to the overall performance of the solution and reflect the
objectives of the project focusing on the areas of improvement. The requirements imposed on the
SANSA enabling components as a consequence of the KPIs are out of the scope of this document and
will be presented in D2.4 “Requirements specification for the key enabling components”.
The end-to-end KPIs that will be assessed for SANSA are the following:
Aggregated throughput
Backhaul network resiliency
Delay
Spectrum efficiency
Energy efficiency
Geographical coverage
It should be pointed out that the project does not aim to improve delay, however it is included as a
KPI to ensure compliance with current expected performance. The assessment of delay and bit error
rate is based on the user perspective and the type of service that is delivered. SANSA shall employ all
the necessary mechanisms to adapt the backhaul network and fulfil these requirements. This topic
will be studied in depth in WP4 where we will be looking at the different traffic classification, load
balancing and routing algorithms.
In the sections that follow we present the end user Quality of Sevice (QoS) requirements for the
different types of service, define the system end-to-end KPIs, and set the corresponding target
values.
5.2 End user QoS requirements per service type ITU-T has introduced the G.1010 Recommendation which defines the user-driven performance
requirements for different types of service that are independent of the underlying technology and
network used [14].
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 85 of 131
SANSA-645047
D2.3
The main parameters that impact the user’s perceived quality of service are:
Delay: The delay perceived by the user is the time it takes between the request and delivery
of information. The achieved throughput is also taken into consideration here, as the user
perceives its impact on the experienced delay. Delay is a very important parameter that
influences the user satisfaction of the service and it is highly dependent on the service type.
Delay variation: The experienced delay is not always fixed, but it varies over time. Delay
variation can be mitigated in delay-sensitive applications with the use of buffering. However,
buffering may result to additional delay being added to the user experience, affecting the
perceived QoS.
Information loss: In all types of service information loss is directly linked to the quality of
information received by the user and is therefore a key requirement.
The following service types are highlighted in [14] with a detailed breakdown of their quality levels in
[15].
Conversational voice: This category includes all services that need to relay voice irrespective
of the technology used. All three parameters of delay, delay variation and information loss
present challenges for this service type. Delay can cause echo which is usually solved with
echo cancellers. High delay can also make the conversation difficult when it starts becoming
noticeable by the conversing parties. The human ear is sensitive to delay variance so it is
common practice to use buffers to remove the variation among the arrival of the different
packets. Lastly, it is important for the information loss to be kept at a level that doesn’t
affect the natural sound of the human voice so it can be easily recognised.
Voice messaging: Voice messaging is different to the conversational voice in the sense that it
is not real-time communication so the delay requirements can be relaxed. In this context,
delay is referring to the time between requesting to record or replay a voice message and
the starting to do so.
Streaming audio: The main requirement when streaming audio is to maintain high quality
while reducing the need for re-buffering at the receiver. Information loss is less critical than
for conversational voice (since retransmissions are possible) but more important than for
voice messaging in order to maintain uninterrupted audio flow. As with voice messaging,
there is no requirement for interactive communications so the user can tolerate higher
delays. However, unlike voice messaging audio streaming can be subject to re-buffering
which impedes the overall quality of experience.
Videophone: This service combines two-way video and audio communication such as Skype
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 86 of 131
SANSA-645047
D2.3
video call or other video conferencing applications. Due to the conversational nature of this
service, the user requirements are the same as in conversational voice applications in terms
of delay, however there is an additional requirement for lip-synch which refers to the
synching between the audio and video traffic. As far as quality is concerned, the information
loss rate achieved should be better than that for the conversational voice as both video and
audio is included in this service.
One-way video: This service is equivalent to streaming audio so the user requirements are
similar.
Web browsing: Web browsing refers to accessing the HTML content of webpages and the
main indicator is the time between requested and viewing the desired content. Therefore,
the delay is more tolerable compared to videophone or conversational voice.
Bulk data transfer/retrieval: As this type of service refers to file downloading, the user can
be more lenient with the delay requirements especially as the size of the downloaded file
increases.
High priority transaction services: High priority transactions are the ones required for e-
commerce such as online payments. The delay for these should be similar to that for web
browsing because the user needs to feel confident about the responsiveness of the system.
Low priority transaction services on the other hand are primarily one-way (such as SMS
service) and can function with much higher delays of up to 30s.
Command/control: Very low delay and no information loss are the key requirements for
command/control applications as the majority of these services are time critical. Interactive
gaming is a type of such applications with very strict delay requirements to ensure the
desired quality for the user.
Still image: There is a variety of encoding formats for still images with different
requirements for information loss. The delay expected from the user is comparable to the
one for file transfers.
E-mail access: The main requirement for email access is the responsiveness of the local
server to the user requests. In this context, the delay that can be tolerated by the user is a
few seconds whereas the server-to-server communication can take several minutes.
Figure 5-1 classifies the different service types presented based on their delay and information loss
requirements.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 87 of 131
SANSA-645047
D2.3
Figure 5-1: Model for user-centric QoS categories [14]
Based on the characteristics of the different types of service that were described, Table 5-1
summarises the performance targets for the various audio, video and data applications. It should be
pointed out that these targets are set based on the end user perception of the QoS, so they are not
linked with any specific underlying technology. For most services preferred and acceptable targets or
limits are available which helps to evaluate whether a specific technology can be employed for a
specific service. This model is a good fit for the SANSA backhaul network because it enables us to
evaluate the performance of every service based on the actual user needs, map it to the available
terrestrial and satellite links and select the best solution. Throughout the project though, it will not
be possible to evaluate the performance of all the services presented below, but only the
performance of a number of representative services.
Table 5-1: Performance targets for audio, video and data applications [14]
Backhaul network resiliency SANSA reconfigured network up
and ready in <10 seconds.
Delay
Per service type targets.
Expected improvements of up to
20-30% over SoTA routing
solutions especially under
backhaul environments.
Spectrum efficiency 10-fold improvement within the
considered Ka band segments.
Energy efficiency Up to 30% improvement
compared to benchmark.
Coverage 95-99% EU coverage
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 95 of 131
SANSA-645047
D2.3
6. Identification of Research Challenges
6.1 Specific challenges associated with SANSA integrated terrestrial and satellite backhaul node
SANSA integrated and satellite backhaul node creates a number of research and technical challenges
as presented below.
6.1.1 Efficient allocation and management of resources in terms of cost and energy The existence of two links on a number of nodes is posing challenges to the efficient allocation and
management of resources. The way resources are going to be allocated, taking into consideration
the additional cost and capacity that the satellite link introduces into the system and the way this
can be balanced in order to have the most efficient use in a way that additional cost is surpassed by
the performance enhancement.
Apart from cost efficiency of the integrated backhaul link, energy efficiency is also a question within
the SANSA frame. Management of links and related equipment should make sure that those links are
used in a power efficient way as one of the major project objectives is to reduce power consumption
up to 30% compared to existing backhaul networks.
6.1.2 Investigation of handover capabilities between terrestrial and satellite backhaul nodes
As it has already been mentioned, one of the main characteristics of SANSA is the coexistence of
satellite and terrestrial links at the same node. One of the challenges created by this coexistence is
the handover capabilities of the node. Hence, the node should be able to provide this handover in a
manner transparent to the user. This means that overload and failure monitoring mechanisms
should be employed by the integrated backhaul node in order to be able to constantly identify
overloaded links and redirect traffic to the less congested ones.
6.1.3 Interoperability of the iBN One of the main challenges that SANSA project will have to deal with is the interoperability of the
terrestrial link and the satellite link. This node will have to seamlessly provide connectivity between
two totally different links, a terrestrial microwave link and a satellite link, both operating at the Ka
band. The different nature of the two links requires that the design is taking into consideration
several aspects of the heterogeneity between the satellite link and the terrestrial link. The
components, the interfaces, the routing capabilities as well as traffic classification techniques are
designed in a manner that there is no biased routing decision and the best path is selected under
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 96 of 131
SANSA-645047
D2.3
any circumstances. Moreover, careful network planning and interference management is required at
a node level in order to avoid interference.
6.1.4 Components design SANSA network promises to deliver a self-reconfigurable, smart network that will enhance backhaul
capabilities. In order to achieve the objectives set, new components will have to be developed.
These components as they were already described are the HNM, which will centrally manage the set
and the smart antenna.
6.1.5 Traffic classification One of the main SANSA features is its hybrid nature. SANSA will take advantage of both satellite and
terrestrial connections; hence a mechanism to classify traffic and send data through the optimal
connection will be needed. There are several techniques that can be applied in order to classify
traffic (port based, DSCP, policy based etc.). The decisions will be based on each service’s unique
characteristics such as bandwidth requirements, latency and jitter sensitivity. What can be told a
priori is that the satellite connections can enhance video services that are bandwidth hungry but not
latency sensitive; the type of traffic expected to be dominant in 2020 [17]. Terrestrial connections
can support services that require low latency and low jitter, like voice services. The service
requirements as expressed in the previous chapter will play a key role in the design of a traffic
classification mechanism for the hybrid network.
6.1.6 iBN interfaces with satellite and terrestrial links SANSA will use already existing equipment for the satellite and terrestrial connections. As a result
SoTA satellite modems will be used to support satellite connections. The same applies for the
equipment used to support terrestrial connections. These devices will need to communicate with
the new components, which will support the existence of the integrated backhaul node at central
and local level. Thus, the design of the interfaces to support this communication is another challenge
for the proper functionality of the integrated backhaul node.
6.1.7 iBN monitoring performance The iBN will be responsible for decisions at a local level as far as the SANSA network is concerned. As
a result, constant accurate monitoring of the network resources and link status for satellite and
terrestrial link as well as proper connection with the HNM should be designed. Real time
undisrupted monitoring of the node status (signal levels for terrestrial and satellite connections,
available bandwidth etc...) is of paramount importance.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 97 of 131
SANSA-645047
D2.3
6.2 Specific challenges associated with SANSA smart antennas
6.2.1 Transceiver architectures In contrast to communication systems below 6GHz, SANSA terminals operating at Ka band are
expected to have a large number of antennas in order to cope with the path loss and interference
mitigation requirements. This large increase in the number of antennas leads to additional research
challenges.
Considering an operational bandwidth of 1 GHz available in the SANSA spectrum sharing scheme, an
antenna array with 50 elements would require an underlying hardware able to work at 50 GHz of
bandwidth. This bandwidth is unaffordable with the current off the shelf FPGAs and ASICs. As a
result, digital precoding and filtering becomes unfeasible in SANSA terminals even if low complexity
precoding and receiving schemes are deployed.
As a result, the system designer shall target transceiver architectures whose number of digital
entries is reduced severely. Where the beamforming is done in the analog domain, the baseband
processor would only require one digital entry, thus minimizing the hardware bandwidth processing
requirements. It is important to note that even though the beamforming operation is done in the
analog domain, its control requires certain digital inputs. For that case, and depending on the
operational mode, the transmitter shall be able to compute the analog beamforming matrix and
send it to the beamforming network. This can be the case of an analog phase only architecture
where the transmitter shall compute the different phases considering the different number of
control bits.
Unfortunately, certain operation modes required in SANSA terminals require very complex analog
units. This is the case of multibeam transmission (i.e. two different symbols simultaneously sent with
different beamformers) that require 2N phase shifters, power amplifiers and N combiners. This can
exponentially increase the analog subsystem cost and complexity.
Consequently, an intermediate solution where certain processing is done at the digital domain and
the rest in the analog one is more convenient. This way, the analog processing unit will transform N
analog signals to M digital entries, with M<N, leading to a substantial reduction of the digital
processing complexity.
The research challenges of the overall transceiver design are the following:
i) Which is the ‚optimal‘ M and N? In other words, how much processing shall be done in
the analog and digital domain.
ii) What is the most efficient analog processing unit yet preserving a low cost? i.e. ‘Phase
only array‘, ‘switching matrix’, ‘spatial feeding scheme’, etc.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 98 of 131
SANSA-645047
D2.3
iii) Given a hybrid architecture, what are the optimal precoding/filtering digital-analog
matrices?
6.2.2 Angle Coverage Considering the large path loss inherent from the Ka-band transmission, it is essential to take into
account the height of the SANSA terminals. In contrast to sub-GHz transmissions, microwave
communication links propagation is no longer located in the same plane with antennas and is
instead distributed randomly in three-dimensional space.
Under this context, SANSA terrestrial transceivers shall be able to point in a three dimension basis.
Covering 360º degrees in azimuth and -90º to 90º in elevation can be a very challenging design. In
addition, depending on the use cases the angle coverage requirement can be reduced. For instance,
in rural scenarios where large levels of re-configurability are not expected, the antenna would need
shorter angle coverages. On the contrary, urban scenarios would require large angle coverages as
the node density will increase exponentially.
6.3 Specific challenges associated with Hybrid Network Management
6.3.1 Configuration management The HNM must encompass resource control and traffic planning functions for the hybrid network, by
interacting with iBNs. This management will happen at different time-scales. There are
reconfiguration processes taking place at small time-scale (seconds or fractions of seconds), at node
level, while other changes will be made at medium time-scale (in the order of minutes) or even at
larger scales (manual or planned network reconfigurations).
The network nodes in SANSA network can be hybrid, in the sense of having both terrestrial and
satellite terminals, and the HNM must be aware of the beams being deployed throughout the
network at any certain moment.
The number of reconfigurable antenna/modem/sat-modem/sat-hub parameters needed in the
frame of SANSA needs to be confirmed, but will probably be related to:
6.3.1.1 Terrestrial system configuration This function consists in the configuration of the resources of the backhauling network, including not
only physical but also upper layer parameters. Modifications could be global or affecting only parts
of the topology. The difficulty lies on keeping the network stability when performing such
modifications. The amount of available network capacity is closely dependent on spectrum
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 99 of 131
SANSA-645047
D2.3
resources. If these are considered as fixed, only ‘small’ modifications can be performed in the
network to optimize efficiency, according to network conditions.
6.3.1.2 Satellite system configuration In addition to terrestrial resources, SANSA network can make use of the satellite system for network
resiliency or optimization. A GEO satellite, using Ka band, with DVB-RCS2/S2x air interface, will be
considered in SANSA. Satellite resources should be used only when the efficiency gains justify it. This
requires that the HNM periodically calculates alternative topologies with or without using the
satellite links, and their corresponding cost. Also, it could be practical for certain services, such as
CDNs, the use of the satellite component when the terrestrial resources are scarce or when better
synchronization is required (regardless of the physical terrestrial topology).
6.3.2 Frequency plan management The scope of this function will strongly depend on the frequency reuse strategy, between the
satellite and terrestrial segment. There could be reconfiguration plans including restrictions to avoid
spectrum reuse or, on the contrary, to maximize the bandwidth. Complexity for the HNM lies on
applying these policies locally, depending on cell or coverage areas location.
6.3.2.1 Interference management Several interference scenarios can be considered for SANSA, normally originated by the terrestrial
signal. Interferences can be originated from terrestrial or satellite transmitters, and may affect both
UL and DL beams. To mitigate this problem, the function of the HNM could detect interferences and
‘repair’ the network as soon as possible, by switching down the source or changing to a more secure
frequency plan.
For the satellite system, location of the interferer position could be challenging when appropriate
tools are not available at the Satellite Control Centre.
Interference from satellite DL signal to the terrestrial transmission seems not to be an issue, because
antennas are very directive.
6.3.3 Fault management The monitoring function will enable the HNM to become aware of the events in the network
topology, which may trigger automatic reconfiguration actions to continue meeting the established
KPIs:
A terrestrial or satellite beam can be switched off.
A certain node can report a congestion state, for instance when the HNM receives traffic
performance measurements.
A traffic classification strategy can be modified as a link status changes. Although this
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 100 of 131
SANSA-645047
D2.3
modification could be performed locally by the iBN, further modifications in other nodes
could be required, if not done automatically.
Static configuration actions are also foreseen in the HNM, when a new network topology is to be
defined. This operation, manually performed by SANSA operator, consists on establishing new links
or modifying existing ones. It implies nodes reconfiguration, since a new set of beams (or beams
characteristics) needs to be propagated throughout the whole network.
It is envisaged that smart antennas can reconfigure themselves to improve link conditions. In
principle, this can be done independently of the HNM stored configuration, which aims mainly to
store antenna pointing and power (or more specifically, destination nodes and bandwidth).
6.3.4 Performance management The HNM should monitor the network performance by obtaining certain subsystem indicators from
the node elements and offer global performance indicators to the operator. According to the
network size, the HNM could receive a big volume of indicators and require high processing power
to extract the system performance.
6.3.5 Network management (topologies) The HNM must implement topology reconfiguration. In principle, the implementation should be
independent of the scenario (rural, urban, etc.). This functionality will be able to define a combined
ring/tree topology, starting with a node matrix having certain constraints (e.g. each iBN can only
connect to other neighbours if they are reachable). The objective is that no nodes in the mesh
network are left un-connected, while covering all the needed area. Moreover, for topology central
nodes, the bandwidth requirements will be higher. At macro-cell level, it can be considered that links
among eNBs are fixed, and will not be configured by the HNM. The HNM will take into account
micro-cells associated to smart antennas topologies.
Other function related to topologies is the ability to respond to link failures, by using alternate
connections among iBNs. This will be possible only when there are reachable neighbour nodes that
allow for an alternate topology communication enabling to bypass the broken link. As the number of
beams from each terrestrial antenna is fixed, this operation will possibly imply to remove one
‘healthy’ link to be able to construct the alternate path(s).
One important HNM feature will be topology modification in response to traffic congestion events,
at macro/micro cell level. To this aim, it must be configured the combination of aggregated capacity
available at each iBN, considering both terrestrial and satellite links for the hybrid nodes. Satellite
links, when established, can relieve terrestrial links when congestion events happen.
The HNM must support a multicast architecture. That is, it should be able to establish both PtP and
PtMP links/beams whenever needed.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 101 of 131
SANSA-645047
D2.3
6.3.6 Interface with the core network (EPC) Implementing user traffic interfaces between the EPC and the HNM is not foreseen in the frame of
SANSA. It can be assumed that edge nodes are directly connected via fibre links to the EPC, and
therefore to the S-GW. These nodes will be hybrid, and therefore the packets received or directed
from/to another satellite terminal will be conveniently sequenced before they are de-capsulated at
the S-GW or the eNB.
6.3.7 SANSA CDN use cases To be able to support CDNs scenario, the HNM should prepare the network to receive the multicast
contents, that is, assure that all the nodes can receive the multicast signal. This involves:
Configuring PtMP links at the regulated frequency (e.g. @18 GHz), selecting the access
scheme (normally will use TDMA/OFDMA), and using shorter link distances than those
employed for PtP links.
Decide the better multicast topology (terrestrial-only, satellite-only, hybrid). This involves
the selection of the multicast “master” node for the transmission.
Check for the best energy-saving scheme.
Configure the nodes for conveniently routing the multicast traffic, allocating resources and
routing policies where needed.
6.3.8 Mobility management This point concerns the handling of mobile terrestrial cell nodes, varying the physical topology
without HNM control. The HNM should keep track of all the cells and adapt the network topology to
maintain connectivity with the mobile or nomadic nodes. Depending on coverage, some nodes will
only be reachable through a satellite link, but other possible scenario could on a change of the
existing terrestrial links.
6.3.9 Security management The HNM must support security options for multi-profile operation (according to different roles, e.g.
a satellite operator and a terrestrial operator).
6.4 Specific challenges associated with spectrum sharing The purpose of the present report is to identify and discuss the research challenges associated with
the terrestrial-satellite spectrum sharing for purposes of backhauling. We summarize the trends in
the research community and identify open research problems.
Satellite operators are facing the spectrum scarcity of conventional C/Ku bands and they are already
demanding more spectrum in higher frequency bands. In SANSA, we focus the research effort on Ka
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 102 of 131
SANSA-645047
D2.3
band where terrestrial backhauling bands of 18 GHz and 28 GHz are shared with the satellite-to-
Earth and Earth-to-satellite, satellite bands. In doing so, terrestrial operators will be able to cope
with the increase of traffic demands and will share part of the band licensing costs with satellite
operators, whereas the latter will take part in the mobile backhaul market.
More precisely, as shown in D2.1 [1], current CEPT recommendations assign the frequency bands of
19.7-20.2 GHz and 29.5-30 GHz for exclusive Space-to–Earth and Earth-to-Space satellite
communications, respectively. The objective of SANSA is to develop novel technology that will allow
deploying satellite terminals in the 17.7-19.7 GHz band without the need of occupying part of the
satellite exclusive bands. In a similar way, they will be also deployed in the exclusive terrestrial sub-
bands within the 27.5-29.5 GHz.
Here, we focus on the research challenges associated with the spectrum sharing in hybrid satellite-
terrestrial backhauling networks. These research challenges have been classified in two general
categories: 1) Spectrum awareness techniques which focus on acquiring knowledge about the
spectrum utilization in the adjacency of a terminal, 2) Spectrum exploitation techniques which focus
on exploiting this knowledge for interference mitigation and/or resource allocation.
These two groups of techniques and the related existing literature are detailed in the following
sections.
6.4.1 Spectrum awareness techniques Spectrum awareness solutions are used to obtain relevant information and knowledge of the
surrounding radio environment.
In a first step, SANSA project will address database-assisted shared spectrum techniques to
efficiently manage the shared access to the spectrum. In addition, spectrum sensing will be explored
to improve the quality of the dynamic information stored in the databases and combined with
network-level radio environment mapping for an improved hybrid access performance.
In this section we focus our attention on research challenges associated to spectrum sensing,
spectrum cartography and radio environment mapping.
6.4.1.1 Spectrum Sensing Spectrum sensing involves making observations of the radio frequency spectrum and reporting on
the availability of spectrum. This can be done in a decentralized mode in which each intelligent
backhaul node makes a decision based on its own measurements or in a cooperative way (either
with a centralized fusion node or just sharing the information with few neighbours), in which
multiple sensing nodes cooperate [17]. The latter has been shown to be effective in relaxing the
sensitivity requirements on individual secondary users and improving the overall sensing
performance [17].
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 103 of 131
SANSA-645047
D2.3
In this section, we focus our attention in spectrum sensing challenges related to single-device
spectrum sensing methods. In SANSA context, the single-devices refer to the intelligent backhaul
nodes. Among the most important research challenges that spectrum sensing techniques must face
are [19]:
Restricted sensing ability: Most devices do not have specific RF transceiver to sense the environment and therefore they typically access the spectrum following a two-stage “listen-before-talk” protocol in which the device performs sensing and transmission independently in two different time intervals. There is an evident trade-off between the sensing capabilities and the throughput that can be achieved by these devices [20], [21].
Wideband sensing: The traditional way for sensing a wideband spectrum is channel-by-channel sequential scanning [22], which introduces large latency. One way to solve the latency problem, at the cost of implementation complexity, is to use an RF front-end with a bank of narrow band-pass filters. One of the emerging trends in the research community is to directly sense a wide frequency range at the same time [23]. Obviously, the scanning of a wide band of frequencies implies high sampling rates which are not affordable by the standard Analog-to-Digital Converter (ADC). Exploiting the fact that the spectrum of interest contains only a small number of active frequencies relative to the band-limit [24], a promising alternative to alleviate the sampling bottleneck is the use of sub-Nyquist sampling techniques also known as Compressive Sensing (CS) [25]. Wideband sensing together with CS research is known in the literature as Wideband Spectrum Sensing (WSS) and has attracted extensive attention due to its enormous potential of sharing spectral resources [26], [27].
Practical imperfections: In practice, there may occur several imperfections such as noise uncertainty, channel/interference uncertainty, and transceiver hardware imperfections like amplifier nonlinearity, quantization errors, and calibration issues [28]. These imperfections may severely deteriorate the performance of the employed spectrum awareness mechanism. In this case, investigation of the methods to counteract the effect of these imperfections is an important research challenge to be addressed from the practical perspective.
Interference measurement: This refers to the fact that one network node transmitting in a shared spectrum cannot be aware of the effect of its transmission on other active receivers operating in the same band. One possible solution is to use a separate sensor network distributed over the area of interest which captures measurements from many fixed sites to create an interference map [29]. The cooperative approach is discussed later on in Section 6.4.1.2.
6.4.1.2 Spectrum Cartography Spectrum cartography is the process of constructing a map showing radio frequency signal strength
over a geographic area (Figure 6-1). This map is commonly known as Radio Environment Map (REM)
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 104 of 131
SANSA-645047
D2.3
[30], [31]: a database containing information on the radio environment. The research challenges
related to this case are listed below:
Sharing information: REM concept is based on geo-localized measurements captured by many devices sharing the same spectrum. In such a multi-user environment, cooperation is essentialto exploit the spatial diversity. However, cooperation in the spectrum sensing phase is not straightforward since the spectrum sensing information needs to be distributed over a subset or all active devices. This might introduce delays and signalling overhead.
Decision fusion: Assuming that the sensing information is well received by all devices, another interesting research path is how to combine all the information (decision fusion algorithms) to produce an accurate map.
Learning: The REM information can be updated with observations from the cognitive nodes. Therefore, learning mechanisms and knowledge management techniques [32] which have been the objective of recent research can be applied as well in this context.
Localization of emitters: The knowledge of the emitters’ location can improve the accuracy of the final estimate of the REM. The process of REM construction becomes easier if information about the incumbent users’ locations is known beforehand by using some estimation methods [33].
Sparsity-based cartography: In general spectrum cartography literature, the spatial sparsity of the multiple sensing devices over a geographical area is not taken into account, neither the sparsity on the frequency domain of the spectrum measurements. Recent works proposed to combine CS technology with spectrum cartography to improve the Radio Environment Map (REM) construction process [34].
6.4.1.3 Spectrum awareness through regulatory databases The process of REM construction can be improved with the help of national and international
spectrum regulators, since they usually register the licensed systems into databases. The main
advantages of the availability of regulatory databases are that it can provide information about the
radio spectrum environment and hence can reduce the required sensing burden. Some challenges
related to use of regulatory databases are listed below:
Accuracy of the listed parameters: In practice, the actual values such as power levels, antenna patterns, etc., may vary from that of the specified values in the database. In such cases, databases may be considered as an initial step and it can be updated based on the outcomes of spectrum awareness mechanisms.
Temporal-variations on the spectrum use: Databases cannot capture the temporal spectrum opportunity. Again, the information provided by the database should be considered as side-information to other sensing mechanisms.
Out-of-date: Maintaining and updating of databases over time is significant for the operator. However, in SANSA, the involved terrestrial FS links and the satellite network are not
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 105 of 131
SANSA-645047
D2.3
dynamic as in a mobile system and, thus, it is not necessary to update the database so frequently. In any case, verification of available database via measurements can be considered.
Figure 6-1: Example of spectrum cartography3
3 Extracted from CROWNCOM 2010 keynote speech by Berna Sayrac (Orange Labs).
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 106 of 131
SANSA-645047
D2.3
Figure 6-2: Classification of Interference mitigation approaches
6.4.2 Spectrum exploitation techniques
6.4.2.1 Interference mitigation techniques Figure 6-2 presents the classification of several interference mitigation techniques in the context of
spectrum sharing scenarios [35]. The main techniques are: cognitive beamforming, cognitive
interference alignment, and cognitive zone. A short description about these techniques and the main
research challenges involved with these techniques are provided in the following subsections.
6.4.2.2 Cognitive beamforming The main difference between cognitive beamforming and the conventional beamforming problem
arises due to the introduction of interference constraints in order to restrict the interference
towards/from the victim/interfering stations. In this context, cognitive beamforming approaches
have been widely studied with different secondary network optimization objectives such as
SINR/rate balancing [36], [37], sum rate maximization [38], and power minimization with QoS
constraints [39]-[41].
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 107 of 131
SANSA-645047
D2.3
Cognitive beamforming can be applied in the considered SANSA scenarios with the objectives of
either controlling the interference or maximizing the SINR of the desired link. For this purpose,
additional constraints on the aggregated interference at the victim/interfering FSS stations/terminals
need to be taken into account while designing beamforming at the SANSA FS stations. Further,
depending on the spectrum sharing scenarios, either transmit beamforming or receive beamforming
or the combination of both can be investigated.
Cognitive beamforming techniques can be further applied in combination with power control [42]-
[44] and user scheduling [45]-[46][47] in order to enhance the exploitation of the unused spectral
resources. Similarly to the conventional beamforming scenarios, cognitive beamforming techniques
should be robust to the array response vector mismatch, calibration errors and channel
uncertainties. This robustness can be incorporated by employing either a stochastic or a worst case
approach [48]. Further, it is usually difficult to acquire the perfect Channel State Information (CSI)
knowledge at the secondary transmitter in practice due to limited training, less cooperation
between primary and secondary systems, and quantization issues. In this context, the aspects of
robust cognitive beamforming considering the imperfect channel state information have received
significant attention in the spectrum sharing literature [49], [50].
The main research challenges for the application of cognitive beamforming techniques in SANSA
spectrum sharing scenarios are summarized below [51].
Channel State Information: For the Channel State Information (CSI) acquisition process, investigating effective feedback techniques in order to reduce the feedback burden from the secondary/primary receivers to the secondary transmitter is an emerging research topic.
Implementation complexity: Many existing non-robust techniques may fail in creating a desired beam pattern in case of the imperfect CSI, hence causing harmful interference to the primary users. Further, there arises the issue of additional complexity while employing robustness in the beamforming design problem. To this end, the investigation of robust and practically realizable techniques is another important research challenge.
Uncertainties in the array response vector: Besides CSI robustness, beamforming solutions should also be robust to the uncertainties in the array response vector, inaccurate Direction of Arrival (DoA) information, and transceiver hardware imperfections such as phase noise, quantization errors etc.
Interference threshold: Computational efficient solutions need to be investigated in order to solve the cognitive beamforming problems in the considered SANSA scenarios.
Direction of Arrival (DoA) of devices: The acquisition of accurate DoA information of the desired and interfering sources/receivers is crucial for implementing cognitive beamforming in the considered scenario. In practice, this information can be obtained either from the databases or by employing a suitable DoA estimation algorithm.
Calibration: Accurate calibration of the antenna array equipped at the SANSA FS station is
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 108 of 131
SANSA-645047
D2.3
crucial, especially while creating nulls to the directions of the interfering/victim sources. In case of calibration errors, suitable compensation mechanisms need to be investigated.
6.4.2.2.1 Cognitive Interference Alignment In spectrum sharing context, Interference Alignment (IA) can be used as an interference mitigation
tool which aligns interference in space, time or frequency domain using suitable precoding
techniques. In this approach, signals transmitted by all users can be designed in such a way that the
interfering signals fall into a reduced dimensional subspace at each receiver. Each receiver can then
apply an interference removal filter in order to project the desired signal onto the interference free
subspace. With this phenomenon, the number of interference-free signalling dimensions of the
network can be substantially increased [52].
In the context of spectrum sharing networks, IA techniques can be broadly classified into non-
cooperative [53], [54] and cooperative [55], [56]. Further, IA techniques can be CSI-aware and blind
as depicted in Fig. 1. The main research challenges involved with cognitive interference alignment
technique are specified below [57], [58].
Channel State Information: In many cases except in the distributed IA, the global channel knowledge is required to carry out the IA operation. In this context, it’s a crucial aspect to investigate suitable blind and semi-blind IA techniques in order to reduce the overhead for acquiring the sufficient channel knowledge.
Channel uncertainty: The penalty of residual channel uncertainty at the transmitters and the impact of channel correlations are other aspects to be explored for IA implementation in practice.
Synchronization: IA techniques require strict synchronization in order to avoid any timing and carrier frequency offsets between cooperating nodes. In this context, suitable synchronization or compensation approaches need to be investigated for the realization of distributed IA in practice.
Low SNR scenarios: Investigation of new IA algorithms which can provide better sum capacity in moderate or low SNR region is another interesting research challenge [58].
Dimensionality of interference networks: Another main limitation for the IA technique is the requirement of the large dimensionality of interference networks. The practical achievable scheme which requires finite dimensions for the case of multiple non-intended receivers is still an open research problem.
6.4.2.2.2 Cognitive Zone
In the spectrum sharing scenario, a Cognitive Zone (CZ) is usually designed around the primary receiver based on its interference threshold, within which secondary users are not allowed to reuse the frequencies used by the primary user. However, in practice, the interference may occur in both directions, i.e., from the primary transmitters to the secondary receivers and from the secondary
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 109 of 131
SANSA-645047
D2.3
transmitters to the primary receiver. Depending on the considered scenarios and the level of interference between these two systems, CZs can be created around one of these receivers or both.
The CZ method has been investigated in the literature in various settings [59]-[62]. The size of the CZ has a great impact on the Quality of Service (QoS) of the primary system since it affects the level of secondary interference that needs to be tolerated and on the secondary system’s capacity since it affects the available amount of primary spectrum at a given location [62]. As depicted in Fig. 1, the CZ approach can be divided into: (i) fixed CZ, and (ii) dynamic CZ. In the first approach, the size of the CZ is fixed based on the worst case scenario, whereas in the second approach, the size of the CZ can be varied based on the time-varying information of the surrounding radio environment such as instantaneous PU behaviour. The complexity of the database for a dynamic CZ becomes higher and the adaption of the CZ is relatively a slow process depending on the level of cooperation between primary and secondary systems [63]. The main research challenges involved while applying the CZ method are mentioned below.
Unpredictable propagation conditions and inaccuracy of the database: Static CZ method may not always guarantee the perfect avoidance of the co-channel interference. The received interference level may vary based on different factors such as terrain variations, environmental conditions, achieved antenna gain patterns, etc. Since the CZ method is mostly based on database created using some propagation models, the accuracy of the database is crucial for the interference protection guarantee. Furthermore, signals could occasionally travel further than expected if (unpredictable) propagation conditions are not properly accounted for.
Terrain enabled propagation models: The consideration of the terrain-based propagation models while investigating Exclusion/protection contours in SANSA spectrum sharing scenarios is one interesting aspect to be explored.
Multi-tiered CZs: Defining the proper boundaries considering realistic propagation environment for the implementation of multi-tiered CZs in SANSA spectrum sharing scenarios is another important research challenge.
Combination of exclusion/protection zones and spectrum awareness mechanism: The sensing measurements can be considered to progressively adjust the exclusion/protection zones. How to combine these two methods effectively in the considered SANSA spectrum sharing scenarios is another interesting aspect.
Combination of exclusion/protection zones and resource allocation strategies: CZ method can be further combined with the power control approach in order to exploit the unused spectral resources in an efficient manner. Another promising approach is to combine this method with the carrier allocation method in which the secondary users within the CZ (designed for a particular frequency) can switch their operating frequency to some exclusive frequency. In this context, investigation of suitable joint approaches is another aspect which
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 110 of 131
SANSA-645047
D2.3
can be considered in SANSA spectrum sharing scenarios.
6.4.2.3 Optimization of spectrum and radio resources The resource allocation in both the communication networks and in the wireless backhaul plays an
important role in enhancing the overall system performance. On top of the interference mitigation
schemes, the resource allocation in spectrum sharing based networks is responsible for
distributing/allocating the available spectrum in combination with applying the adequate power
control strategies. In that sense, the scarce wireless backhaul frequency spectrum should be
allocated wisely so that no spectrum waste is caused due to more spectrum being allocated than
what is required. At the same time, the allocated spectrum for backhauling should not be
inadequate which would degrade the system performance significantly. The research challenges
related to the resource allocation in spectrum sharing context can include the limitation generated
from:
Imperfect spectrum sensing information: Designing the resource allocation algorithm while considering that the sensing information is perfectly known may lead to inefficient distribution of the resources and cause harmful interference to the parties that share the spectrum. Any allocation process should consider the sensing false alarm probability where the amount of the resulted interference should be adapted accordingly.
Imperfect channel state information: There is always some uncertainty in the channel state information due to the feedback channel errors, limitation, or unreliability. Accordingly, the resource allocation in the SANSA transmitters should be developed in order to reduce the negative impact of the lack of this information.
Cross-Layer resource allocation: While independent optimization per individual layer can lead to system performance improvement, exploiting the synergies between several layers may result to even better enhancement and adaptability to the system variations. Accordingly, SANSA backhauling nodes can take into account several sets of information from the MAC and network layers, in addition to the possible environmental and traffic changes.
Priority consideration: Design of well-defined utility functions to balance the different constraint and enable the usage of machine learning schemes in addition to game theoretical model can be a possible approach depending on the nature of the optimization requirements.
Computational complexity and green communications: In addition to the minimization of the transmit power, the design of an efficient resource allocation algorithm with low computational complexity helps in reducing the processing time for the backhauling nodes and accordingly reducing the latency and the consumed energy.
Inter-network interference consideration: Depending on the sharing schemes and on the way of distributing the spectrum, i.e. centralized, distributed or cluster-wise, the
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 111 of 131
SANSA-645047
D2.3
interference between the different SANSA nodes should be included considering the different possible interference caused by the communication between the terrestrial nodes and/or the communication between the satellite and terrestrial nodes.
6.5 Specific challenges associated with integrated backhaul service delivery
The service delivery model is a challenging area for the operators. The ownership of the components
and the business model needed to decide who is running the service, who is the owner of the Hybrid
Network Manager, and how data from backhauling network can access the core network. The
different approaches are a matter of the network architecture design, where the various
components will be placed and where the connection to the core network will be implemented.
Various fields of research depend on the way the service is going to be delivered and by the route
data is going to follow to access the internet.
6.5.1 Mobile terrestrial operator lead In this model, the responsible for the connectivity between the backhaul network and the core is the
terrestrial operator. The PGW, the SGW, and the HNM are entities that are present in the MNO
network. Satellite operator receives data coming from terrestrial nodes and routes them through its
network to the MNOs network. All the data handling in terms of PGW, SGW and HNM is performed
at MNO’s premises.
In order to briefly describe this scenario, data packets from the HeNB, are directed through either
the terrestrial or the satellite link following the path that was decided at the node. This decision is
the result of the interoperability function. If data flow is routed terrestrially, then all packets are sent
directly to the MNO network following the path designed by the routing algorithm. This decision will
be based on the current topology status and QoS requirements as they will be provided from the
HNM to each smart node.
If the data is sent via satellite, it is received by the satellite hub and sent to the IP gateway. From
there it is routed to the MNO’s network where the SGW and PGW are present to onwards route it to
the internet.
In this scenario all important components belong to the mobile network operator as well as the
HNM, which is responsible for the routing and topology rearrangements. The satellite operator is
transparent or almost transparent to the packets routed through its network (Figure 6-3).
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 112 of 131
SANSA-645047
D2.3
Figure 6-3: Terrestrial Operator Lead
6.5.2 Hybrid Terrestrial Satellite Operator Service The second service model is the model where satellite operator and MNO are accessing the internet
separately. This is a hybrid model since components like the PGW and the SGW are present in both
satellite and mobile network operator premises. This allows data to access the network directly
either through the satellite operator or the mobile network operator. The HNM is placed again at
the MNO’s premises and all the SANSA network management and configuration is performed by the
terrestrial operator. This model offers a wide field for virtualization research for entities like the
PGW and the SGW. The level at which PGW and SGW are going to be migrated to the satellite
operator’s network is a challenge. Big challenge is if images of these entities going to be identical to
the original ones and if users are going to have the same rights and same type of access to the
functionalities that PGWs and SGWs provide (Figure 6-4).
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 119 of 131
SANSA-645047
D2.3
Survey," Communications Surveys & Tutorials, IEEE , July 2015.
[29] P.J. Kolodzy. “Interference Temperature: A Metric for Dynamic Spectrum Utilization”,
International Journal of Network Management, 16:103–113, Apr 2006.
[30] Yilmaz, H.B.; Tugcu, T.; Alagoz, F.; Bayhan, S., "Radio environment map as enabler for
practical cognitive radio networks," in Communications Magazine, IEEE , vol.51, no.12,
pp.162-169, Dec, 2013.
[31] Zhiqing Wei; Qixun Zhang; Zhiyong Feng; Wei Li; Gulliver, T.A., "On the construction of Radio
Environment Maps for Cognitive Radio Networks," IEEE Wireless Communications and
Networking Conference (WCNC), pp. 4504-4509, Apr, 2013.
[32] V. Atanasovski, J. van de Beek, A. Dejonghe, D. Denkovski, L. Gavrilovska, S. Grimoud, P. Mahonen, M. Pavloski, V. Rakovic, J. Riihijarvi, B. Sayrac, "Constructing radio environment maps with heterogeneous spectrum sensors," in IEEE Symp. in Dynamic Spectrum Access Networks (DySPAN), pp. 660-661, May 2011.
[33] L. Bolea, and et al, "Context Discovery Mechanisms for Cognitive Radio," IEEE Vehicular Technology Conference (VTC Spring), May 2011.
[34] Jayawickrama, B.A.; Dutkiewicz, E.; Oppermann, I.; Fang, G.; Ding, J., "Improved performance of spectrum cartography based on compressive sensing in cognitive radio networks," in IEEE Int. Conf. Communications (ICC), pp. 5657-5661, Jun, 2013.
[35] S. K. Sharma, “Interweave/Underlay Cognitive Radio Techniques and Applications in Satellite Communication Systems”, Ph.D. dissertation, University of Luxembourg, Luxembourg, 2014, http://orbilu.uni.lu/handle/10993/18973.
[36] A. Tajer, N. Prasad, and X. Wang, “Beamforming and rate allocation in MISO cognitive radio networks,’’ IEEE Trans. Signal Process., vol. 58, no. 1, pp. 362{377, 2010.
[37] K. Cumanan, L. Musavian, S. Lambotharan, and A. Gershman, “SINR balancing technique for downlink beamforming in cognitive radio networks,’’ IEEE Signal Processing Letters, vol. 17, no. 2, pp. 133-136, 2010.
[38] L. Zhang, Y.-C. Liang, Y. Xin, and H. Poor, “Robust cognitive beamforming with partial channel state information," IEEE Trans. Wireless Commun., vol. 8, no. 8, pp. 4143-4153, August 2009.
[39] K. Phan, S. Vorobyov, N. Sidiropoulos, and C. Tellambura, “Spectrum sharing in wireless networks via QoS-aware secondary multicast beamforming,’’ IEEE Trans. Signal Process., vol. 57, no. 6, pp. 2323-2335, 2009.
[40] E. Gharavol, Y.-C. Liang, and K. Mouthaan, “Robust downlink beamforming in multiuser MISO cognitive radio networks with imperfect channel-state information,’’ IEEE Trans. Veh. Technol., vol. 59, no. 6, pp. 2852-2860, July 2010.
[41] H. Du, T. Ratnarajah, M. Pesavento, and C. Papadias, “Joint transceiver beamforming in MIMO cognitive radio network via second-order cone programming,’’ IEEE Trans. Signal Process., vol. 60, no. 2, pp. 781-792, 2012.
[42] L. Zhang, Y.-C. Liang, and Y. Xin, “Joint beamforming and power allocation for multiple access
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 120 of 131
SANSA-645047
D2.3
channels in cognitive radio networks," IEEE J. Sel. Areas in Commun., vol. 26, no. 1, pp. 38-51, 2008.
[43] H. Du, T. Ratnarajah, M. Pesavento, and C. Papadias, “Joint transceiver beamforming in MIMO cognitive radio network via second-order cone programming," IEEE Trans. Signal Process., vol. 60, no. 2, pp. 781-792, 2012.
[44] R. Zhang and Y.-C. Liang, “Exploiting multi-antennas for opportunistic spectrum sharing in cognitive radio networks," IEEE J. Sel. Topics Signal Process., vol. 2, no. 1, pp. 88-102, 2008.
[45] K. Hamdi, W. Zhang, and K. Letaief, “Joint beamforming and scheduling in cognitive radio networks," in IEEE GLOBECOM, 2007, pp. 2977-2981.
[46] W. Zong, S. Shao, Q. Meng, and W. Zhu, “Joint user scheduling and beamforming for underlay cognitive radio systems," in 15th Asia-Pacific Conf. on Commun., 2009, pp. 99-103.
[47] A. Massaoudi, N. Sellami, and M. Siala, “A two-phase scheduling scheme for cognitive radio networks based on opportunistic beamforming," in IEEE VTC Spring, 2013, pp. 1-5.
[48] A. Gershman, N. Sidiropoulos, S. Shahbazpanahi, M. Bengtsson, and B. Ottersten, “Convex optimization based beamforming," IEEE Signal Processing Mag., vol. 27, no. 3, pp. 62-75, 2010.
[49] L. Zhang, Y.-C. Liang, Y. Xin, and H. Poor, “Robust cognitive beamforming with partial channel state information," IEEE Trans. Wireless Commun., vol. 8, no. 8, pp. 4143 -4153, August 2009.
[50] G. Zheng, K.-K. Wong, and B. Ottersten, “Robust cognitive beamforming with bounded channel uncertainties," IEEE Trans. Signal Process., vol. 57, no. 12, pp. 4871-4881, Dec. 2009.
[51] S. K. Sharma, S. Chatzinotas, and B. Ottersten, “Cognitive Beamforming for Spectral Coexistence of Hybrid Satellite Systems," in Cooperative and Cognitive Satellite Systems, S. Chatzintoas and et al, Eds Elseiver, 2015, ch. 13.
[52] S. A. Jafar, “Interference alignment- a new look at signal dimentions in a communication network," Foundations and Trends in Communication and Info. Th., vol. 7, no. 1, pp. 1-134, 2010.
[53] S. A. Jafar, “Interference alignment- a new look at signal dimentions in a communication network," Foundations and Trends in Communication and Info. Th., vol. 7, no. 1, pp. 1-134, 2010.
[54] S. Perlaza, N. Fawaz, S. Lasaulce, and M. Debbah, “From spectrum pooling to space pooling: Opportunistic interference alignment in MIMO cognitive networks," IEEE Trans. Signal Process., vol. 58, no. 7, pp. 3728-3741, 2010.
[55] B. Abdelhamid, M. Elsabrouty, and S. Elramly, “Novel interference alignment in multi-secondary users cognitive radio system," in IEEE Symp. Computers and Communications (ISCC), 2012, pp. 785-789.
[56] B. Guler and A. Yener, “Interference alignment for cooperative MIMO femtocell networks," in IEEE GLOBECOM, 2011, pp. 1-5.
[57] S. K. Sharma, S. Chatzinotas, and B. Ottersten, “Cognitive interference alignment for spectral coexistence," in Cognitive Radio and Networking for Heterogeneous Wireless Networks, D.
D2.3: Definition of reference scenarios, overall system architectures, research challenges,
requirements and KPIs Date: 01/02/2016
Page 121 of 131
SANSA-645047
D2.3
Benedetto and et al, Eds. Springer, 2014, ch. 2. [58] El Ayach, O., Peters, S., Heath R.W., J.: The practical challenges of interference alignment
wireless Communications, IEEE 20(1), 35–42 (2013). [59] V. Deslandes, J. Tronc, and A.-L. Beylot, “Analysis of interference issues in integrated satellite
and terrestrial mobile systems," in 5th ASMS conf. and 11th SPSC workshop, Sept. 2010, pp. 256 -261.
[60] M. Vu, N. Devroye, and V. Tarokh, “On the primary exclusive region of cognitive networks," IEEE Trans.Wireless Commun., vol. 8, no. 7, pp. 3380-3385, July 2009.
[61] Z. Wei, Z. Feng, Q. Zhang, and W. Li, “Three regions for space-time spectrum sensing and access in cognitive radio networks," in IEEE Globecom Conf., Dec. 2010, pp. 1-6.
[62] S. K. Sharma, S. Chatzinotas, and B. Ottersten, “Cognitive beamhopping for spectral coexistence of multibeam satellites," in Int. J. Satellite Commun. and Networking, March 2014.
[63] Weiss, M.B.H.; Altamimi, M.; McHenry, M., "Enforcement and spectrum sharing: A case study of the 1695–1710 MHz band," Cognitive Radio Oriented Wireless Networks (CROWNCOM), 2013 8th International Conference on , vol., no., pp.7,12, 8-10 July 2013.
[64] BATS FP7 project http://www.batsproject.eu/ [65] UWB ESA project [66] https://networks.nokia.com/ [67] F. Rusek, D. Persson, B. K. Lau, E. G. Larsson, T. L. Marzetta, O. Edfors, and F.
Tufvesson,“Scaling up MIMO: Opportunities and challenges with very large arrays,” IEEE SignalProcess. Mag., vol. 30, pp. 40–60, Jan. 2013
[68] 3rd Generation Partnership Project; Technical Specification Group Radio Access Network; Evolved Universal Terrestrial Radio Access (E-UTRA);Further advancements for E-UTRA physical layer aspects (Release 9), http://www.3gpp.org, 3GPP TR 36.814 V9.0.0 (2010-03)