-
ACTS Project 215Cellular Radio Access for Broadband Services
CRABS
D2P1B
SPECIFICATION OF NEXT-GENERATION OF LMDSARCHITECTURE
Project Number: AC215
Project Title: CRABS
Deliverable Type: K
CEC Deliverable Number: AC215/TEL/RD/DR/P/D2P1B/b1
Contractual Date of Delivery to the CEC: Nov. 1998
Actual Date of Delivery to the CEC: .2. February 1999
Title of Deliverable: Specification of next-generation of LMDS
architecture
Workpackage contributing to the Deliverable: WG2
Nature of the Deliverable: R
Abstract:This report describes the specification of the next
generation LMDS system. It contain a description ofthe overall
system as well as detailed simulation and anlysis of radio
sub-systems and protocols.
Keyword list: Broadband wireless access, LMDS radio sub-systems,
millimetre radio propagaion,modulation and coding schemes,
frequency plans and re-use strategies.
Editor: H. Loktu (TEL)
Contributions from: I. Bucaille (TOM), A. Burgess (EBL), G.
Coppola (PRM), D. Evans (PRL), V.Ferrero (POL), R. Germon (PBN), O.
Koudelka (IAS), T. Kourtis (DEM), A. Nordbotten (TEL), H.Loktu
(TEL), J. Norbury (RAL), M. Orifice (POL)
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
i
SUMMARY
In Project 215 Cellular Radio Access to Broadband Services
(CRABS) in the ACTS program,Working Group 2 has dealt with issues
related to system architecture and simulation studies. The
mainobjectives has been to study and specify the system
architecture for the next generation system ofwireless broadband
access systems with particular focus on Local Multipoint
Distributions Systems(LMDS). Next generation systems within this
context is understood as the commercial equipment andsystems
available on the market within 1 to 2 years from finalisation of
the project.
Targetting a combined residential and SME market scenario, a
medium to high interactivity terminalwas identified as the major
design objective. This was concluded to be very suitable for the
high-end ofthe residential customers, small offices or home offices
and the low-end of the SME customers.Furthemore, to enter the
residential market a low-cost approach to terminal design is
regarded asmandatory. To achieve this goal extensive re-use of
widely supported standards assumed to a necessarylimitation in
devising a solution within the time frame of the next generation of
LMDS equipment.
The proposed next generation LMDS system is capable of
supporting a full service access network withuser bit rate better
than 8 Mb/s on the uplink and 34 Mb/s dowlink. Using the DVB as the
baselinestandard familiy, low-cost user terminals will most likely
occur. It is designed for quasi-error-freeoperation and offers
availability figures better than 99.99 %. A cellular approach is
recommended withtypically 2000 residential customers within a
single cell.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
ii
TABLE OF CONTENTS
SUMMARY................................................................
................................ ................................
..................... I
TABLE OF
CONTENTS................................................................
................................ ..............................
II
GLOSSARY OF TERMS AND
ACRONYMS................................................................
............................. V
1.
INTRODUCTION................................................................
................................ ................................
..... 1
2. SYSTEM REQUIREMENTS AND DESIGN OBJECTIVES
................................................................ .
2
2.1 EVOLUTION OF MVDS AND LMDS IN AN EUROPEAN PERSPECTIVE
..............................................................
2
2.2 TARGET LMDS USER CLASSES AND SEGMENTS
................................................................
............................ 2 2.2.1 User
classes................................................................
................................ ................................
....... 3 2.2.2 Geographic user
segmentation................................................................
................................ ........... 4
2.3 LMDS SERVICE DEFINITIONS
................................................................
................................ ...................... 5 2.3.1
Classification of target services
................................................................
................................ ......... 6 2.3.2 User class
service configuration
................................................................
................................ ........ 7 2.3.3 Definition of
network
availability................................................................
................................ ....... 8
2.4 LMDS TERMINAL
CONFIGURATIONS................................................................
................................ ............ 9 2.4.1 IDU and ODU
terminal requirements
................................................................
................................ 9 2.4.2 User classes terminal
options................................................................
................................ ............. 9
3. LMDS SYSTEM ARCHITECTURE
................................................................
................................ ..... 11
3.1 OVERVIEW OF LMDS SYSTEM ARCHITECTURE
................................................................
........................... 11
3.2 INTERNAL TRANSPORT
NETWORKS................................................................
................................ ............. 12 3.2.1 Connection
to external networks
................................................................
................................ ...... 12 3.2.2 LMDS base station
feeder
network................................................................
................................ ... 13 3.2.3 Microwave BS feeder
networks
................................................................
................................ ........ 15
3.3 CELLULAR LMDS ACCESS NETWORK
................................................................
................................ ........ 18 3.3.1 The concept of a
LMDS cell
................................................................
................................ ............. 18 3.3.2 Cellular
subscriber coverage
................................................................
................................ ........... 19 3.3.3 Cellular
network capacity
................................................................
................................ ................ 20 3.3.4 Access
network
availability................................................................
................................ .............. 23
3.4 IN-HOUSE DISTRIBUTION NETWORKS
(IHDN)................................................................
............................. 25 3.4.1 Wireline options and
requirements for
IHDN................................................................
................... 25 3.4.3 Radio Local Area Network (RLAN)
................................................................
................................ .. 30
4. FREQUENCY PLANS AND REUSE STRATEGIES
................................................................
........... 33
4.1 INTRODUCTION
................................................................
................................ ................................
........ 33
4.2 FREQUENCY
PLAN................................................................
................................ ................................
..... 33 4.2.1 Dual frequency/dual polarisation
plan................................................................
............................. 33 4.2.2 Channel plan for
downlink................................................................
................................ ............... 34 4.2.3 Two or
more operators in the same franchise
area................................................................
........... 36
4.3 SERVICE MAPPING ONTO FREQUENCY
PLANS................................................................
............................... 36 4.3.1 Broadcast
channels................................................................
................................ .......................... 36
4.3.2 ATM data services
................................................................
................................ ........................... 38
4.4 MULTI-LOBE ANTENNAS TO ACHIEVE GREATER CAPACITY IN EACH
CELL................................. ..................... 41
4.5 FREQUENCY PLAN INTERFERENCE ISSUES
................................................................
................................ ... 42 4.5.1 Interference between
near adjacent cells
................................................................
......................... 42 4.5.2 Frequency allocation schemes to
limit interference in near adjacent cells
....................................... 43 4.5.3 The effect on
interference situations when using diversity improvement
........................................... 44
4.6 OTHER FREQUENCY BAND ISSUES
................................................................
................................ .............. 44 4.6.1 Bandwidth
allocated to backbone system
................................................................
......................... 44 4.6.2 Alternative cell plan which
produces better C/I
values................................................................
..... 44 4.6.3 Band plan for the extra 1 GHz of spectrum from 42.5
to 43.5GHz....................................................
46
5. LMDS RADIO
SUB-SYSTEMS................................................................
................................ .............. 49
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
iii
5.1 RADIO SUB-SYSTEM DESIGN FRAMEWORK
................................................................
................................ .. 49
5.2 TRANSMISSION AND
MULTIPLEXING................................................................
................................ ........... 51 5.2.1 Multiple
access schemes and framing structures
................................................................
.............. 51 5.2.2 Modulation and coding schemes
................................................................
................................ ...... 52
5.3 ANTENNA CHARACTERISTICS
................................................................
................................ .................... 55 5.3.1
Directive Base Station
antennas.................................................................
................................ ...... 55 5.3.2 Omni-directional
Base Station antennas
................................................................
.......................... 57 5.3.3 User Terminal
antennas................................................................
................................ ................... 58 5.3.4 Base
station antennas with more complex pointing system.
.............................................................. 59
5.3.5 Bandwidth considerations
................................................................
................................ ................ 60 5.3.6 Dual
linear (H/V) polarisation antennas
................................................................
.......................... 60
5.4 MILLIMETRE-WAVE EQUIPMENT AND TECHNOLOGIES AT 40 GHZ
................................................................ 60
5.4.1 Base station
................................................................
................................ ................................
..... 60 5.4.2 Consumer Premise Equipment
(CPE)................................................................
............................... 62 5.4.3 Review of 40 GHz MMIC
Technologies................................................................
............................ 65
6 RADIO MEDIUM ACCESS AND CONTROL
................................................................
...................... 69
6.1 PROTOCOLS FOR MEDIUM ACCESS
(MAC)................................................................
................................ . 69 6.1.1
LISSY................................................................
................................ ................................
............... 69 6.1.2
C-TDMA................................................................
................................ ................................
.......... 71 6.1.3
DAVIC................................................................
................................ ................................
............. 71
6.2 PROTOCOL PARAMETERS COMPARISON
................................................................
................................ ...... 72
6.3 SIMULATION RESULTS
COMPARISON................................................................
................................ ........... 73 6.3.1 Simulation
hypothesis
................................................................
................................ ...................... 74 6.3.2
End-to-end delay comparison
................................................................
................................ .......... 75
6.4 CONCLUDING REMARKS
................................................................
................................ ............................
76
7. CONLUSIONS AND
RECOMMENDATIONS................................................................
..................... 78
REFERENCES................................................................
................................ ................................
............. 80
APPENDIX A REGULATORY ISSUES, STANDARDS AND REFERENCE MODELS
....................... 82
A.1
INTRODUCTION................................................................
................................ ................................
........ 82
A.2 SPECTRUM
ISSUES................................................................
................................ ................................
.... 82 A.2.1 CEPT
................................................................
................................ ................................
.............. 82 A.2.2
WRC................................................................
................................ ................................
................ 82 A.2.3 ETSI Project BRAN Spectrum
Matters................................................................
............................. 83 A.2.4
DECT................................................................
................................ ................................
.............. 83
A.3 MPT1560
................................................................
................................ ................................
............... 83
A.4 ETSI AND DVB STANDARDS
................................................................
................................ ................... 83 A.4.1 DVB
................................................................
................................ ................................
................ 83 A.4.2 DVB LMDS Interaction Channel
Specification
................................................................
................ 84
A.4.3 Other Relevant DVB Specifications
................................................................
................................ . 86 A.4.4
DECT................................................................
................................ ................................
.............. 86 A.4.5 Additional ETSI Specifications
................................................................
................................ ........ 87
A.5
DAVIC................................................................
................................ ................................
................... 87 A.5.1 Introduction
................................................................
................................ ................................
.... 87 A.5.2 DAVIC LMDS Specification
................................................................
................................ ............ 88
A.6 ETSI PROJECT BRAN
................................................................
................................ ............................. 90
A.6.1 Introduction
................................................................
................................ ................................
.... 90 A.6.2 HIPERACCESS
................................................................
................................ ...............................
91
A.7 NETWORK ARCHITECTURE
................................................................
................................ ....................... 91
A.8 REFERENCE MODELS
................................................................
................................ ...............................
92
A.9 SUMMARY
................................................................
................................ ................................
............... 92
APPENDIX B - SUMMARY OF SIMULATION
RESULTS................................................................
..... 93
B.1 INTRODUCTION
................................................................
................................ ................................
........ 93
B.2 DOWNLINK RECEIVER CONSIDERATIONS
................................................................
................................ ... 93
B.3 ANTENNAS
................................................................
................................ ................................
.............. 94
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
iv
B.4 ARCHITECTURE FOR MASTER ANTENNA PASSIVE OPTICAL
NETWORKS.......................................................
94
B.5 INTERACTIVE
UPLINK................................................................
................................ ...............................
94
B.6 CHANNEL
................................................................
................................ ................................
................ 94
B.7 NETWORK CONNECTION
................................................................
................................ .......................... 95
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
v
GLOSSARY OF TERMS AND ACRONYMS
AAL ATM Adaptation LayerABR Available BitRateACTS Advanced
Communication Technologies and ServicesATM Asynchronous Transfer
ModeAWGN Additive White Gaussian NoiseBAP Broadband Access PointBER
Bit Error RatioBISDN Broadband Integrated Services Digital
NetworkBoD Bandwidth on DemandBS Base StationBWA Broadband Wireless
AccessC4 Communication, Computing, Contents and Consumer
electronicsCBR Constant BitRateCDMA Code Division Multiple
AccessCEPT Conferenc Europenne des Postes et Tlgraphes et
TlcommunicationsCORP CorporationCNR Carrier-to-Noise RatioCPE
Customer Premises EquipmentCPW CoPlanar WaveCRABS Cellular Radio
Access to Broadband ServicesDAVIC Digital Audio VIdeo CouncilDBS
Direct Broadcast SatelliteDECT Digital Enhanced Cordless
TelephoneDQPSK Differential Quaternary Phase Shift KeyingDRO D..
Running OscillatorDTH Direct-To-HomeDVB Digital Video BroadcastERC
European Radiocommunications CommitteeETS European
Telecommunications StandardETSI European Telecommunications
Standards InstituteFDD Frequency Division DuplexFDDI Fibre
...InterfaceFDM Frequency Division MultiplexFDMA Frequency Division
Multiple AccessGMSK Gaussian Minimum Shift KeyingGW GateWayHBT
Hetrojunction Bipolar TransistorHEMT High Electron Mobility
TransistorHH HouseHoldHFR Hybrid Fibre RadioIF Intermediate
FrequencyIDU InDoor UnitIHDN In-House Distribution NetworkIP
Internet ProtocolIRD Integrated Receiver DecoderISDN Integrated
Services Digital NetworkISM Industrial, Scientific and MedicalISO
International Standardisation OrganisationISP Internet Service
ProviderITU International Telecommunications UnionLAN Local Area
NetworkLED Light Emitting Diode
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
vi
LISSY Local network Interconnection using Satellite SYstemsLMDS
Local Multipoint Distribution SystemLNA Low Noise AmplifierLOS
Line-Of-SightMAC Medium Access ControlMAPON Master Antenna Passive
Optical Network architectureMCC Main Co-ordination CentreMMDS
Multi-channel Multi-point Distribution SystemMMIC Millimetre-wave
Monolithic Integrated CircuitMPEG Motion Pictures Expert GroupMPT
Ministry of Post and Telecommunications(UK)MVDS Multipoint Video
Distribution SystemMWS Multimedia Wireless SystemNIU Network
Interface UnitNRE Non Recurring EngineeringODU OutDoor UnitOFDM
Orthogonal Frequency Division MultiplexPABX Private Automatic
Branch eXchangePAR Peak-to-Average RatioPDH Plesiochronous Digital
HierarchyPHY PHYsical layerPON Passive Optical NetworkPOTS Plain
Old Telephone ServicePSTN Public Switched Telephone NetworkQoS
Quality of ServiceQAM Quadrature Amplitude ModulationQPSK
Quaternary Phase-Shift KeyingRF Radio FrequencyRLAN Radio Local
Area NetworkSDH Synchronous Digital HierarchySME Small Medium
EnterpriseSOHO Small Office/Home OfficeSONET Synchronous Optical
NETworkSTB Set-Top BoxTCP Transport Control ProtocolTDM Time
Division MultiplexingTDMA Time Division Multiple AccessTS Transport
StreamTV TeleVisionUMTS Universal Mobile Telephone SystemVBR
Variable BitRateVoD Video on DemandVSAT Very Small Aperture
AntennaWAN Wide Area NetworkWDM Wave Division MultiplexingWRC World
Radio ConferanceWWW World Wide Web
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
1
1. INTRODUCTION
The main objective of this deliverable is to provide a
specification of the next generation of LMDS(Local Multipoint
Distribution System) systems with focus on deployment in a
pan-European market. Inthis context next generation system is
understood as being commercially available within 1 to 2 yearsfrom
the finalisation of the project. Furthermore, this is also intended
to be the final report on theresearch activities conducted within
Working Group 2. As such this document will not resemble asystem
specification in the strict sense but merely reporting
recommendations and guidelines forpractical and economic system
design.
The emerging arenas for broadband access in Europe are currently
characterised by a ongoingderegulation of the traditional market
segments for services. This implies that the traditional division
ofmarkets into telecom, datacom and broadcast services is breaking
down and are gradually replaced by amultiple service scenario in a
strongly competitive marketplace. Broadband access networks are
goingdigital and consequently a diverse range of both wireline and
wireless technologies will be capable ofcovering a wide range of
services. As a consequence of this trend operators of broadband
accessnetworks are also in position to target a broader range of
user segments in the market.
This collapse of the traditional market structure is often
referred to as C4 convergence and constitutes aconsiderable
challenge for all players in the market of broadband access. To
provide profitableoperation in this environment, the operators of
access networks must in general be able to operate at lowtake-up
rates and with rapidly changing service requirements. To be
successful under thesecircumstances the demand for a highly
flexible, scaleable and cost-effective access technology isobvious.
In the context described above broadband wireless access (BWA)
networks offer a verypromising/exciting alternative which is
expected to fulfil most of the above mentioned requirements.
In terms of customers, the major objective is to specify a
digital LMDS system suitable for deploymentin the residential and
small-to-medium enterprise (SME) markets. The main source of
revenue frombroadband access in the residential has traditionally
been broadcasting of TV programs.Therefore, the provision of
digital MPEG-2 based TV is a mandatory requirement and consequently
thewell established DVB standards are chosen as a initial framework
for system design. Furthermore, high-speed Internet is believed to
become a mandatory requirement for the SMEs and as well as
forresidential customers in the near future. A further increase of
this service bouquet is envisaged and toenable a flexible handling
of a highly dynamic service mix , ATM has been chosen as the basis
for theswitching functionality of the system
To be able to offer both telecom and broadcasting services
within the same band, there is an obviousadvantage from a
regulatory point of view that fixed services (FS) and broadcasting
services ( BS) areallocated on a coprimary basis in a given
frequency band This requirement in fulfilled in the band 40.5-42.5
GHz which is target band for the CRABS project.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
2
2. SYSTEM REQUIREMENTS AND DESIGN OBJECTIVES
2.1 Evolution of MVDS and LMDS in an European perspective
During the first half of the 1990s, the concept of Multipoint
Video Distribution Services (MVDS) forcellular television was being
discussed actively in Europe. The CEPT organisation had early
identifiedthe band 40.5 to 42.5 GHz (as well as lower frequency
bands) for such services which lead topublication of the CEPT
recommendation TR/52-01 [1] published in 1992. This recommendation
hasby the end of November 98 been ratified by the majority of CEPT
member countries. At that timeMVDS where mostly considered for
analogue TV distribution systems, with a possibly limited elementof
interactivity for services such a video on demand (VoD). The
standard for analogue MVDS in UKcalled MPT 1550 [3] emerged already
in 1993.
In parallel the standardisation of systems for digital TV
distribution in Europe progressed rapidlyheaded by the EBU body DVB
and a standard for digital TV distribution via satellites [2]
emerged in1994. This standard was quickly adopted by satellite
broadcasters and satellite operators in thetransport part of the
distribution network because it reduced the distribution cost per
TV programdramatically. As the price of DVB compliant set-top-boxes
(STB) dropped it became evident that thiswould pave the way for a
digital MVDS standard for cellular TV distribution based on DVB
standards.This would allow more TV channels, transparent
interconnection of the transport and access networkand reuse of the
DVB compliant STBs altogether an low-cost approach suitable for
targeting theresidential market. Besides, the rapid growth in
commercial Internet services at that time demonstrated afuture
demand for an interactive system. Recognising that, UK Radio Agency
made a revision of MPT1550 based on an interactive digital approach
named MPT 1560 [4] published in June 1996. Thisstandard was later
with minor adjustments adopted by ETSI as the first European
standard for digitalinteractive MVDS systems [5].
At this stage the band 40.5-42.5 GHz still was allocated for
broadcast services (BS) by ITU-R withboth terrestrial broadcasting
and satellite broadcasting on a coprimary basis. This was changed
byWRC-97 where fixed services (FS) was allocated on a coprimary
basis with the other two [6] Thisdecision prepared the band for
interactive services from a regulatory point of view and as such
enablesthe convergence of telecom, datacom and broadcast services
within a single band. This intention wasfurther enhanced by the
CEPT proposal late this year which recommends this band for
multimediawireless systems (MWS) including MVDS.
Considering the significant change in the use of the 40.5-42.5
GHz band, the term MVDS do no longerconvey an adequate description
of the actual services expected to be provided by a next
generationsystem. In the US the chosen acronym for wireless BWA
systems with more or less the same servicecapability, is LMDS.
Initially launched in 28 GHz band in the US, these systems are
still applicable inthe 40 GHz range. Consequently, ETSI has adopted
this term and a standard for the interactive pathwas issued this
year [7]. In general, LMDS is now in a broader context understood
by ETSI as a digitaltwo-way BWA system above 10 GHz with an
asymmetric transmission bit rate. Hence, in the followingthe term
LMDS will be used for the purpose of specifying the next generation
system in the chosenfrequency band.
2.2 Target LMDS user classes and segments
When specifying the target services for the next generation LMDS
system, one should keep in mind thepossible deployment scenarios
for LMDS systems. In general the regulatory environment today
ischaracterised by deregulation of the traditional market
structure. This development will generate ahighly competitive
marketplace where different technologies partly will deliver the
same services.Consequently, the LMDS systems must be capable of
delivering service quality (and availability) with acompetitive
price-quality( and availability) relationship, covering a broad
range of regulatory
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
3
100 10
1
.0 0
.1B
USI
NES
SR
ESID
ENTI
AL
Downlink Capacity [Mb/s]
Uplink C
apacity [M
b/s]SME
CORP.
0.1 1.0 10 100
HH
SOHO
Figure 1 Segmentation into LMDS user classes
regimes. These would range from a single service up to a full
multimedia service configuration.Furthermore, various combinations
of user classes and geographic segments will be encountered
leavingthe system with a very complex and partly contradictory set
of requirements.
2.2.1 User classes
The users are traditionally divided into two categories, namely
business and residential users. Businessusers are different from
residential users in many aspects. The major part of business
customers aredemanding symmetric connections whilest the major part
of the residential users are characterised byvery asymmetric demand
from a broadband services point of view. Another important feature
is thelarge difference in investment and tariff payment
capabilites.
There is a variation within each category and to obtain a more
useful representation of the customers,each user category is
divided into two classes. Business users are split into corporate
users (CORP.)and small- and medium-sized enterprises (SME) whilest
residential users are divided between small-office and home-office
(SOHO) and private households (HH). A graphical representation of
the typicalmaximum uplink and downlink capacities requested by
these user classes are shown in figure 1.
2.2.1.1 Corporations (large businesses)
Corporations have large demand for capacity ranging up symmetric
155 Mb/s connections andoccasionally above. These connections
usually carry aggregate traffic with low burstiness factor
oftenused for interconnection of LANs and PABXs by use of leased
lines. A new an interesting application ishigh-speed connection to
Internet.
2.2.1.2 Small- and medium-sized enterprises (SME)
SMEs are defined as businesses with up to 250 employees. The
majority of businesses throughout theEuropean market are SMEs and
as such they constitute an important group of customers. Due to
theconsiderable number of employees the connections carry aggregate
traffic typically up to symmetric E1
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
4
(2 Mb/s) and E2 (8 Mb/s) capacities mostly by leased lines. The
high-end of this class will or hasalready migrated to higher
capacities like symmetric E3 (34 Mb/s) or STM-1 (155 Mb/s).
2.2.1.3 Small-office and home-office users (SOHO)In practice,
the SOHO class contains both residential and business users but has
more in common withthe residential users when considering the
traffic characteristics. Due to the low number of
individualsinvolved, the burstiness factor is relatively high and
switched connections are preferred instead of leasedlines. To cover
several simultaneous services including at least multi-party video
conferenceing and webbrowsing, the maximum connection capacity
should be in at least 768 kb/s uplink and 2 Mb/s downlink.
2.2.1.4 Private households (HH)
The connection capacity required by private households are to a
large extent determined by entertain-ment services. Reception of
broadcast digital TV requires typical 6 Mb/s downlink to ensure
qualitywhich compares with analogue PAL transmission. In uplink
direction, the expected demand for mid-quality video conferencing
or simultaneous connections of various type, requires a maximum
connectioncapacity in order of 384 kb/s.
2.2.2 Geographic user segmentation
In devising a optimum system design, knowledge about local
demographics is required. Due to largevariations between different
countries, it is impossible to make a simple model covering a
pan-Europeandeployment scenario However, to be able to suggest an
efficient cellular structure (including cell size)for a given
environment, a simple model must be established.
A widely used approach is to classify the users according to the
average density and range for a givenarea. The average is
calculated per cell area in km2 and the range is include to give a
first impression ofthe corresponding variance. A segmentation model
covering European households is shown in Table 1
Table 1 Typical Household (HH) densities in Europe
One should keep in mind that the local peak density could be
very large compared to the average value.There are no well defined
boundaries between the proposed demographic segments and
particularly thetransition from suburban to urban area depends very
much on local conditions.
Similar tables could be made for SOHO and SME user classes where
densities in general would bemuch lower that that of private
households. The SOHO user would display similarities to the HH
usersin the sense that home office part would mostly be located in
private households. Throughout
Demogr. segment Rural Suburban Urban City centreHH density mean
100 1000 3000HH density range 5 - 500 500 - 3000 1000 - 8000 8000 -
30,000
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
5
0.01 0.05 0.1 0.2 0.5 1 2 3 5 8 100
500
1000
1500
2000
2500
3000
3500
4000
HH
's pe
r c
ell
HH Density [1000/km 2 ]
LM DS cell size
0.5 km 2
1.0 km 2
2.0 km 2
4.0 km 2
8.0 km 2
25 km 2
Figure 2 Relationship between HH density and number ofHH per
cell with different cell sizes
Europe the 99 % of all enterprises have less the 250 employees
[8] and consequently belong to the SMEand SOHO classes. Per country
basis the number of enterprises amounts to 5-10 % of the total
numberof private households.
The aggregate amount of traffic generated within a single cell,
is partly determined by the number ofusers connected. The number of
users are related to user class density by the inverse of the cell
size insquare kilometres. To predict the approximate cell size
range related to the different demographicsegments, the number of
HH per cell is plotted against HH density for cell size areas from
0.5 km2 to 50km2. Assuming a square cell configuration with 0.5, 1,
2, 4, 8 and 25 km2 cell size, the correspondingmaximum ranges (or
diagonal of the square) are approximately 1, 1.5, 2, 3, 4 and 7
km.
If we assume a maximum number of HH per cell of 2000 the results
in Figure 2 suggests that for therural segment a cell maximum.
range of 3 km and up is adequate. For the suburban case between 1
and3 km is appropriate, whilest the urban case gives 0.7 to 2 km.
Using the mean densities given in Table 1we arrive at 6, 2 and 0.9
km as typical average cell maximum. ranges. As already pointed out,
these areaverage figures and large local variation must be
expected.
2.3 LMDS service definitions
Specifying a target service scenario for the next generation of
LMDS systems, one should keep in mindthat the services as such are
in general not specific to a LMDS system. Rather the specific
features ofLMDS could be the service attributes by which the actual
services are delivered to the customer.Service attributes being
employed in the following simplistic approach are uplink and
downlink peakbitrate, traffic class and network/service
availability.
The bitrates are specified according to established standards
like fractional E1 (N*64 kb/s), E1 (2Mb/s), E2 (8 Mb/s) and E3 (34
Mb/s). This is done only for the matter of compatibility with
existingequipment and does preclude the use of other bitrates. The
traffic classes options are defined
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
6
Service Classes Uplink (kb/s) Downlink NetworkMean Peak Mean
Peak Traffic Avail.
Low grade 64 64 64 64 CBR 99.99%Video Conf. Medium grade 384 384
2.048 2.048 CBR 99.99%
High grade 2.048 2.048 8.448 8.448 CBR 99.99%On Demand 16 64
3.200 4.096 VBR 99.99%
Video/TV Broadcasting - - 8.192 8.448 CBR 99.9%Contribution 8192
8192 64 64 CBR 99.99%IP-based 16 64 16 64 VBR 99.9%
Voice PSTN/ISDN 64 64 64 64 CBR 99.99%PABX Interconnect 2.048
2.048 2.048 2.048 CBR 99.99%Low grade 8 64 64 512 ABR 99%
Web Medium grade 64 512 512 2.048 VBR 99.9%High grade 512 2.048
2.048 8.192 CBR 99.99%Cellular backhaul 2.048 2.048 2.048 2.048 CBR
99.99%
Transport LAN Interconnect 8.448 8.448 8.8448 8.448 CBR
99.99%ATM UNI 25.6 25.600 25.600 25.600 25.600 V-CBR
99.99%TV-multiplex distrib. - - 34.368 34.368 CBR 99.9%
Table 2 Service classification by link capacity and network
characteristics
as packet-oriented (VBR and ABR) and circuit-oriented (CBR)
referring to the established terminologyfor ATM multiplexing. The
last attribute is network availability which refers to the
percentage of a yearfor which the network is should be available
for delivery of services with the specified performance.This do not
guarantee availability of the service as such because it depends on
traffic loading conditionsetc.
2.3.1 Classification of target services
2.3.1.1 Video services
Video services are expected to become a very important feature
of the next generation broadbandscenario. Broadband systems are in
general foreseen to enable the transition from voice to video
basedcommunications between individuals and between producers and
consumers as well. Most videoservices are circuit-oriented and
require high network availability except for TV broadcasting
whichusually operate with medium availability (99.9 %). Video
conferencing is a typical case wheredifferentiated service quality
is expected. Hence, a tree service grade levels is proposed to
cover thedifferent user classes and configurations. For the medium
and high grade options multi-party videoconferencing is suggested
in which asymmetric peak bit rates must be accommodated.
2.3.1.2 Voice services
Voice related service have been and will be important part of
the next generation LMDS servicescenario. In addition to
traditional PSTN/ISDN based telephony for communication
betweenindividuals, a low to medium grade option is becoming
available through IP-based telephony.PSTN/ISDN will occasionally
require extremely high network availability (above 99.995 %)
whichusually would not be achievable for next generation LMDS
systems. From a market point of view therealso seems to be a
considerable demand for aggregate voice traffic for instance
through interconnectionof PABXs.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
7
Service Classes HouseHold SoHo SMELow High Low High
Low Qual. 1 1 0 0 0Video Conf. Mid Qual 0 2 2 1 0
High Qual. 0 0 1 2 2On Demand 0 1 2 2 1
Video/TV Broadcasting 2 2 1 0 0Contribution 0 0 0 1 2IP-based 0
1 1 2 2
Voice PSTN/BRI-ISDN 2 2 2 1 0PABX Interconnect 0 0 0 1 2Low
grade 2 1 0 0 0
Web Medium grade 1 2 2 1 0High grade 0 0 1 2 2Cellular backhaul
0 0 0 1 2
Transport LAN Interconnect 0 0 1 2 2ATM UNI 25.6 0 0 0 1
2TV-multiplex distri 0 0 0 0 1
Table 3 Typical user class service configuration according
toprimary(2), secondary(1) and optional (0) services
2.3.1.3 Web browsing and Internet access
The far most important service for the next generation LMDS
system is believed to be access to Internetand the world wide web
(WWW). To cater for the expected diversity in service requirement,
threeservice grade levels are proposed. They are differentiated in
bit rate, traffic class and networkavailability. The bit rate
peak-to-average ratio (PAR) is proposed to be largest for low and
partly themedium grade service reflecting a presence of low
interactivity users. The high grade users will be moreinteractive
and may request broadband streaming capability and large file
transfers both uplink anddownlink. Hence, the bit rate PAR is
generally lower and the asymmetry between uplink and downlinkwould
be weaker than in the case of a low or medium grade service.
2.3.1.4 Transport networking
The inherent broadband capabilities in a next generation LMDS
system, makes it suitable for transportnetwork of for instance
narrowband traffic. Cellular backhaul (or feeder networks) for GSM
or futureUMTS networks is a evident service and interconnection of
LANs is another. These services arecharacterised by high network
availability and symmetric connections using a circuit-oriented
approach.In case of satellite TV broadcasting, the possibility of
redistributing a full MPEG-2 multiplexedconnection containing for
instance 4 to 8 digital TV channels is an interesting opportunity.
This requiresa downlink only and could enable a seamless
integration of broadcasting into the next generation
LMDSsystem.
2.3.2 User class service configurationThe different types of
users specified in section
2.2.1 User classes, will request a different configurationof the
services defined in Table 2. To be able to confine the LMDS system
design for an adequate set ofuser classes and segments, a target
service configuration for each user class must be defined.
Aproposal for service configuration of the defined user classes are
shown Table 3. The indicatednumbering suggests a ranking in
primary(2), secondary(1) and optional(0) services. This does
not
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
8
preclude the operation of other service configurations but
merely constitutes a framework for optimisedLMDS system design.
2.3.2.1 Small and medium enterprises (SME)
SMEs are characterised by a group of people performing some kind
of production activity whichusually generates a significant revenue
stream. Their major use of services relates to a
professionalenvironment and they will therefore mainly request high
grade broadband services. SMEs are expectedto be the major user of
interactive broadband video service in both contribution, video on
demand andmulti-party video conferencing. They will have modern
computer facilities and will probably be themajor user of
IP-telephony due to their demand for long distance phone calls. Due
to their size they alsomay require transport network facilities
There is still a monopoly in most countries on symmetricbroadband
services (E1 and above). Consequently, a large potential for leased
line replacement isenvisaged.
2.3.2.2 Small Office HomeOffice (SOHO)SOHOs have much in common
with the low-end of the SME user class. A professional element is
beingpresent either as a stand-alone business or as residential
home office connected to a SME or CORP.Consequently, one may expect
medium or secondary high grade services being requested and
Internetaccess and video conferencing will be important.
IP-telephony will be used if long distance phone callsare vital.
SOHO users also have much in common with the high end of the HH
user class.
2.3.2.3 Private households (HH)
There are tree major services for private households. They are
PSTN/ISDN telephony, low and mediumgrade Internet and TV
broadcasting. The latter have been the major source of revenue for
broadbandaccess in this user class so far. And that will probably
not change within the timeframe of the nextgeneration LMDS systems.
The high-end users will have more data-centric service requirements
addingvideo conferencing as a primary service.
2.3.3 Definition of network availabilityThe purpose of define
network availability is to provide a design objective for
specification of theLMDS radio up- and downlink. From a radio point
of view this figure relates to time percentage of anaverage year in
which the radio link carrier-to-noise ratio (CNR) exceeds a fixed
design figure.Inversely, the outage time is defined as 100 % minus
availability shown the time of a average year inwhich connection is
not available. A summary of different availabilitys and
corresponding outage timesare given in Table 4
Grade Availability Outage timeLow 99 % 3d 15h 40minMedium 99.9 %
8h 46minHigh 99.99 % 52minExtreme 99.999 % 5min
Table 4 Availability and corresponding outage time inpercentage
and time of an average year
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
9
Figure 3 A generic system Reference Model for
InteractiveSystems
2.4 LMDS terminal configurations
2.4.1 IDU and ODU terminal requirements
User terminals for the next generation of LMDS system will in
general be divided into a outdoor unit(ODU) and a indoor unit
(IDU). The ODU will consist of a roof-mounted antenna closely
integratedwith the microwave components necessary to do filtering,
amplification and frequency translation ofsignals both for in the
uplink and downlink direction whilest the IDU performs the baseband
functionsrelated to uplink and downlink transmission.
The ODU should be one piece of equipment to enable simple
installation at the customers premises. andthe physical interface
should be at a L-band intermediate frequency (IF). This would
enable possible re-use of IDU equipment manufactured for other
frequency bands and access technologies. The IDUshould in principle
have a common design for at least LMDS frequency bands above 20
GHz.According to the ETSI reference model [9] shown in Figure 3,
IDU should logically consist of anetwork interface unit (NIU) and
set-top-unit (STU). In DVB terminology the IDU is referred to as
aset-top-box (STB) and another often used term is residential or
commercial gateway (GW) most oftenused with reference to the
business users. Preferably the STB should consist of a single piece
ofequipment but that does not seem feasible for all service
configurations. envisaged.
2.4.2 User classes terminal options
To devise a single/generic user terminal suitable for the whole
range of user classes and serviceconfigurations solution is not
regarded as a viable solution for the next generation LMDS system.
Ingeneral a LMDS system tailored to a specific service
configuration provides a poorer cost-performancerelationship when
used for largely different service configurations. Especially, when
narrowbandservices only is delivered by a true broadband system,
one should usually expect to find better way ofproviding delivery.
However, the possibility to add narrowband service onto broadband
services, couldcreate a competitive narrowband case if accepted
from a regulatory point of view.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
10
B A S ES T A T I O N
C U S T O M E RU N I T
D V B
W W LD E C T
H I G HD A T ARATEA T M
M P E GC O D I N G
L O W R A T ER E T U R N
P A T H
I S D N U S I N GT D M A / F D Dfor 64 kb i t /s
channe l s
low cost un i tsw i th N x ISDNC H A N N E L S
Capac i tyequ iva lentto 2 Mbi t /s
o r more
Pr iceappropr ia te to
S M E s
interact ivelink
2 Mb i t / s
Desktop computer
Voice
Desktop computer
Ethe
rnet
Server
T V & S T B
L A N
Fax
34 Mb i t / sdown l ink
Figure 4 LMDS user terminal options.From top and downwards
isshown the low, medium and high interactivity configuration
Considering the user class service configurations proposed in
Table 3 there is tree obvious targetterminal options as shown in
Figure 4.
The low interactivity terminal covering the low-end of the HH
class, which are mainly requestingentertainment oriented service
based on TV broadcasting and low grade Internet service. This
approachis compliant with the architecture adopted in the ETSI
reference model in Figure 3. and requires thatuplink is provide
through PSTN or ISDN. The medium interactivity terminal will cover
the high-endHH, SOHO and parts of the low-end SME users. Both
uplink and downlink would typically requirefrom fractional to full
symmetric E1 capability. And finally, the high interactivity
terminal coveringmost of the SME user class requesting up to
symmetric E2 capability. The medium and highinteractivity terminals
both provide an integrated radio uplink with true broadband
capabilities.
In terms of service configuration and terminal architecture the
medium and high interactivity terminalshave much in common.
Furthermore, introduction of broadband services in the market is
expected totake place in the business user class. Hence, especially
SMEs will be important in the first phase oflaunching next
generation LMDS systems. Later as the terminal cost goes down, SOHO
and high-endHH users will be included. Consequently, a common
terminal for medium and high interacitivity usersare proposed for
the next generation LMDS system. The low interactivity users should
be served with aseparate hybrid solution with a broadband radio
downlink and uplink provided by a narrowband accesstechnology.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
11
3. LMDS SYSTEM ARCHITECTURE
3.1 Overview of LMDS system architecture
A LMDS system is in the following understood as consisting of
all network elements necessary todeliver broadband services to
users within a LMDS franchise. Hence, the LMDS system does
notdirectly deal with provision of services. The choice of a
network centric approach follows directly fromthe interpretation of
the acronym LMDS as a generic delivery mechanism for local
services. And assuch the LMDS system must to a large extent be
based on a service independent architecture and avoidservice
specific elements as far as possible. An example of a LMDS system
architecture is shown inFigure 5.
From a network point of view, a LMDS system will in general have
a four layer structure. The firstlayer originated/starts from
the
LMDS control centre which is the broadband access point
(BAP)connecting the LMDS system to external networks. This node in
the core network is usually theentrance point for service provision
by for instance ISPs. The LMDS Control Centre is connectedthrough
an internal transport network to the LMDS main co-ordination centre
(MCC) both whichoccasionally may be located at the same place.
The LMDS main co-ordination centre defines the transition to the
second layer and is the networkentrance point to the actual
geographical area in which services are delivered by the LMDS
system. Atthis level, satellite broadcast TV may be inserted to
lower the demand for transport network capacitybetween the control
and main co-ordination centres. An overlaid transport network
originates from themain co-ordination centre providing connectivity
with/to the different cells within the LMDS franchise.It will
provide intra-cell connectivity and the practical implementation is
very much dependent on theparticular deployment scenario. The main
co-ordination centre are usually located at one of the basestations
at the edge of the LMDS franchise in which the cell are defined as
a master cell.
Each cell is feed by a LMDS base station which defines the
transition to the third level. The main co-ordination centre is
connected to the base stations through the overlaid transport
network. A single basestation may feed several cells and provides
connectivity to a large number of user network terminationsusing
radio communications in the 40 GHz band. This connectivity may be
obtained by directcommunication or indirectly by use of transparent
radio repeaters if necessary. This third network levelof the LMDS
system is hereafter defined as the LMDS access network or LMDS
network for short.This complies both with the interactive reference
model in Figure 3 and the general reference modeladopted by the
ETSI BRAN project [8].
To be able to provide the different services at the physical
location where the users want to connect, anin-house distribution
network (IHDN) is required. This part of the LMDS system defines
the fourthnetwork level. The actual implementation of IHDN network
may be done using for instance radio,coaxial or optical
cables/technologies. In supporting multi-user dwelling units in a
multi-servicescenario, the design of a cost effective and practical
IHDN network is of vital importance. Generally,this level starts at
the NIU as shown in Figure 3.
Although the LMDS system is described as having four network
layers, there will in practice be caseswhere a transparent
transition between the different levels are envisaged. This is for
instance done whenemploying hybrid fibre radio (HFR) technologies
at level two to feed the base stations. From a networkpoint of
view, the layer structure will collapse into a simpler structure
which may be advantageous froma economic point of view.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
12
cell clusters inurban areas
multichanneldigital satellite
broadcastreceiver
main co-ordination
centre
trunk telecommunicationstraffic eg telephone,
internet, videoconferencing ,
SME leased lines, etc
larger cell inrural area with
near omnidirectional
antenna
micro cell
fibre/coaxconnection to main
cell clusters
radio connectionto more remote
cells
LMDS basestations
" m u s h r o o m s h a p e "
Fig .3 : Sc h em e o f th e mu sh ro o m an te nn a.
D
Figure 5 Overview of LMDS system architecture
3.2 Internal transport networks
3.2.1 Connection to external networks
Connection to external networks goes through the LMDS Control
Centre. If the Control Centre and theMain Co-ordination Centre are
located at the same place, the external networks are feed directly
by theLMDS base station feeder network. Conversely, the traffic
concentrated at the Main Co-ordinationCentre would be relayed to
the Control Centre using a high capacity trunk connection. An
exceptionmay be experienced if we consider an external satellite
broadcasting network where a direct reception atMain Co-ordination
Centre would relax the capacity requirement on the inter-connection
of the MainCo-ordination Centre and the Control Centre.
The choice of transport network technology for this part of the
LMDS system depends on severalfactors. First of all, the
demographic characteristics of the area will impose an upper limit
on therequired aggregate transport capacity. In areas with high
traffic volume, a Control Centre will usuallybe established within
reasonable distance and consequently a fibre optical connection
would often bepreferred. If the distance to the Control Centre is
large, a connection using point-to-point radio may beadequate when
the traffic volume is not exceeding the transport capability of
radio. This constitutes atypical case in the early stage of
deployment and will be applicable for low number of customers
in
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
13
general regardless of the actual demographic characteristics. In
practice radio is expected to offer aeconomic solution in the range
STM-1 (155 Mb/s) to STM-4 (622 Mb/s). For the case of broadcastonly
or isolated cells far away from the Control Centre, a satellite
connection may be adequatedepending on the particular LMDS cell
service configuration
3.2.2 LMDS base station feeder networkThe LMDS base station
feeder network corresponds to level two as defined in section
3.1 Overview ofLMDS system architecture. To deliver services
provided by external networks, each LMDS cell must beconnected to
Main Control Centre via a BS feeder network. In a multi-cell
deployment scenario, there isa vast range of possible configuration
for feeder network design. To devise an optimised solution,detailed
knowledge about the expected service configuration profile for the
customers within eachLMDS cell must be available.
Considering a multi-cell LMDS network, the preferred approach is
to connect the LMDS cells to theMain Co-ordination Centre using an
overlaid inter-cell transport network. This network areimplemented
by connecting the cells in a hierarchical manner. The BS at the
Main Co-ordination Centreis connected to one or more base stations
which in turn again are connected to one or more base stationson
the next level. In general, a nested tree structure is produced
when repeating this procedure. In somecases, this structure is
referred to a master-slave configuration. However, the base
stations mayestablish both inter-cell and intra-cell communication
independently, effectively bypassing the BS at theMain
Co-ordination Centre. Thus, the master-slave terminology is not
recommended in the general case.
Due to the short distance between the Main Co-ordination Centre
and the nearest(master) BS, thisconnection will usually be
implemented using either coaxial or optical fibre technologies. As
for theinter-cell network, a microwave radio or optical fibre
technologies are usually preferred. In principle,both alternatives
may be used to implement a physically transparent inter-cell
network in which theLMDS access network and the BS feeder network
only differ in the characteristics of the transportmedium.
Generally, the transition between the inter-cell and LMDS access
network is non-transparenteither with different physical
transmission formats or with regenerating base stations to allow
fordistributed network functionality like local switching.
Three options for inter-cell network implementation are shown in
Figure 6. The upper and middlealternatives are both based on
microwave radio and show transparent (upper) and
non-transparent(middle) implementations/configuration. In the
transparent case the downlink signal in the leftmost cellis
received by off-air pickup by the adjacent cells. Consequently, the
same information content isdelivered in each cell. The LMDS cell
uplinks are concentrated at the BS and relayed to the same cellfrom
which the downlink was received. Optionally, the uplink may be
provided by wireline alternativesif low uplink capacity is
requested. The non-transparent alternative employs a separate
microwave inter-cell backbone where the transport stream may be
reformatted and remultiplexed before entering the nextnetwork
level. The lower alternative shows the optical fibre option which
may be transparent as well asnon-transparent. A typical transparent
option is the emerging HFR technologies where all digitaloperations
are performed at the Main Co-ordination Centre and the base
stations only act as passiverepeaters. Non-transparent
implementations are very similar to the microwave radio alternative
andSDH connections carrying ATM in different configurations are
typically used.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
14
Figure 6 Options for inter-cell networking in a multi-cellLMDS
network configuration
(a) (b)
Figure 7 BS feeder network topologies for continueous
andsystematic radio coverage
To provide a continueous radio coverage throughout a wide area,
a systematic procedure for designingthe BS feeder network must be
employed. The choice of network topology and the technological
solutionare here determined by the actual service configuration and
traffic volume. Assuming a regular four cellclustering around a
single BS, two possible systematic feeder network topologies are
proposed in.Figure 7. Figure 7a shows a daisy-chain configuration
where consecutive base stations are arranged inparallel paths with
duplex transport connections. This topology is suitable for
implementation oftransparent feeder networks and especially for the
fibre optical solution were civil works cost can beshared when one
optical fibre are feeding each BS. We also see that the uplink and
downlink feeder linkpaths map into the same route. If a
non-transparent implementation is preferred, this can be done
usinga single high-capacity SDH ring along each chain. An
alternative solution is shown in Figure 7b whichis suitable for
non-transparent implementation especially for SDH microwave feeder
networks. Theuplink and downlink feeder link paths here map onto
several different routes and consequently feedernetwork route
diversity is an inherent feature. Note the maximum number of
simplex feeder link
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
15
connections at each BS is four and the total number is lower
than for the daisy-chain configuration. Thisconfiguration may be
implemented using several overlaid microwave SDH rings. Due to the
largenumber of branches in the topology, this configuration is in
general not suitable for fibre opticalimplementation.
3.2.3 Microwave BS feeder networks
3.2.3.1 Broadcast services delivered by off-air pickupThe first
cell cluster would normally be fed through a fibre/coax connection.
However, it might bedesirable to feed the next cell cluster or even
an individual cell in the system, through off-air pick-up ofthe
broadcast channels, until it is economic or timely to install a
fibre/coax link . Assuming the four cellclustering configuration in
Figure 7, the distance between two adjacent base stations would be
twice theinternal cell distance. Choosing a maximum radio path
length of 2km, two neighbouring base stationswould be separated by
at least 4 km. We need to have a line of sight between them as
shown in Figure 6and the required outage for broadcast services is
about 99.9 %. The additional free space loss fromdoubling the range
is 6dB and the increased rain margin of 7.5 dB in rain zone L. If a
typical customerantenna with 35 dB gain was deployed to perform the
off-air pick-up, the overall margin on thedownlink would be reduced
by 11.5 dB in case of rain zone L.
The rainfall margin in zone L must according to Figure 13 be 25
dB for 2 km radio path length at 99.99% availability and 44 dB at 4
km. At 4 km, 99.9 % availability requires 17 dB rainfall margin. If
thefirst cell cluster provide an availability of 99.99 %, a
broadcasting channel availability of 99.9 % mayby achieved directly
at the next cluster because the new margin is 25-6 = 19 dB which is
2 dB above therequirement of 17 dB. To achieve 99.99 % at 4 km, an
additional margin of 44-25+6 = 25 dB must beprovided which could
only be reached with a 35+25 = 60 dB gain antenna. In terms of
practical size ofthe antenna, a availability of 99.95 % would be
more adequate. It is possible that the separationsbetween the base
stations may be more than 4 km with a reduction in the availability
achieved in thesecond cluster. For instance, a separation of 8 km
would increase the rain margin required to provide a99.9 % service
to 32 dB. An antenna with a gain of 54 dB would be needed for the
off-air pick-upreceiver to provide 32 dB margin.
In principle the broadcast signals could be cascaded to a
further cell clusters through a tandemconnections. However the
availability drops in each hop of the cascade. The availability
achieved in thenth hop is a1*a2*.*an , ai is the availability
achieved in the ith cell. If the availability on each hop is
say99.95 % then the fourth cell in the cascade cannot be better
than 99.8 %, i.e. still close to the broadcastrequirement. It is
likely that the second cell cluster would be treated in much the
same way as a microcell. The off-air signal would be amplified and
re-transmitted on the orthogonal polarisation in thesecond cell
cluster, which would use a near omni-directional antenna for this
initial service for thebroadcast channels. This method could be
used, even if broadcast on demand had been implemented,provided all
the services demanded were also transmitted in the first cell
cluster. As all the servicesrequired in the network would also be
available in the first cell cluster, some break even point would
bereach, where the demands for spectrum from interactive services
in the initial cell cluster could renderthis arrangement
inefficient and the fibre connection desirable to connect the
second cluster.
3.2.3.2 Interactive services delivered over ATM
In a mature system the interactive ATM based traffic would be
connected to the Main Co-ordinationCentre through a fibre link.
However, in the initial stages of development connections between
basestations might be achieved, as with the broadcast channels,
through a radio link, using a transparentrepeater at the second
cell cluster. The problem is similar but not identical to the
broadcast on demandsituation. Again, the initial cluster is
connected to the MCC with a fibre/coax link and a radioconnection
is installed between the first and second cluster. One method of
relaying services would be toassume that all the users were in the
same cell cluster and relay all the services both clusters, with a
line
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
16
of sight(l-o-s) link operating in orthogonal polarisation, as
previously described using off-air pick-up.Some reduction in the
availability (99.95 %) would be experienced but this would not
detract muchfrom the benefits of a speedy roll-out of service
achieved through a radio connection. The accessprotocols would need
to accommodate the delays produced by longer path lengths of up to
about 10 km(if the base station separation is 8 km).
A more complex but spectrally efficient procedure would be to
split the operations between the first andsecond clusters. Then the
services for the second cell cluster are carried to the next cells
in a separatedfrequency band which allows the first cluster to have
access to all the allocated interactive channels. Theservices would
now be connected through a conventional LOS duplex link as shown in
Figure 6 (i.e. nolonger be relayed through off-air pick-up). Higher
gain antennas at the first cell cluster would beinstalled improving
the availability achieved. At the second cell cluster, services
would be re-radiatedthrough transparent repeaters on orthogonal
polarisation. Initially this approach would be successful,until the
demands of the second cell reach saturation in terms of the band
width required on the backbone. The four cell cluster arrangement
would require four time the traffic capacity(i.e. bandwidth) onthe
backbone as that in each cell. Furthermore, full potential of the
frequency reuse arrangements couldnot be realised with this method.
All these methods use transparent repeaters at the second
andsubsequent base stations, which retransmit the same modulation
schemes used on the air interface forthe in-cell operations.
A further improvement in the radio connection efficiency could
be achieved through using higher levelmodulation schemes on the
backbone. If the modulation scheme on the backbone achieved 4
bits/Hz interms of spectral efficiency (i.e. 64 QAM), then equal
bandwidth would be allocated to the backboneand four cell cluster.
However this arrangement would need both regenerative repeaters at
the secondbase station and a greater fade margin (12 dB more) to
achieve the same availability. This complexitywould need to be
compared with a fibre/coax connection in terms of cost, time scale
for installation andperformance. If regenerative repeaters are used
for the inter-cell connections, then one possibility wouldbe to use
out-of-band frequencies for the backbone network. This then becomes
a regulatory issue and itmight be considered unacceptable for a
service which had already been allocated at least 2 GHz ofspectrum
and possibly 3 GHz. Any encroachment on other parts of the spectrum
would be prohibiteduntil a very strong case for the request could
be made. However once regenerative repeaters are used forthe
connections between base stations, then using equipment developed
(at 38 GHz with a G.703interface) for mobile base station
interconnection could be attractive in cost terms
3.2.3.3 Capacity and link margins for microwave radio
backboneSome examples of the achieved performances for backbone
link connections are shown below in Table 5as a function of path
length, antenna gain, climatic zone and modulation scheme. The RF
parametersused are similar to those for the base station systems.
The bit rate is assumed to be a standard 34Mbit/s, RF power 17 dBm
and noise figure 7 dB. Table 5 shows the margin achieved with
theseparameters for link lengths from 4 to 10 km and compares these
values with those required to achieve99.99 % availability in
climatic zone E, H, and L. In zone E the 99.99 % service is just
achieved onpaths up to 10 km. This is in contrast with the
performance in zone L, where only the 4 km QPSKscheme meets the
design goal.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
17
Modulation Radio path Antenna Achieved Margin required
(dB)Scheme length (km) Gain (dB) Margin( dB) Zone E Zone H Zone
L
4 50 20 27 48QPSK 6 46 28 37 65
8 45 44 36 47 8110 42 43 55 94
16-QAM 8 38 36 47 8164-QAM 8 32 36 47 81
Table 5 Achieved and required margins for 99.99%microwave
backbone availability
The estimation of the traffic requirements (in terms of total
bit rates per cell) indicate that a fullyoperational cell might
generate up to 100 Mbit/s in the busy hour. In a four cell cluster
this wouldquadruple. If the backbone to cell interface is a
transparent repeater, then 400 MHz bandwidth would berequired to
carry backbone the traffic on a one bit/Hz basis. This seems
excessive as only 600 MHz isthen left for broadcast services. A
practical limit might be to assume that the backbone never
carriesmore than 155 Mbit/s, i.e. about the limit of the proposed
interactive spectrum allocation. In thisarrangement the backbone
links would always occupy the orthogonal polarisation to that used
in thecell. Some interference to users who fell within the l-o-s
beam would be inevitable due to the muchhigher power densities and
through scatter into the orthogonal polarisation. However the
percentage ofaffected users would be small, as both the beam width
of the l-o-s links is small (~100 m at 8 km range)and local terrain
and building would add further protection. When the number of users
in a particularcell cluster reaches about one third full capacity,
that seems an opportune time to switch to a fibreconnection for the
backbone.
The situation becomes more complex if the radio backbone extends
to multiple hops. It would bedifficult to use much more than the
above 155 Mbit/s on the backbone for interference reasons,
astransmissions on the backbone on orthogonal polarisation would
interfere with the in-cell services. Thusthe full capacity of all
the feeder links would be limited to 155 Mbit/s for all practical
purposes.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
18
3.3 Cellular LMDS access network
Base station
Radio repeater
Base station
Macro cell Micro cell
Figure 8 Basic LMDS network configuration
3.3.1 The concept of a LMDS cellRadio systems at 40 GHz have two
main features. Their main advantages are the very large
bandwidthwhich is currently available and the small physical size
of high gain antennas. The bandwidth availableof 2-3 GHz can
provide nearly three orders of magnitude more capacity when
compared with existingcellular mobile systems. This advantage is
however qualified by the very limited range (a fewkilometres) over
which such systems can provide an effective service, due to
propagation effects andblockage considerations. This second
restriction confines the system design to a cellular approach,
whencoverage of anything larger than a very small area is
considered.
The basic LMDS network configuration is shown in Figure 8. A
radio cell is in this context defined as acontinuous geographical
area throughout which the LMDS network will provide broadband
accessaccording to specified quality and performance criteria
covering a certain percentage of the potentialcustomers. To provide
a suitable framework for LMDS cell design, a geographical area must
becharacterised in terms of local topography, demography and
meteorology. Combining the geographicalmodel with principles for
radio network design, we may derive a radio cellular structure
where each cellprimarily is described by its size (in square
kilometre) and shape. Among the factors determining cellsize and
shape, we shall later discuss the relationship to
cellular subscriber coverage, single-cellnetwork capacity and
access network availability
In providing cellular broadband access, three major network
elements are necessary for flexible radiocell design. Those are the
cell base stations. micro-cell radio repeaters and customer
premisesequipment (CPE) as shown in Figure 8. The base stations
connects the cell to the internal LMDStransport network and provide
connections to the individual CPEs. Usually the base stations are
locatedin cell centre or at the cell edge depending each particular
scenario. Note that there might be severalbase stations covering a
single cell. CPEs provide the LMDS network termination at the
customerlocation. To provide connection to areas without LOS to the
base stations, micro-cell radio repeatersmight be necessary. The
maximum range of the micro-cells is typically 0.1 to 1.0 km whilest
the macro-cells illuminated by the base stations cover 1 to 10
km.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
19
0 1000 2000 3000 4000 5000
Range from transmitter (m)
0
10
20
30
40
50
60
70
80
90
100%
Co
ver
age
+0m:Signal OK +0m:Signal obstructed by trees +1m:Signal OK
+1m:Signal obstructed by trees +2m:Signal OK +2m:Signal obstructed
by trees +3m:Signal OK +3m:Signal obstructed by trees +4m:Signal OK
+4m:Signal obstructed by trees +5m:Signal OK +5m:Signal obstructed
by trees
Figure 9 Radio coverage in typical north-European suburban
environment as function ofdistance between BS and user.The
different curves show the antenna height above the rooftop
3.3.2 Cellular subscriber coverage
In general radio systems at 40 GHz relies on line-of-sight (LOS)
conditions to ensure proper operation.Hence, radio transmission at
this frequency will be blocked by buildings, vegetation and
terrainintercepting the radio wave path between the base station
and the CPEs. This problem. constitutes oneof the major challenges
of radio planning at 40 GHz. Providing 100 % subscriber coverage
within thecell area may be achievable from a technical point of
view but will in general not be a viable solutiondue to the
substantial economic cost involved. However, this does not preclude
100 % coverage incertain areas with favourable local
conditions.
3.3.2.1 Cellular design requirements and limitations
The practical limitations of subscriber coverage, is not
necessarily a major problem. In the emergingcompetitive market for
broadband access, different access technologies will partially
target the sameservices and geographic areas. Hence, the
general/total broadband access subscriber coverage would inthe long
term approach 100 %. From a regulatory point of view it should then
be suitable to adopt amoderate requirement on minimum subscriber
coverage for LMDS networks. This would allowoperators to lower the
cost of network deployment significantly enabling low-cost delivery
of services tothe subscribers. To achieve a fair market penetration
in a competitive scenario, the cellular subscribercoverage should
be as high as possible. It should also compare with other broadband
access networks interms of subscriber coverage. Consequently, a
minimum subscriber coverage above 50 % is necessaryand for the
purpose of LMDS cell design a subscriber coverage of 70 % is
suggested.
Individual CPE LOS conditions are largely dependent on the radio
path elevation angle and the height ofthe CPE antenna. For a given
BS antenna height we easily see that the elevation angle decreases
withincrease radio path length. Thus, for constant CPE heights, the
subscriber coverage are non-uniformthroughout the cell, decreasing
from the BS and outwards. This can be partly be compensated
byincreasing the CPE antenna heights as you move away from the BS.
BS antenna heights are estimated torange 20 to 40 m and may be
restricted by local environment restrictions. CPE antennas could be
easyand cheap to mount and heights up to 2-3 m above the roof line
and should be manageable forresidential customers. For business
customer CPE heights could be much higher at least 5 m or abovethe
top of the building. There is also a matter of visual user
acceptance of roof-top installations whichmay limit at the CPE
antenna heights in the case of residential customers.
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
20
3.3.2.2 A simulation study for a typical LMDS deployment caseA
typical example of a simulation of the effect of obstructions is
shown in Figure 9, where a threedimensional data base, containing
building, terrain and trees, has been used in conjunction with
raytracing methods. The percentage of buildings, which would
receive a LOS signal from a BS antenna at30 m height, is shown as a
function of height of the user antenna above the roof level (upper
set ofcurves). A second set of curves (lower set) shows the
percentage of paths which pass through a tree.The simulation was
produced for a fairly typical small town (Malvern, UK), which is
located in rollingterrain with a predominance of two and three
storey buildings. According to these characteristicsMalvern belongs
to the overlapping section between a suburban and urban case. This
is a typicaldeployment case where LMDS systems are expected to
perform well.
From Figure 9 we see that the general comments concerning
subscriber coverage stated previously areconfirmed. To met the
design requirement of 70 % subscriber coverage, we see the maximum
cell radiopath length is between 1.5 to 2.5 km. Assuming a CPE
antenna height of 3 m above the roof line, thecorresponding figure
is 2km. This establishes a upper bound on cell size fed by a single
BS. If a edge-fed approach is used, the upper bound is equal to the
maximum radio path length. However, if a centre-fed configuration
is assumed, the upper bound becomes twice the maximum radio path
length.Conversely, one may obtain much better subscriber coverage
with a centre-fed approach (comparedwith a edge-fed) when operating
with the same cell size and shape. For this particular case we
alsonotice that the obstruction by trees are very low. Therefore,
the major obstruction mechanism would bebuildings or other man-made
objects.
3.3.2.3 Cellular network implementation guidelines
A centre-fed approach generally leads to a 7-frequency re-use
pattern which in this context is a poorutilisation of the radio
spectrum. Hence, this approach is suitable for rural areas where
low trafficdemand allows use of larger cells and less radio
spectrum. Alternatively, if an isolated cell is planned,full
frequency re-use is possible and deployment in urban and suburban
areas is feasible as well. Inrural areas, the blockage are mainly
due to vegetation and terrain and improved coverage or
enlargedcells are usually achieved by illuminating an edge-fed cell
with several base stations. In case of verydifficult terrain
formations (e.g. highly rolling terrain, narrow valleys),
micro-cell radio repeaters maybe necessary to provide subscriber
coverage. In urban to city centre environments the blockage
ispredominately due to buildings and extensive use of radio
repeaters in combination with multiple basestations would be needed
to reach high subscriber coverage figures.
3.3.3 Cellular network capacity
3.3.3.1 Limitations on aggregate user bitrate
The maximum available LMDS network capacity is primarily
determined by 3 factors. They areavailable spectrum bandwidth,
spectrum re-use and spectral efficiency. Available radio spectrum
at 40GHz amounts to 2 GHz with current allocations ranging from
40.5-42.5 GHz. Utilising bothpolarisations, the unable spectrum
from network point of view is 4.0 GHz. There is a possibility
thatan additional 1 GHz might be added later extending the band up
to 43.5 GHz. In practice if a multiplecell deployment is envisaged,
some kind of spectrum re-use has to be done to avoid interference
betweencells. The degree of re-use is defined by the frequency or
spectrum re-use factor and for a LMDSnetwork values in the range 1
to 7 will cover most deployment scenarios. A re-use factor of
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
21
0 0.5 1 1.5 2 2.5 3 3.5 4 4.50
2
4
6
8
10
12
14
16
S pec trum re-us e
S pec tra l E ffic ienc y [b /s /H z ]
LMD
S c
ell
ca
pac
ity [G
b/s
]
= 1= 2= 4= 7
Figure 10 LMDS cell capacity for 2 GHz of radio spectrum using
both polarisations
0 0 .5 1 1 .5 2 2 .5 3 3 .5 4 4 .50
200
400
600
800
100 0
1200
1400
1600
1800
2000
Ac
tive
us
ers
per
ce
ll
S p ec t ra l E ffic ienc y [b /s /H z ]
A ve rage c apac it y pe r us e r
= 2 M b /s= 4 M b /s= 8 M b /s= 16 M b /s= 51 M b /s
Figure 11 Simultaneous number of active users per cellwith
frequency re-use factor of 4 and 2 GHz of spectrum
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
22
one is obtained if only an isolated cell is established and a
re-use factor of 7 corresponds to the case ofcentre-fed cells
similar to the approach employed by mobile cellular networks. The
spectral efficiencyrefers to the density with which information is
carried over the radio channel. It is measured in terms ofbit per
second per Hz (b/s/Hz) and is primarily defined by the coding and
modulation formats used. Theaverage spectral efficiency for a LMDS
network is estimated to be in the range 1.0 to 4.0 b/s/Hz. A
barchart showing the relationship between LMDS network capacity and
spectral efficiency for 4 practicalspectrum re-use factors is shown
in Figure 10.
From the Figure 10 we see that the capacity per cell reach as
high as 16 Gb/s for isolated cells. Hence,the available radio
spectrum is capable of supporting for instance two duplex 2.5 Gb/s
STM-16transport network connections for high density local traffic.
In a multi-cell deployment scenario, afrequency re-use factor of 2
to 4 is likely to be obtained for the urban and suburban case.
Assuming anaverage spectral efficiency of 1 to 2 b/s/Hz, the
corresponding cell capacity will be in the range 1 to 4Gb/s being
able of supporting for instance two duplex 622 Mb/s STM-4 transport
network connections.For the rural case the cell capacity will be
below 1 Gb/s in which one or more duplex 155 Mb/s STM-1connections
may be supported.
3.3.3.2 Limitations on number of simultaneous active usersThe
number of simultaneous active users handled by the LMDS network is
an important design figure.Referring to the proposed user class
configurations in Table 3, it is evident that no simple
procedureexists for derivation of the number simultaneous users
support by the LMDS network. In general, therewill be an exhaustive
number of possible ways to configure a cell in terms of users and
services. Toestablish an understanding of how the major factors
influence the number of simultaneous active users,an approach based
on a probable range of average user bitrates may be adopted. This
approach will notproduce the definite answer but merely indicate
the range over which the number of active users usuallyvary. A bar
chart showing the relationship between spectral efficiency and
number of active user foraverage user bitrates in the range 2 to 51
Mb/s, is shown in Figure 11. The assumptions are a frequencyre-use
factor of 4 and 2 GHz of available spectrum.
With the stated assumptions, the bar chart shows that the number
of simultaneous users may reach ashigh as 2000. In the range 1 to 2
b/s/Hz which seem to be the most likely option in practice, the
figureis below 1000. According to the user class service
configurations in Table 3, an average duplex userbitrate of 2 Mb/s
may be interpreted as a border between the business and residential
segment. Hence,we may state that a practical LMDS network may
support at least 1000 HHs or up to 1000 SME percell. The typical
average duplex user bitrates will in practice be much lower than 2
Mb/s for HHs andhigher than 2 Mb/s for SMEs. Consequently, as a
first order conclusion we may suggest that the LMDSnetwork will
typically support up to several 100 SMEs or several 1000 HH per
cell. In a mixedresidential and business scenario, the figures
would have to be reduced. Also, if for instance half thespectrum is
used for broadcasting of digital TV, the corresponding numbers
should be reduced with afactor of two. Considering a multi-licence
scenario, a similar reduction may be envisaged.
One should notice that the upper bounds of simultaneous active
users does not restrict the deployment ofLMDS networks in terms of
customer or HH densities. The mapping between the number of
activeusers and customer densities requires knowledge about radio
coverage, service penetration the useractivity rate (that is the
ratio between active and total number of customers)
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
23
Figure 12 Climatic zones in Europe according to
ITU-RP.837[11]
3.3.4 Access network availability
As stated in section 2.3.3 Definition of network availability,
the access network availability is in thiscontext based on the
yearly CNR statistics as determined by radio events. This
constitutes a frameworkfor predicting a best-case lower bound of
LMDS network availability. and other processes havingimpact on
network availability (like MAC protocols), will not be capable of
improving this lowerbound.
There are several radio propagation phenomena contributing to a
reduction of the nominal CNR. The byfar most important one in terms
of reduced signal level, is rainfall attenuation. Rainfall
attenuation ischaracterised by large and slow variations and the
fading process is generally not considered to befrequency selective
over the bandwidth of a LMDS up- or downlink. Annual predictions
models forrainfall attenuation based on [12] show that its scales
strongly with rain rate and radio path length. Thelatter shows that
there is a relationship between rain attenuation statistics and
maximum radio pathlength within LMDS cell and furthermore the
corresponding cell size.
The major meteorological parameter employed in calculating the
attenuation prediction is point rainrate.Ideally, accurate
knowledge about local rainrate statistics is required to derive
correct predictions.However, due to lack of such data in most
places, the ITU-R recommendation P.837 [11] is used as ageneral
framework for the predictions. ITU-R P.837 prescribes a division of
the earth in climatic zoneswith largely similar rainrate
statistics. Recommended climatic zones in Europe is depicted in
Figure 12.The map shows that the Mediterranean area along with
Turkey, parts of Eastern Europe and theBalkans occasionally
encounter very intense rainfall (zone K and L).
-
AC215/TEL/RD/DR/P/D2P1B/b1 Deliverable Report D2P1B
24
05
10152025303540
0,0 0,5 1,0 1,5 2,0 2,5 3,0 3,5 4,0 4,5 5,0
0
5
10
15
20
25
30
35
40
EFGHJKL
05
10152025303540
(a)