Analysis of the Impact of TD-LTE on

Mobile Broadband

XI CHEN

Degree project in
Communication Systems
Second cycle
Stockholm, Sweden 2013
ANALYSIS OF THE IMPACT
OF TD-LTE ON
MOBILE BROADBAND

CHEN Xi
xiche@kth.se

Supervisor and Examiner: Jan Markendahl

Master of Science Thesis

KTH Information and Communication Technology
Communication Systems
SE-164 40 KISTA
KTH School of Information and Communication Technology
TRITA-ICT-EX-2014:7
The Master of Science degree project on the subject of communication systems is open to
pubic scrutiny on September 18, 2013, 9:00-10:00 (Wednesday) at CoS conference room
Motala, 4th floor, Elevator C, Electrum (Isafjordsgatan 22), Kista.

© CHEN Xi, October 2013

Tryck: Universitetsservice US AB
Abstract
The mobile broadband services have developed rapidly over the years, becoming critical revenue
components of operator business. To accommodate users’ growing interests on mobile broadband
services and applications, global operators have tried to reorganize their former voice-centered
networks by focusing more on the reinforcement of mobile data network performances and
capacities. TD-LTE, as one of the emerging powers for the mobile broadband solutions, has gained
global attentions and momentums.
Different from previous work about the technical evaluations or market predictions, the thesis
aims to provide a techno-economic analysis to TD-LTE system, linking its technological characters
to the market opportunities and implementation strategies. In order to help TD-LTE operators
identify the profit potentials of the system, the services and applications that TD-LTE could enable
are discussed together with the analysis of the terminal developments, which are critical to end users’
acceptance of TD-LTE. The network deployment strategies are examined and the methods of
implementing a decent TD-LTE mobile data network are analyzed with cost efficiency
considerations.
The analysis find out that the availability of spectrum resource are of most importance for the
adoption of TD-LTE technology and the sustainable growth of TD-LTE business relies on the
differentiation strategy of services and applications. The feasibility study shows that TD-LTE could
enable varied network deployment scenarios, including integrated network solution with legacy
networks, and convergent network solution with LTE FDD. The cost analysis find out site
infrastructure sharing is beneficial for cost reduction during a national level rollout, especially when
the traffic volume increases. Reusing the coverage and capacity of legacy network is mostly effective
in the initial phase, and TD-LTE deployment pace should accord with the data prediction on the
live network. For compact and indoor scenarios, TD-LTE femto could be a cost effective
alternative for mobile broadband access, however the bottleneck of the solution is the limitation on
the number of concurrent connections for each femtocell.

I
Acknowledgment
The thesis project has been a challenging yet rewarding journey, which has introduced me to a new
field, that I am always interested in but lack the opportunity to engage in before. Upon its
completion, I would like to express my gratitude to those who have accompanied me along the way,
providing guidance and support.
First and foremost, I would like to thank my supervisor Assoc. Prof. Jan Markendahl for his
guidance, advices, encouragement and patience over the whole course of the project. His constant
and timely suggestions and feedbacks always guide me to the right direction for both the research
and the writing of the report. I would also like to express my appreciation to all the researchers and
fellow students, who help me in all sorts of manners and show their interests in my work, especially
Óscar García and Juan Pablo who take times to review my work and act as my opponents. I am also
very grateful for the excellent research, learning and communication platform that the Department
of Communication Systems and Wireless@KTH have provided me for my work.
Finally, I would like to thank all my families and friends home and abroad, especially my parents
and my cousins for all their support and trust.

CHEN Xi

Stockholm, October 2013

II
Abbreviations

3GPP The 3rd Generation Partnership Project

CAPEX Capital Expenditure
CMCC China Mobile Communications Corporation
CPE Customer-Premises Equipment
CSFB Circuit Switched Fallback
DwPTS Downlink Pilot Time Slot
E-UTRA Evolved Universal Terrestrial Radio Access
EPS Evolved Packet System
FDD Frequency-Division Duplexing
GP Guard Period
GPS Global Positioning System
GSM Global System for Mobile Communications
HSPA High Speed Packet Access
IMS IP Multimedia Subsystem
IMSI International Mobile Subscriber Identity
LAU Location Area Update
LSB Location Based Services
LSTI LTE/SAE Trial Initiative
LTE Long Term Evolution
M2M Machine to Machine
MBMS Multimedia Broadcast Multicast Service
MiFi My Wi-Fi
MIIT Ministry of Industry and Information Technology
MIMO Multiple-Input and Multiple-Output
MNO Mobile Network Operator
MWC Mobile World Congress
NDRC National Development and Reform Commission
NPV Net Present Value
OPEX Operating Expenditure
R&D Research and Development
TAU Tracking Area Update
TDD Time-Division Duplexing
UE User Equipment
UMTS Universal Mobile Telecommunications System
UpPTS Uplink Pilot Time Slot

III
UTC Coordinated Universal Time
VoIP Voice over IP
WiMAX Worldwide Interoperability for Microwave Access

IV
Table of Contents
Abstract I!
Acknowledgment II!
Abbreviations III!
Table of Contents V!
List of Figures VII!
List of Tables VIII!
1 Introduction 1!
1.1 Background 1!
1.2 Problem formulation 2!
1.3 Related work 4!
1.4 Contributions and scopes 6!
1.5 Thesis outline 7!
2 Methodology 9!
2.1 Methodology description 9!
2.2 Approach 10!
2.3 Data collections 11!
2.4 Cost modeling and assumption 11!
2.5 Case study 12!
3 TD-LTE System 13!
3.1 Introduction to TD-LTE system 13!
3.1.1 LTE system overview 13!
3.1.2 FDD and TDD 15!
3.1.3 Key concepts in TD-LTE system 16!
3.2 Evaluate TD-LTE system 18!
3.2.1 TD-LTE spectrum 18!
3.2.2 Coverage 19!
3.2.3 Asymmetric character 20!
3.2.4 Synchronization 20!
3.2.5 Channel reciprocity 20!
3.2.6 Leverage from other systems 21!
3.2.7 Summary 21!
4 TD-LTE Market Opportunities and Ecosystem 23!
4.1 Performance evaluation of MBB services powered by TD-LTE system 23!
4.1.1 TD-LTE network performance 23!
4.1.2 TD-LTE enabled MBB service performance 26!
4.2 TD-LTE enabled new services and applications 27!
4.2.1 Downlink biased services 27!
4.2.2 Uplink biased services 28!
4.2.3 Specific scenario and service packages 28!
4.2.4 TD-LTE market and operation strategies 29!
4.3 The development of chipset and terminal equipment 31!
4.3.1 Introduction to TD-LTE terminal products 32!

V
 4.3.2 Chipset and terminal development 32!
4.3.3 Trends and challenges of TD-LTE terminal products 34!
5 Network Deployment Strategies and Cost Analysis 35!
5.1 Coexistence and interoperability between TD-LTE and legacy networks 35!
5.2 Convergence of TD-LTE and LTE FDD 36!
5.3 Cost analysis 37!
5.3.1 Cost analysis for reusing legacy network infrastructure and capacity 38!
5.3.2 Cost analysis for HetNet deployment using TD-LTE and LTE FDD system 44!
6 Case Study 50!
6.1 China Mobile 50!
6.1.1 CMCC’s choice of TD-LTE 50!
6.1.2 CMCC’s strategies on TD-LTE services and applications 51!
6.1.3 CMCC TD-LTE terminal products 52!
6.1.4 CMCC Network deployment progress and strategies 53!
6.2 Denmark Hi3G 54!
7 Conclusion and Future Work 57!
7.1 Research question discussion 57!
7.2 Conclusions 59!
7.3 Discussion and future work 60!
References 61!
Appendix 63!

VI
List of Figures

Figure 2.1 Interrelation among topics and workflow 9!

Figure 3.1 System architecture diagram of LTE (EPC and E-UTRAN) 14!
Figure 3.2 TDD and FDD working principal 15!
Figure 3.3 TD-LTE frame structure (Frame structure Type 2)[20] 16!
Figure 5.1 Annual network production cost 42!
Figure 5.2 Production cost per GB 42!
Figure 5.3 Network implementation cost by areas 43!
Figure 5.4 Macro and Femto solution production cost from year 1-5 47!
!

VII
List of Tables
Table 3.1 TD-LTE uplink-downlink configuration[20] 17!
Table 3.2 Configuration of special subframes in OFDM symbols[21] 17!
Table 4.1 TD-LTE asymmetric peak throughput estimation 26!
Table 4.2 TD-LTE enabled services and applications 30!
Table 5.1 Cell range and coverage of HSPA and TD-LTE 39!
Table 5.2 Cost structure of HSPA/TD-LTE sites 40!
Table 5.3 Reuse scale of Case B-E 41!
Table 5.4 Macro site number requirement 46!
Table 5.5 Femtocell number requirement 46!
Table 5.6 Cost structure of Macro LTE sites 46!
Table 5.7 Network cost comparison 47!
Table 5.8 Production cost for different sectors 48!
Table A.1 Cost analysis of CASE A 63!
Table A.2 Cost analysis of CASE B 63!
Table A.3 Cost analysis of CASE C 64!
Table A.4 Cost analysis of CASE D 64!
Table A.5 Cost analysis of CASE E 65!

VIII
 1 Introduction
The mobile operator business has experienced dramatic changes over the years with the flourishing mobile
Internet development. Starting from 3G eras, the market have shown growing interests on mobile broadband
for the changes and benefits it brings to people’s way of communication. As one of the promising candidates for
future mobile broadband solution, TD-LTE has high potentials to help mobile operators keep competitive as
mobile broadband dominates new ways of communication. However, the adoption of a new technology usually
involves careful evaluation and comparison, hence the thesis aim to help operators to analysis the system though
both market opportunities and network deployment strategies.

1.1 Background
During the past years, the development of mobile broadband is unprecedented, occupying an
increasingly important role in mobile operators’ business. However, as people enjoy the
revolutionized way of communication and manifold services and applications it brings, the
data traffic through mobile network has also grown over the years and pushed the boundary
of current mobile networks’ capacity. The gaps between user demands and network capacity
would even deepen, as more smart devices with aggressive data functionalities are introduced
to the market. According to the estimation from Cisco in [1], the global mobile traffic will
experience a 13-fold growth in 2017 over five years term. For developed markets like Europe
and North America and emerging markets in Asia and Pacific regions, the data growth is
especially astonishing.
Beside the traffic tsunami, the mobile subscribers also expect significant improvement over
the network performance, since stable and enhanced network infrastructure lays the
foundation for the growth of innovative services and applications, which have gained much
tractions among end users. A nearly fixed network experience is expected on the mobile
devices. Hence, operators need to maintain their customers by providing a satisfactory
network experience through, for instance, higher data rate, lower latency and seamless
connections.
It is a dilemma for operators to enhance the network capacity and performance, while at
the same time reduce the network construction and operation costs. Besides, the traditional
revenue source of voice and SMS services have been shrinking and constituted an increasingly
smaller proportion of total revenues. Even for the burgeoning mobile data services, operators
face strong competition from Internet companies, as the increasing data traffics do not always
brings correspondent revenues, and large amounts of profits flow directly to application
providers and Internet companies instead of mobile operators. To rectify the adverse
situation, mobile operators became to realize that more focuses need to put upon mobile
broadband and corresponding services and applications, as they would presumably shape the
future profit margin of mobile operators. Hence, despite of the cost pressure brought by
upgrading mobile broadband services, investing on the data network is still what many global
operators do to tackle those challenges. During the deployment progress, a few general
strategies need to be defined, for instance the method of mitigating the network upgrading
and operational cost, the way to generate revenues through innovative service and
applications and how to provide a satisfactory mobile broadband experience though network
implementation and radio technologies.
TD-LTE along with other technologies like LTE FDD and WiMAX, are therefore
introduced to the telecom market, hoping to help operators stay competitiveness in the value
chain. Different from other technologies, TD-LTE has gained global momentums based on
several of its key advantages, for instance, the utilization of unpaired and affordable spectrum
resource, flexible UL/DL data rate, seamless upgrade from existing systems like WiMAX or
TD-SCDMA, leverage of LTE FDD ecosystem. According to the latest industry report from
GTI, there are 16 commercial TD-LTE networks launched by far and the commercial TD-
LTE contracts have reached the number of 47 [2]. Estimation from Heavy Reading suggests
that TD-LTE system will accumulate 158 million subscribers globally, penetrating 37% of
global LTE market [3]. The TD-LTE roll out will also cover several of the most densely
populated areas like China, Japan and India. Global deployments have attracted intense R&D
investment and support from major vendors ranging from infrastructure, chipset, terminal
and test equipment. As a result, scale of mature and competitive products could pave the way
for the future growth of TD-LTE deployment, and relieve the bottlenecks, like lack of
popular smart handsets in the market, where TD-SCDMA have been suffered from.
However, even though TD-LTE has drawn much attention within the industry, and
became a promising candidate for the mobile broadband solution. The development of TD-
LTE is still at its infancy and its prospect is yet to be proven by the markets over the coming
years. But as for an operator, the most important question is whether and how TD-LTE
system could bring profits for the network operations. It is therefore worthwhile to discuss
the potentials and drawbacks of the system, and define general strategies in terms of markets
and network deployments.

1.2 Problem formulation

From operators’ point of view, both the market and network deployment strategies and
decisions should contribute for the long-term competiveness and return of investment. To
achieve long-term competiveness amid the transformation of the telecom industry, operators
could adjust its operation by investing in the next generation network infrastructure,
enhancing the network performance and developing innovative service and applications, all of
which lays the foundation for future growth in the mobile broadband service. While in order
to achieve satisfactory return of investment at the same time, operators need also to reduce
the network deployment cost and identify more profitable service models.
Among all alternatives of mobile broadband access technologies, the operators should
consider cost efficiency, application scenarios, market prospects and network performance to
motivate the selection of mobile broadband technologies among the most prevailing
candidates like FDD LTE, TD-LTE and WiMAX. For future 4G operators, it is important to
identify the pros and cons of the candidate systems and examine whether the character of
such system could be in line with operators’ position and strategies. While for operators who
have already selected TD-LTE system, focus would be put more on how TD-LTE could help
them keep competiveness and growth.
Selection of proper mobile access technologies is important, so does the method of
cultivating the mobile broadband markets and powering the profit growth by such system. In
the same regional markets, TD-LTE operators would usually face direct competition from

2
operators using other mobile access technologies. Therefore, to provide end users with
competitive network experiences would help TD-LTE operators to gain an edge in the
market competition. Based on the network performance evaluation, TD-LTE operators could
identify whether its users could get comparable performance when using standard mobile
broadband service. Beside of traditional mobile broadband revenues, TD-LTE operators
could also explore the add-on values of the system, by developing new types of service and
application, targeting specific groups and segments. Therefore, the revenues from both
traditional mobile broadband service and new types of service and application would help
TD-LTE operators to keep decent growth in profits.
Another issue that TD-LTE operators would concern is the availability of terminal
products. End users’ decision to subscribe to a certain network depends not only on the tariff
and network performance, but also on the preferred terminal products. Therefore, the
popularity of TD-LTE system products is one of the most important prerequisite for the take
off of TD-LTE system. The variety of terminal products could also limit the service and
applications that TD-LTE operator could offer.
Looking back to the infrastructure development of TD-LTE system, the global
deployment is still at its initial phase, and legacy networks like 2G and 3G systems would still
exist in the live network environment. To provide a seamless communication experience to
the end users and explore the benefits of each individual system, the coexistence and
integration strategies of TD-LTE system and legacy networks could be investigated. There is
also a trend in the global operators to provide convergent networks with TD-LTE and LTE
FDD system, raising the question of how to position and intergrade those two versions of
LTE system.
For a newly built system like TD-LTE, it is important for operators to identify the issues
like deployment strategies and construction cost which could however vary in different
phases and implementation scenarios. The issues are highly related to each other, as the cost
of network construction is usually affected or even defined by the deployment strategies
applied. For instance, the deployment strategies of positioning TD-LTE system as the
primary network with full coverage and positioning TD-LTE system as complement network
by reusing the capacity of other networks may result in large difference in deployment costs.
Site infrastructure reuse is another mechanism, which could bring possible reduction on
deployment costs but it could also bring inflexibility of network optimization and expansion.
TD-LTE heterogeneous network solution could have the similar influence to network cost
structure, even though the HetNet solution could bring operators large amounts of redundant
network capacity for indoor environment, its cost efficiency based on actual condition is still
the primary assessment for the deployment strategy. Therefore, the balance between
deployment strategies and construction cost need to be found, based on the circumstances
like user requirement, network deployment and management difficulties, possibility of
network expansion, network investment and profitability. Operators could define the network
deployment roadmap by estimating the cost and profits.
As one of the candidates toward the next generation mobile network, TD-LTE could
provide MNOs with alternative and flexibility in services, applications and network
deployment strategies. It also has the potential to help MNOs achieve growth in profitability,
by exploring new revenues and reducing the network cost. Therefore how to make full use of
TD-LTE system in both technical and business perspective is what the thesis tries to explore.

3
More specifically, it is expected to investigate TD-LTE system through problems under those
two categories:
Market opportunity and drivers
• Why choose TD-LTE, comparing to other mobile access technologies?
• Is the user experience of TD-LTE different from other systems in terms of standard
services or applications? What particular services or application can be facilitated by
TD-LTE system?
• What is the market prospect for TD-LTE enabled chipsets or terminals?
Network deployment and cost analysis
• How to integrate TD-LTE with self-owned FDD LTE, 3G or 2G systems?
• What deployment strategies and cost control mechanisms could be applied for TD-LTE
implementation?

1.3 Related work

Previous reports can be found in several areas related to either the TD-LTE or MBB system
in general, and they can be identified to develop several main topics, including,
TD-LTE system evaluation
Papers [4], [5] are trying to identify the network performance and characteristics of TD-LTE
system in different situation and possible solutions to improve the system performance have
also been proposed. For instance, the uplink to downlink sub frame ratio setting is one of the
key issues of TD-LTE system operation. Different uplink to downlink sub frame ratio can be
used to facilitate different application or user behavior. This is one of the flexibility that TD-
LTE system could offer to accommodate the future challenges of exponential data growth
and make network smarter by proper network planning. Therefore, several simulations are
made to compare LTE FDD system and TD-LTE system with various uplink to downlink
sub frame ratio settings, the results could assist in exploring the benefit of TD-LTE system in
network planning and deployment process. Besides, an adoptive adjusting mechanism is also
discussed in [6] to improve the system performance.
Other potentials of TD-LTE system have also been discussed, for instance the possible
solution to make TD-LTE system greener is discussed in [7], since major operators will
deploy large quantities of base station across the service area for the purpose of coverage
competitions or social responsibilities, bringing enormous energy consumption. TD-LTE
system could exploit its own system character to further reduce the energy consumption to
minimize the cost of operation and be more environments friendly. Another common
opinion about TD-LTE system is that its coverage range could be smaller than LTE FDD
system. This could influence the decision of network investment and increase network
deployment difficulties to improve the overall performance of the systems.
MBB system business strategy and cost analysis
As the current user behaviors have gradually been transferred from the traditional voice based
service mode to data based service mode, operators need heavy investment to improve the
network capability to meet the user's requirement, while at the same time, the data service are

4
no longer as profitable as traditional voice or SMS service. To make it even worse for
operators, the business landscape has changed during the latest development periods of the
mobile broadband, as most of the profit have been obtained by some of the new players
joining the competition, like the on line companies or application providers. To tackle those
challenges and fill the revenue gaps they may bring, the business strategy of implementing
MBB system have been investigated in [8] and [9].
The cost analysis models have been proposed in [8] and [10], to simulate the cost structure
of different network deployment options and technologies. The calculation could help
operators to estimate how different network deployment solution and radio access technology
options could bridge revenue gaps while at the same time bring further growth. TD-LTE
system as one option for MBB system could also be analyzed through cost models, when new
network deployment or service strategies are identified.
Industry reports regarding to TD-LTE system development
The industry reports [3] and [11] tried to describe the current development and future
prospect for TD-LTE system. The current market overview for both TD-LTE and LTE
FDD system is described; LTE system construction has already been spread out from the
leading markets to the whole world, even though most operators begin their network
development with LTE FDD system, TD-LTE system is also favored by some of the main
actors and markets. As more operators like Softbank from Japan, Bharti Airtel from India and
China Mobile from mainland China and Hong Kong, entering the market with TD-LTE
system, it would became a more mature and recognized solution for the mobile broadband.
Markets report [2] has also kept a close watch on actions that each operator takes to prepare
for the launch of service, including the test results for trial networks, the procurement
progress for the network equipment and handsets, network roll out plan, spectrum acquisition
strategies, timing of commercialization and how the operators could position in domestic
markets and influence the development of TD-LTE ecosystems.
Survey results for operators have also been demonstrated in the industry report [3],
showing the willingness of implementation of LTE system for global operators, and how
operators plan to implement the networks. Topics includes the frequency band that operators
favored to implement the service, the preferences between TD-LTE system and LTE FDD
system or mixed network, the network roll out strategies like how to position TD-LTE
system in the entire operational network. The network implementation plan of TD-LTE
system is also unveiled for the operators who have decided or even started with the
implementation process. The network construction and subscription forecast have also been
predicted.
Beside network operators, other players within the TD-LTE ecosystems, like chipset
vendors, handset manufactures and equipment providers are also included for the monitoring
in [11]. How could those actors support with decent network equipment as well as terminals
is also crucial for the market, as previous markets experience has shown that proper handset
support is key to the market share among different technologies. Even though, LTE unlike its
predecessor 3G systems, have evolved to minimize the difference between FDD and TDD
mode, the manufactures still need to continuously contribute for the wholesome growth of
TD-LTE system. However, challenges also bring benefit, the growth of TD-LTE market
could help the vendors to be more successful among competitors. Therefore, key network
equipment vendors have also demonstrated in their reports, for instance [12] and [13], how

5
their TD-LTE solution could help operators improve the network performance and increase
revenue while at the same time minimize the investment.

1.4 Contributions and scopes

Based on previous papers and reports, a few more work could be done to further analyze TD-
LTE system by investigating several of the key issues that MNOs are usually interested in,
instead of focusing on the prediction for the market volume, the pros and cons of the system
will be examined, often related to technological characteristics, network performance
comparison, profit growth prospects, development of ecosystems, and the cost efficiency
evaluation. Generally speaking, the potential of TD-LTE system, about how it will help the
operators turn market challenges into revenue, will be explored.
According to identified previous works, even though network characteristics or evolution
path of TD-LTE system are mentioned, for instance in [5]; or a basic comparison between
TD-LTE and LTE FDD system are carried out in [14] in terms of protocol difference,
frequency band, and data rate, there is still a gap for demonstrating a comprehensive
comparison between TD-LTE system characters and other systems in a techno-economic
point of view, meaning that the technological characters of TD-LTE systems from previous
works will be collected and elaborated by putting them under the business and market
contexts. The efforts aim to provide mobile operators a clear picture about why and how
those technical characters could influence the development of the system and their mobile
broadband business. The technical characters, which have potentials to influence operator
business, could be about spectrum, coverage, asymmetry nature, synchronization mode,
channel reciprocity, and architecture similarity with other systems. By linking the technical
characters with business and market considerations, more intuitive topics which operators
concerns most could be derived, including implementation possibilities and difficulties,
investment cost, resource utilization rate, user experiences, new applications and markets, and
ecosystem development.
The thesis will also extend the discussion from previous network performance evaluation
results, like in [4], and focus more on how those results could affect the basic or similar
service that TD-LTE system could offer, comparing to other systems like LTE FDD system
in particular, and how they could affect the market positioning and differentiation of
operators. Beside the comparison between similar services offered, the study will extend the
discussion in [15] for the new possibilities and applications that TD-LTE system could offer,
by exploring how to implement and promote those possibilities and examine other unique
services and scenarios that TD-LTE system could enable, but have not been mentioned in
previous works. Based on the discussion, the market and operation strategies could be
revealed and the question about how to facilitate the market acceptance of TD-LTE services
and applications could be discussed.
The thesis will fill an identified gap when it comes to the interoperability and practical
integration strategies between TD-LTE and other systems, which are rarely shown in
previous works. Since there have already been operators implementing TD-LTE systems, the
general solution guidance could be provided to position TD-LTE system in the entire
network environment. The discussion could be divided into two parts, where TD-LTE will be
analyzed together with legacy networks like GSM and UMTS, or FDD version of LTE
system. The discussion will focus both on the opportunities and drawbacks of TD-LTE

6
convergent networks, compared to standalone solutions. The opportunities could include
voice solutions like CSFB, ubiquitous communication experiences, cost saving mechanism
through infrastructure sharing, flexibility on tariff setting and service differentiation,
additional capacity, and flexibility on network hierarchical planning. However, convergent
networks also have obvious drawbacks like the inflexibility of network tuning and possible
performance degradation. The influences to operators’ businesses based on the opportunities
and drawbacks will be discussed.
The cost of network implementation is one of the primary concerns of operators, which is
especially true during the financial turmoil, therefore, the study will do similar cost analysis as
presented in [8] but will focus more on the scenario of reusing network infrastructure and
capacity when constructing TD-LTE system on top of HSPA network, and the scenario of
constructing both TD-LTE and LTE FDD system for HetNet. Based on the designed
scenarios and following cost calculation, more general network deployment would be derived,
helping operators to tailor its own network deployment strategies based on the experiences
and discussion from the paper.
In general, the market challenges of MNO, especially the TD-LTE operators, are coming
from several different aspects, including the end users' drive on higher data rate, newly
flourishing applications' request for higher traffic volume, competition from other MNO or
even landline providers, competition among different mobile access technologies, pressure
from the investment of network infrastructure and network management, and strikes on the
traditional voice and low rate data connection based profit model. Thus, the thesis will mainly
stand on the MNO's point of view. It will identify the market challenges more specifically and
try to answer how to address the market needs by implementing TD-LTE system. The
deployment options will be analyzed to explore the possibility of further cost efficiency.
Besides, the TD-LTE system enabled service type and profit model will be discussed. Topics,
which are general to all MBB system will however not be discussed in details. The study is
specifically focused on the issues related to TD-LTE system, and expected to gain insight
about market trends and solutions for MNOs and serve as a reference for decision making of
network investment and business models applied.

1.5 Thesis outline

The scope of the thesis are constituted by two parts, including market opportunities and
network deployment strategies, aiming to help operators find out both opportunities and
challenges from the initial phase, implementation phase and go-to-market phase. The thesis
starts to introduce the topics based on market backgrounds and define the scope of the thesis
according to previous works in Chapter 1.
In Chapter 2, the logical flow of the thesis is demonstrated, explaining how the research
questions could be answered and linked to each other. Concerning a new technology like TD-
LTE, the general method of analysis has been defined and explained. Latter in the chapter,
the market analysis approach, information sources, cost analysis assumptions are also
presented for validation purpose.
Chapter 3 tries to provide a general view about TD-LTE, explaining the basic and key
concepts of the system. An evaluation of TD-LTE system based on a techno-economic view
is presented and tries to explain why TD-LTE system could be beneficial for some operators.

7
 The market opportunities are shown in Chapter 4, focusing on the services and
applications that TD-LTE could enable. In order to evaluate those services and applications,
the network performance of TD-LTE is discussed as an input. However, the market
development of TD-LTE systems does not only depend on the services and applications,
therefore the terminal and chipset development observations are also presented.

In Chapter 5, the thesis starts discussions regarding to the network deployment issues. The
implementation strategies are then being demonstrated based on technological or market
discussion, and most importantly cost analysis.

Chapter 6 tries to continue the discussion from previous chapters, by applying them to real
operators, including China Mobile and Hi3G Denmark.

Chapter 7 concludes the work by reviewing the key concepts related to research questions,
and possible future work is also listed.

8
 2 Methodology
In order to answer the research questions, general methodologies are described in this section, including the
workflow and interrelations among different topics related to RQs. Business model by Chesbrough and
Rosenbloom, and three market research approaches are also mentioned for their contributions to the analysis in
the thesis. Some important methods when dealing with data processing, cost analysis and case study have also
been described.

2.1 Methodology description

The purpose of the thesis is to help operators analyze and utilize TD-LTE, and discuss what
does TD-LTE mean to the mobile broadband industry. To analysis and identify TD-LTE,
there are several issues could be included, for instance technical strength and weakness,
network performance, potential for new services, market prospects and strategies. To
expound the methods of utilizing the system, it shall cover issues related to network
integrations, system positioning, deployment strategies and cost control mechanism.
Therefore, the thesis is constructed in the following orders to help achieve the objectives
of the thesis and answer the research questions,
1. Identify the technical characters of TD-LTE, and explore how they could influence
the cost structure, network deployment process, industry development and market
prospects of the system.
2. Expand the discussions regarding market potentials and industry prospects.
3. Through simulated models, explore possible network implementation strategies.
4. Discuss how the identified market and implementation strategies could be applied to
specific global operators.
More specifically, the interrelation among the topics within the scope of the thesis could
be illustrated in Figure 2.1.

Figure 2.1 Interrelation among topics and workflow

As TD-LTE is a technology at its infancy, it is therefore suitable to apply the methods

brought by Chesbrough and Rosenbloom in [16], which intend to define a business model for
the analysis of an early stage technology. The business model definition gives an insight about
how a new technology could generate value for a company and what is the main interest of

9
companies when it comes to the adoption and application of a new technology. However, not
all the elements in the business model are discussed with the same depth, depending on the
scope of research questions. Therefore, how those elements are used within the scope of the
research questions could be listed more specifically,
• The value proposition of TD-LTE help explain the add-on and exclusive value that the
system could offer to end users though innovative and specific services and
applications.
• As many of the service and applications have clear target, the market segments could
therefore be identified. However, when TD-LTE is offer as a standard mobile
broadband solution, the target group have no clear distinction to other technology,
hence the specific segments requiring customized service have been put on with more
focuses.
• TD-LTE operators’ existing assets within the firm value chain, which could help deliver
the service and application, is discussed, for instance legacy networks.
• The cost structure and profit potential are estimated in different manners. A selection of
service scenarios provides insights about the cost structure of TD-LTE when applying
different deployment strategies, while profit potential are discussed generally without
explicit calculation.
• The position of TD-LTE operators in the value network is discussed in three main
aspects, as TD-LTE operators depends on regular subscribers and specific business
segment as their main source of value generation, and are influenced by the offer and
development of suppliers of infrastructure and terminals. It also faces the erosion of
values caused by complements like mobile Internet companies.
• The competitive strategy of TD-LTE operators is discussed, as they face the competition
from operators with other mobile broadband technology in the same markets.
As mentioned by Markendahl in [17], the business model lacks elements about technical
design and functionality, which is important as a background for the discussion of value
proposition and market segments, therefore the technical characters of TD-LTE system is
also identified.

2.2 Approach
One of the main objectives of the thesis is to investigate the relations between technology and
economics, more specifically how the technology character of TD-LTE system would
influence the market strategies and positioning of TD-LTE operators, therefore a few market
analysis approach are implemented to help answer the research questions.

Comparative analysis approach

The comparative analysis approach is to identify the nature of an entity though analysis of the
similarity and difference of objects within the same category. This method helps to identify
the most compelling and unique characters through connection instead of isolation. The
approach are used to answer the RQ of “Why choose TD-LTE, comparing to other mobile
access technology?” and “Is the user experience of TD-LTE different from other systems in
terms of standard services or applications?”, both of which often use LTE FDD system as a
benchmark, as the most popular global mobile broadband access solution.

10
 The same approach is also used to define the target of case study, as the differences in size,
policy and market of China Mobile and Hi3G could result in different network deployment
and business strategies, even the same technology is being used.

Qualitative and quantitative analysis

The combination of qualitative and quantitative analysis helps obtain both the whole picture
and interesting details. To define the network deployment strategies of TD-LTE system, the
qualitative analysis are used to describe the relation and integration method between TD-LTE
and other networks, like 2G, 3G and LTE FDD. Therefore, it is possible to provide a rich
description about each integration possibility, for instance voice solution, coverage and
demands complement, mobility management, site reuse possibilities and heterogeneous
network.
Quantitative analysis will help solve the problems about how user demands, population
density, network positioning and deployment strategies like site reuse, femto solution could
shape the production costs of TD-LTE service, which provide direct basis for the selection of
network deployment strategies under different scenarios and tariff setting.

Deductive reasoning
The process to deductive reasoning is to obtain new conclusions based on the collection and
reasoning of existing information. The method is used to solve the questions about TD-
LTE’s potentials to provide both standard and particular services and applications. TD-LTE’s
technical characters and actual network performances are presented as the premises, and
based on that, the possibilities and potentials of TD-LTE enabled service and applications are
deducted as the new conclusion.

2.3 Data collections

The data collection of the thesis is mainly based on secondary data sources, often providing
information with a large scope and high quality, which could be difficult to be obtained with
limited resources. Here list the main sources to support the discussion and analysis in this
thesis,
• Publications and Books
• Industry reports and presentations
• Press releases, annual reports and media interviews
• Statistics from authorities
To validate the secondary sources, especially those might contains subjective opinions like
media interviews, double confirmation is made through different sources. The original of the
sources could be traced back though reference list and footnotes.

2.4 Cost modeling and assumption

The purpose of cost modeling is trying to discover network deployment strategies and cost
control mechanism based on the cost calculation of typical scenarios, and discuss the issues
that TD-LTE operator might face during the network implementation process. Based on
given scenarios, different network deployment alternatives are presented and compared cost
wise for varied customers, areas or phases. The goal is to help operators to adopt suitable
network deployment strategies and cost control mechanism according to own situations,
11
therefore the discussion would also include possible issues of TD-LTE implementation in the
given scenarios as precautions for future deployment.
Two scenarios are selected for the cost modeling, including TD-LTE network
implementation based on existing HSPA network, and TD-LTE femtocell implementation. In
the first scenario, a simulated country is under analysis based on the form of a medium size
European country, with some data, e.g. area, population density, mobile broadband business
development or user demands, originated from countries like Sweden or Demark, where there
is one major national hub like Stockholm or Copenhagen. Five different deployment
alternatives are selected based on the depth of reuse of existing network and the comparison
metrics include CAPEX, OPEX, annual network production cost, and production cost per
GB. The second case will compare TD-LTE femto solution with LTE FDD macro solution
in a simulated science park based on the data of Kista Science City. The comparison metrics
include CAPEX, OPEX, NPV and production cost per GB and production cost per month
per user, which is a comparable counter to ARPU.
The discussion is mainly conducted for the network side cost, and the cost of spectrum is
not considered in the analysis. But in the first case, the total uplink and downlink bandwidth
of both HSPA and TD-LTE network is set to be identical as 20MHz. The operators may
need to take the additional cost of spectrum resources into consideration in real situation.
Besides, for simplification reasons, the network expansion option neglects the possibility of
adding an additional carrier for HSPA network.

2.5 Case study

The motivation for the case study is to apply the general conclusion of research questions
into real operator business and verify whether those findings are still legitimate in different
situations. Beside of verifying the common conclusions, the cases study will also aim to find
out specific phenomenon and solutions, which are unique for each case based on specific
market situations. The case study will select two typical operators for analysis, China Mobile
and Hi3G Denmark. China Mobile is the representative for incumbents, while Hi3G is the
representative of newcomers to the business. How the respective operator positioning and
market situation would influence the operator’s TD-LTE implementations and strategies
could be an interesting topic to discuss. To resonate with research questions, the topics in
each case include analysis of each operator’s reasons for the choice of TD-LTE, their
strategies on TD-LTE services and applications, actions on the development of TD-LTE
terminal products, and progresses for network implementation as well as discussions about
the network deployment strategies they could follow.

12
 3 TD-LTE System
For operators aiming to provide mobile broadband services, could TD-LTE be a suitable solution? To address
this issue, the chapter tries to introduce and evaluate TD-LTE system by compare it with other alternatives
like LTE FDD system. When it comes to the evaluation of TD-LTE, the discussion start with the
technological characters, but also try to elaborate the findings by relating them to issues that matters most to
operator business, hence those technical details are linked to four major topics, i.e. cost of implementation,
network deployment issues, market and industry development, network performance and applications, all of
which will also be further discussed in later chapters.

3.1 Introduction to TD-LTE system

3.1.1 LTE system overview
The structures of traditional mobile cellular systems are mainly designed for the voice
services, and have served the world users for decades. Starting from 3G eras, the data based
network structures have been integrated as one of the core components for the whole system,
which also cultivated the mobile Internet business and services mode. With the flourish of
mobile broadband business, mobile networks have soon been overwhelmed by the growing
data demands and user expectations. Hence, a dedicated data based mobile network system
need to be designed to meet the gaps between network capability and user expectations. LTE
is such a system, which is designed for the IP-based data service, and revolutionary from its
predecessors. The network design of LTE follows a few requirements from several aspects
[18]:
• Data rate: The data rate requirement refers to peak data rate and cell edge bitrate.
System peak data rates are related to system bandwidth and antenna mode. Therefore,
it is required to achieve 100Mbps for downlink with 20MHz bandwidth and 2 receive
antenna at UE. For uplink it is required to achieve 50Mbps for uplink with 20MHz
bandwidth and 1 transmit antenna at UE. This requirement is important for the
increasing numbers of service and applications with high data consumption rate, like
high definition video service and cloud services. Those services and applications are
on the rise in terms of both numbers and data consumptions, making it a prerequisite
to include the peak data requirement for a future-proof mobile system. While at the
cells edges, users could hardly achieved the peak data rate, therefore it is also required
for LTE to achieve a improved cell edge data rate based on the current deployment,
comparing to its predecessors, guaranteeing the users with a minimum experiences
even in poor radio environments.
• Capacity: Capacity is important for system dimensioning, especially for a loaded
network environment. Sufficient amount of capacity ensures that the systems could
support enough spontaneous connections with reasonable data rate, and it is often
measured with user throughput and spectral efficiency. Therefore, it is required to
significantly increase user throughput and spectral efficiency up to 2 to 4 times from
Release 6, for both uplink and downlink with basic antenna configuration, i.e. 2T2R
for downlink and 1T2R for uplink.
13
 • Latency: Latency is another measurement for system performance, and it is
important for real time applications, for example instant voice and messages, video
conferencing, and emergency alarm reporting system. As one of the primary
performance indicator, it requires the latency in user plan less than 10ms, and control
plan latency (camp state to active state) less than 100ms, which is also much more
strict than previous systems.
• Spectrum: Spectrum resources are crucial for mobile network deployment, yet
limited in most parts of the world. Therefore scalable spectrum is introduced to LTE
system, binging more flexibility for mobile operators to deploy a system tailored with
its specific network status and customer expectations. LTE systems could support
1.25 MHz, 1.6 MHz, 2.5 MHz, 5 MHz, 10 MHz, 15 MHz, 20 MHz system bandwidth,
and both paired and unpaired spectrum.
Besides the primary system enhancement, there are also other requirements included for
LTE system design. For mobility requirement, LTE should optimize for low mobile speed
and also support high speed up from 350km/h to 500km/h for specific scenarios like high-
speed train services. LTE should also support internetworking with 3G and other non-3GPP
systems, which helps mobile operators provide a seamless communication experiences based
on its legacy network and LTE systems. Further requirements include reduced production
costs on the network equipment and reduced complexity on the terminal products, enhanced
MBMS in E-UTRA systems, enhanced IMS and core network, and LTE should also support
various type of services, like VoIP voice solution.
Based on the requirements, LTE have developed a flat architecture, which minimized the
system complexity. RNC from 3G systems have been removed from LTE, and its radio
functionalities have been moved to base stations. This structure could reduce the system
latency and production cost as well.

Figure 3.1 System architecture diagram of LTE (EPC and E-UTRAN)

In Figure 3.1, LTE contains two major components, EPC and E-UTRAN. E-URAN is the
access point of UE, dealing with most radio related operations, likes radio ciphering,

14
integration protection, load balancing, interference coordination, and synchronization. EPC
communicate with E-UTRAN with S-GW and MME. S-GW will transfer the user data
packets from and to E-UTRAN, while MME handles the signaling in control plane. The
design of LTE architecture is IP- based, therefore circuit switch architecture no longer exists
in EPC. eNodeB connects MME and S-GW through S1 interfaces, and eNodeB could also
connect each other through X2 interfaces. To reduce the loads of EPC, eNodeBs could
process handover between adjacent nodes through X2 interfaces.
In order to be qualified as a 4G standard defined by ITU, LTE have also been developed
further into LTE-Advanced. Through some of the key technologies like carrier aggregation,
high order MIMO and relay nodes, LTE-Advanced could achieve much higher peak data rate
and cell edge bitrate, and enhance system capacities though improved spectral efficiency, and
increased spontaneous active connections[19].

3.1.2 FDD and TDD

LTE has defined two duplex mode FDD and TDD. FDD utilize paired spectrum resources,
where spectrum resources are divided into two identical bands. Uplink and downlink signaling
and data are allocated to delicate band respectively and the guard band lays in-between the
uplink and downlink band to prevent interferences. Due to the uplink and downlink have
been assigned with own spectrum resources, FDD system is continuous in time domain for
both uplink and downlink transmission.

Figure 3.2 TDD and FDD working principal

TDD does not require paired spectrum, since it will assign all its bands to both uplink and
downlink. To distinguish uplink and downlink transmission, TDD system split its time
resources, and switch the system operation frequently between uplink and downlink mode.
Due to the uplink and downlink transmission utilize the same frequencies, TDD system will
assign a guard period between uplink timeslot and downlink timeslot in case that the
synchronization is not accurate enough, which help prevent the interferences between uplink
and downlink transmission. Therefore, TDD system is continuous in frequency domain, but
discontinuous in time domain, as in Figure 3.2.

15
 In LTE systems design, the difference between FDD and TDD mode are minimized.
Besides the differences in physical layer, they have the same design and are transparent to the
layers above. However, the tiny differences between TD-LTE system and LTE FDD systems
define the different characters, performances and applications. For instance, FDD have
identical bandwidth for uplink and downlink, which suits the symmetric application. While
TDD does not have to assign identical time-frequency resources to uplink and downlink,
making it could facilitate asymmetric applications.

3.1.3 Key concepts in TD-LTE system

3.1.3.1 TDD frame structure
Due to the difference between FDD LTE and TD-LTE system, TD-LTE system has its own
frame structure, which could customize the timeslots for uplink and down link transmission,
in Figure 3.3.

Figure 3.3 TD-LTE frame structure (Frame structure Type 2)[20]

Each TD-LTE radio frame takes up 10ms, containing 10 subframes. Subframe 1 and
subframs 6 are special subframes, while the rest are data subframes. Each special subframes
contains 3 special timeslots called DwPTS, GP and UpPTS. DwPTS transmits downlink data
as well as reference signal and control signals, however its length could vary according to
different configurations. UpPTS transmit sounding reference signals and random access
signals; therefore UpPTS will not contribute to the uplink data transmission. GP is the
protection time interval between downlink to uplink switch.
Internet application today has shown asymmetric character between downlink and uplink;
therefore in order to bring flexibility and efficiently utilize time-frequency resources, TD-LTE
system has defined different downlink to uplink allocations. TD-LTE systems support 5ms
and 10ms switch-point periodicities, each has several uplink-downlink configurations as in
Table 3.1.

16
 Uplink-downlink Downlink-to-Uplink Subframe number
configuration Switch-point periodicity 0 1 2 3 4 5 6 7 8 9
0 5 ms D S U U U D S U U U
1 5 ms D S U U D D S U U D
2 5 ms D S U D D D S U D D
3 10 ms D S U U U D D D D D
4 10 ms D S U U D D D D D D
5 10 ms D S U D D D D D D D
6 5 ms D S U U U D S U U D
Table 3.1 TD-LTE uplink-downlink configuration[20]

From Table 3.1, configurations 0 and 6 have uplink preferences, while configuration 2, 3, 4
and 5 allocate more time for downlink in each radio frame. Configuration 1 allocates nearly
the same time-frequency resources for uplink and downlink, making it most similar to LTE
FDD system in terms of symmetry. For 10ms switch-point periodicities, subframe 6 becomes
a regular downlink subframe. For special subframes (S in Table 3.1), 3GPP also defines
different configurations for DwPTS, GP and UpPTS, as in Table 3.2.
Normal CP Extended CP
DwPTS GP UpPTS DwPTS GP UpPTS
0 3 10 1 3 8 1
1 9 4 8 3 1
2 10 3 9 2 1
3 11 2 10 1 1
4 12 1 3 7 2
5 3 9 2 8 2 2
6 9 3 9 1 2
7 10 2 - - -
8 11 1 - - -
Table 3.2 Configuration of special subframes in OFDM symbols[21]

The length of DwPTS, GP and UpPTS could change according to the configuration of
special subframes according to Table3.2, bringing further flexibility to the system. For normal
cyclic prefix (CP), each radio frame contains 14 OFDM symbols for special subframes, and
for extended CP, each radio frame contains 12 OFDM symbols for special subframes.
Different special subframe configurations could affect the coverage capability of TD-LTE
sites.

3.1.3.2 TDD scheduling and acknowledgment

The scheduling and acknowledgement mechanism is more complex for TD-LTE system than
LTE FDD system, due to its asymmetry character. For LTE FDD system, each uplink
subframe could associate with a downlink subframe, therefore the ACK/NACK information
could be rather clear to the systems. While for TD-LTE system, most configurations are
asymmetric. When the uplink subframe excesses to the number of downlink subframes, not
all uplink subframe could find an exclusive downlink subframe to transfer its
acknowledgment, the similar situation happens as the downlink subframe excesses to the
number of uplink subframes. Therefore the solution is that for asymmetric scenarios, an
uplink subframe could schedule several downlink subframes, and a downlink subframe could
schedule several uplink subframes, if there is not sufficient subframes in one direction. The

17
issue with this mechanism is that the accumulated control signal associated to each subframe
of TD-LTE system could be more than LTE FDD system, therefore only less time-frequency
resources could be allocated to payload transmission.

3.1.3.3 HARQ
The design of HARQ is more complicated for TD-LTE system than LTE FDD system, since
TD-LTE system is discontinuous in time domain. For LTE FDD system, the HARQ process
is predictable. When UE send data to eNodeB, and eNodeB receive the data successfully,
eNodeB could send acknowledgment back to UE after processing. When UE receives the
acknowledgment, the HARQ process is finished. While in TD-LTE system, when UE sends
data to eNodeB, eNodeB receives the data successfully and finishes processing the
information; it might still need to wait for a few milliseconds for the coming of downlink
timeslots until it could send the acknowledgment back. The waiting time is related to
subframe configuration and the arriving time of the data, which brings uncertainty to the
system design. The same situation also applies to uplink acknowledgement and negative
acknowledgments. Therefore, the complexity of TD-LTE system transceiver is higher than
LTE FDD system, bringing additional cost to the systems design and manufacture. The
round trip time of TD-LTE system could varies as well, which is a potential risk for the
system performance.

3.2 Evaluate TD-LTE system

From market point of view, it may not be a difficult decision to upgrade MNO’s current
network to the latest technology towards 4G, especially for those operating in developed
markets and many of the emerging markets, where decent data service becomes one of the
main requests from the end users and the selling point for MNO. However, due to the factors
like market strategy, available spectrum, capital strength, operational capability or regulatory
condition, different operators may end up in several different mobile access technologies as
their next generation network solution. Therefore, to identify whether TD-LTE is a really
proper choice for certain operators, much factors could be considered for the decision.

3.2.1 TD-LTE spectrum

ITU has defined the E-UTRAN spectrum bands for LTE FDD and TD-LTE system
respectively. As the deployment of LTE system continue to progress, 3GPP has increased the
bands planned for both FDD and TDD system, from release 9 to release 11. Starting from
release 11, Band 44 has been included for TD-LTE system, which is the lowest frequency
band available for TD-LTE by far. It is good news for TD-LTE operators, especially for
those who plan to construct a stand-alone TD-LTE network; since Band 44 could help those
operators to increase site coverage and reduce construction costs. However, in 3GPP
specifications, FDD systems have been offered with more spectrum bands for operators to
choose.
Availability and global harmonization of spectrum resources is important to radio system
development. At the initial development phase of LTE system, the focuses of global telecom
authorities and operators are mainly on LTE FDD systems. Those preferences have driven
up the auction price of FDD spectrum resources, and with the continuous allocations of LTE
FDD spectrum resources, the remaining FDD spectrums are becoming limited and even
more precious. Based on the previous spectrum auction results, LTE FDD spectrums have
18
higher price per MHz in most countries than TD-LTE spectrums, however the low price
advantage of TD-LTE system could change, if more global operators plan to adopt the
systems. Currently, the majority of TD-LTE bands are 1.8G, 2.3G, 2.6G and 3.5G bands,
while the 700M bands from release 11 has not became mainstream. Based on [22] , Global
operators from Asia Pacific and Europe mainly construct TD-LTE network in 2.3G and 2.6G
bands, making them the most favorite bands for industry product development and
international roaming. Comparing to LTE FDD system, current TD-LTE adopters might
have relative abundant spectrum resources, since TD-LTE could utilize unpaired spectrum
resources, which have not get enough attention before. For the popular 2.6G bands, 2570-
2620MHz bands have been allocated to TD-LTE operators in several European countries,
while in Asia Pacific, some countries intend to allocate the same bands to TD-LTE operation
as their European counterparts, while others intend to allocate all 190MHz bandwidth to TD-
LTE operations[23]. These allocations decisions show that global TD-LTE deployment might
enjoy large continuous blocks of spectrum resources, which is crucial to provide service with
high capacity and data rates.
Beside market advantages, TD-LTE system could utilize fragmented spectrum resources
left in previous deployments. This character will benefit telecom industry even more as carrier
aggregation technology is implemented for LTE-A. However, in TD-LTE systems, due to
guard period, the length of cyclic prefix and different special subframe configuration could
bring varied overheads to the system, and according to existing test results, the spectral
efficiency of TD-LTE is slightly lower than LTE FDD, meaning that the capacity of TD-LTE
system is below to LTE FDD system, by using the same bandwidth and radio configuration.

3.2.2 Coverage
It is often considered that TD-LTE has less coverage range than LTE FDD system in the
same condition. However, the coverage of radio system could be considered as control
channel coverage and traffic channel coverage. The control channel coverage is the minimum
requirement for UE to access the network; therefore the coverage range is relatively large. In
this sense, TD-LTE and LTE FDD have similar coverage. But since LTE system is a data
oriented network, it is meaningless for end user to access the network without minimum data
service guaranteed. When talking about the coverage with minimum cell edge data rate, it is
often to select TD-LTE system with DL/UL configuration 2:2 to compare with LTE FDD
system, assuming the same transmitting power, antenna mode and equivalent bandwidth, for
instance 10M for FDD and 20M for TDD. But in TD-LTE system, the special subframe
could not transmit uplink data; the efficient uplink bandwidth of TD-LTE system is therefore
lower than FDD LTE system. Therefore, for the same minimum cell edge user uplink data
rate, TD-LTE needs to decrease link budget to compensate for the losses in efficient
bandwidth. In this sense, TD-LTE has lower uplink coverage range but higher downlink
coverage range than TD-LTE system. Both TD-LTE system and LTE FDD system is uplink
limited, therefore the minimum uplink performance usually defines the coverage range.
However, there are also other DL/UL configurations, meaning that the uplink and downlink
traffic channel coverage could vary when given the same minimum cell edge user data rate.
Therefore, it is not precise to say that TD-LTE has lower coverage than LTE FDD system.
The coverage range of TD-LTE system differs with subframe configuration and system
requirements of cell edge users.

19
3.2.3 Asymmetric character
Traditional voice service is a typical example of symmetric application, however the data
services have shown different patterns. For LTE FDD system, it allocates the same
bandwidth to both uplink and downlink transmission. Considering the spectral efficiency of
uplink and downlink respectably on given scenario, the user peak data rate and system
capacity are fixed in a certain ratio. However, the actual data traffic does not always follow the
pattern given by the system, and fluctuate according to various factors like time or location,
meaning the time frequency resources could be wasted in either direction. Therefore, the
inflexibility of LTE FDD system could hardly satisfy the diverse traffic patterns, bought by
mobile broadband service and applications with various natures. For TD-LTE system, 3GPP
has defined 7 different uplink and downlink subframe configuration, making it more flexible
to different asymmetric scenarios. TD-LTE system could support both strong uplink biased
scenarios and downlink biased scenarios, as well as symmetric applications.
However, in the current setups, it is often recommend to configure TD-LTE system
downlink to uplink subframe rate and special subframe with identical value for all the cells
under the same geographic area to prevent issues related to interference and synchronization.
Therefore, it still needs to investigate the technical solution of co-existence of TD-LTE cells
with different subframe configuration, and possible self-adoptive algorithms to adjust the
subframe configuration according to the feedback of network traffics. Therefore, TD-LTE
system brings potential to offer customized service and applications based on the user pattern
and requirement with better utilization of time-frequency resources.

3.2.4 Synchronization
Radio systems require frequency synchronization and time synchronization. Frequency
synchronization is relative easy to achieve, requiring the signal frequency of eNodeB is
synchronized with the reference frequency. While time synchronization require the system to
synchronize with coordinated universal time (UTC). For LTE FDD system, only frequency
synchronization is required, but TD-LTE also requires time synchronization with the
accuracy of microseconds, otherwise uplink and downlink transmission could cause
interference to each other.
For LTE FDD system, most eNodeB acquire frequency reference signal for IP based clock
source, and such source could usually support hundreds of eNodeB at the same time. While
in order for TD-LTE system to synchronize, additional implementation are needed, for
instance GPS or IEEE 1588 V2. GPS solution could achieve high accuracy of time
synchronization, but each eNodeB required a GPS signal receiving system, which brings
additional cost of installation and maintenance. Besides, GPS antenna requires clear sky view,
which could bring deployment difficulties for some indoors sites. For IEEE 1588 V2, even
though it is an IP based solution, it is still required to upgrade all the intermediate
transmission equipment to support IEEE 1588 V2 for time synchronization, which could
bring large network upgrade cost and massive operation to some operators, Besides, the
precision of IP based solution could be affected by the transmission network performance.

3.2.5 Channel reciprocity

TD-LTE system uses the same frequency bands for both uplink and downlink transmission,
and the transmission slots are adjacent to each other, therefore the radio channel environment
are highly correlated. eNodeB could estimate the downlink channel condition, based on the
20
receiving uplink information. However, FDD based system use two separated bands to
transfer uplink and downlink information, the frequency dependent channel fading characters
could cause channel reciprocity hard to be exploited, unlike TD-LTE system. The channel
reciprocity could benefit inter cell interference mitigation and multi antenna beamforming
technology, which could efficiently increase cell coverage, capacity and cell edge user
performance. For equipment vendors, the channel reciprocity of TD-LTE systems could also
reduce the complexity of terminal and radio equipment transceivers.

3.2.6 Leverage from other systems

LTE FDD system has been selected by large amount of global operators as the solution for
the next generation mobile broadband. Therefore, it has become a mature technology as
vendors have gained more experiences for the product development and operators also gain
increasing implementation know-how though countless scenarios and conditions. Due to the
scale of LTE FDD system, the cost of both production and implementation has been
reduced. The design of TD-LTE and LTE FDD are same for most of the part, and only a
few differences exist in the physical layer. The radio equipment vendors could develop TD-
LTE capable network products, based on their LTE FDD experiences. As for the core
structure, TD-LTE system could also share with LTE FDD system. The same situation
applies to terminal products, as the TD-LTE compatible and multi mode chipset could be
developed under the same platform of LTE FDD system. Therefore the network deployment
and operation cost of TD-LTE system could be eliminated, and TD-LTE customers could
have more selections on terminal products. As an evolved technology of TD-SCDMA, TD-
LTE could also leverage the development achievement of TD-SCDMA.
As a member of 3GPP standard technology, TD-LTE could interoperate with other 2G,
3G technologies and FDD version of LTE. This brings flexibility for TD-LTE deployment
and application scenarios, as some of the implementation drawbacks of TD-LTE system
could be complement by other widely deployment technologies. Besides, international and
domestic roaming becomes possible between TD-LTE system and other 3GPP standard
technologies.

3.2.7 Summary
In order to evaluate TD-LTE system, it is important to define the metrics of measurement
based on operators’ strength, weakness and network deployment strategies. Basically, the
introduce of TD-LTE system is to tackle the growing data demands and help operators to
mitigate the imbalance between mobile broadband traffics and revenuers, therefore a few
aspects and indicators are most relevant for the objectives. From cost perspective, the
unpaired spectrum unusually cost operators less comparing to paired spectrum resources in
most part of the world. However, most allocated TDD spectrum resources are from high
bands, if the operators could not get low bands of TD-LTE or other technologies for
coverage complement, the network construction cost could offset the benefits from low cost
unpaired spectrum resources. The additional cost could also come from the synchronization
solution, due to the additional GPS system for each sites or upgrading cost of transmission
network. Beside the network site cost, potential cost might come from terminal product
procurement, as TD-LTE system have not reach the scale of LTE FDD system, the profit
margin of operators could be squeezed if TD-LTE operators want to compete with LTE
FDD rivals with similar terminal product line.

21
 From network deployment perspective, it is important for operators to position TD-LTE
system accordingly. Due to TDD spectrums are mostly allocated in high bands, TD-LTE
systems could usually be used as a capacity complement network or indoor high demands
broadband solution. If the operator plans to use TD-LTE as a primary network for capacity
and coverage, the system could still qualify for the job by adjusting the DL/UL configuration
and cell range according to the user pattern. However, the planning should take the network
performance for cell edge users into consideration, even though the goal is to minimize the
network construction. Regardless of how operators position TD-LTE system, it is usually
needed to integrate the system with others to provide end users with a seamless
communication solution, hence the interoperability issues like inter-RAT handover, integrated
voice solution or international roaming should be also be put into focus. To sustain and
improve the network performance, operators need to include network synchronization
planning and multi antenna technology during the deployment, which could include GPS
installation, transmission network upgrade, antenna and feeder system upgrade. For operators
choosing other systems, it might to also be a good idea to keep TD-LTE as an option for
future expansion, due to its abundant spectrum resources and the ability to utilize available
spectrum through radio cognition.
From market or industry development perspective, the mainstream bands of TD-LTE
system is from 2.3G to 2.6G, therefore this affects the development of both network
infrastructure and terminal products. TD-LTE system could leverage the development from
LTE FDD, TD-SCDMA system or other widely deployed 3GPP system, but the
development of TD-LTE system have not reach enough scale to compete with LTE FDD
system. The TD-LTE market and industry still need a few more years to verify whether TD-
LTE system will be widely accepted and successful.
From network performance and application perspective, the spectral efficiency of TD-LTE
are slightly lower than LTE FDD system due to additional overhead, but the difference could
be neglected in most cases, and compensated if the spectrum resources are abundant. The
flexibility bought by asymmetric character could facilitate new types of service and application
in addition to the standard mobile broadband services. Therefore, customization and
differentiation is what TD-LTE system service and application should be built upon.

22
 4 TD-LTE Market Opportunities
and Ecosystem
For a sustainable growth, TD-LTE operators need to retain their traditional revenues and expand their
business scope with new services and applications. By comparing the network performance of TD-LTE and
other systems, the user experiences of standard mobile broadband service and applications could be identified.
Further more, particular and innovated service and applications, which is exclusive to TD-LTE, will also be
discussed. After the discussion of market opportunities based on both standard and particular services and
applications, the development of chipset and terminal products will be summarized, and the market trends and
challenges of TD-LTE chipset and terminal products will be described accordingly.

4.1 Performance evaluation of MBB services

powered by TD-LTE system
The network performance of mobile broadband service is essential for the development of
wireless access technologies. The applicable services and applications over the wireless
infrastructure are largely affected by the network capability that a system could offer. Hence
both the standard or innovative service and applications should be developed based on
several key network indicators. By comparing with other access technologies, the strength and
weakness of TD-LTE system could be identified and used as an input for the market
positioning of the operators. It could then be possible to investigate the methods of
mitigating drawbacks and elaborating advantages of the system against other competitors.
Based on those evaluations, the operator could possibly educate and attract customers by
promoting differentiated services and applications.

4.1.1 TD-LTE network performance

The network performance of radio system could be evaluated in different aspects, and mobile
operators are unusually most interested in the indicators like peak throughput, cell capacity
and latency. Those indicators will have different influences based on the nature of application,
and end users’ mobile broadband experiences could be quantized in those counters.
Peak throughput measures the maximum data rate that an end use could possibly get,
which set the up boundary of application consumption. The applications affected most by the
peak user throughput are those with high date consumption rates, for instance high definition
video streams requires guaranteed data rates for smooth live videos and playbacks. Another
example is cloud service, where the user peak throughput influences the synchronization time
of files. Therefore, when it comes to the testing of user peak throughput, file downloading
and uploading based on TCP and UDP have usually been selected as a typical application.
However, to achieve the peak throughput is never easy, theoretically, it is related to the
factors like bandwidth, multi antenna mode, modulation type and error correction coding.
TD-LTE system has been designed with significant improvement for peak throughput,
comparing to its predecessors. According to [24], ITU-R require the peak spectral efficiency
to be 15 bps/Hz and 6.75 bps/Hz for downlink and uplink respectively, and those targets has
been fulfilled by both TD-LTE and LTE FDD, but under strict condition for instance 4×4
23
MIMO for the downlink with 64QAM modulation scheme and symmetric subframe
configuration for TD-LTE system. As a result, LTE release 8 has targeted 15.3 bps/Hz and
15 bps/Hz for LTE FDD and TD-LTE downlink peak spectral efficiency, hence the
downlink peak spectral data rate is slightly higher for LTE FDD system due to the difference
in physical layer design. However, the uplink peak spectral efficiency haven’t meet the
requirement set by ITU-R, but LTE FDD still gain a little edge over TD-LTE in terms of
uplink peak throughput. As a qualified 4G standard, LTE-Advanced in released 10 have
implemented even more advanced multi-antenna technologies, i.e. 8×8 MIMO for downlink
and 4×4 MIMO for uplink, which sets high standards for network infrastructure deployment
and terminal development, but provides improved peak spectral efficiency. The downlink
peak spectral efficiency have been doubled, and the uplink spectral efficiency have been
improved 4 times, which have surpassed the requirement set by ITU-R.
Ascom has in [14] demonstrated the theoretical calculation for the peak user data rates for
various multi antenna mode and modulation scheme. Taking the most frequently used 20M
Hz for TD-LTE or 2×10MHz for LTE FDD, the downlink peak data rate range from 14.4
Mbps for QPSK and single input single output (SISO), to 172.8 Mbps for highest order
modulation 64QAM and 4×4 MIMO; its corresponding uplink peak user throughput range
from 7.5 Mbps for QPSK and SISO, to 21.6 Mbps for 64QAM and SISO. The results apply
to both TD-LTE and LTE FDD system when no error correction coding is used, and the
peak data rate has been reached a rather high level. The calculation have been verified by the
strictly controlled LSTI lab environment test results shown in [25] and [26]. Given the 2×2
MIMO for downlink and 1×2 MIMO uplink, TD-LTE system could achieve highest
downlink peak spectral efficiency of 8.5 bps/Hz in 64 QAM, comparing to 8.6 bps/Hz of
LTE FDD system, and highest uplink spectral efficiency of 2.7 bps/Hz in 16 QAM,
comparing to 2.8 bps/Hz of LTE FDD systems. The peak spectral efficiency of TD-LTE
system reaches the lowest point of 2.8 bps/Hz and 1.4 bps/Hz for downlink and uplink
respectively, when QPSK modulation is applied. The test result by LSTI shows TD-LTE
peak spectral efficiency could approach theoretical boundary as the radio environment is
strictly controlled in lab environment and the lab test peak spectral efficiency could surpass
the 3GPP designed LTE peak spectral efficiency. The difference between TD-LTE and LTE
FDD is minimum, both of which have been significantly improved over previous 3GPP radio
access technologies.
However, the theoretic calculation and lab test results could hardly been achieved in field
test, let alone in the live network. The field test environments and live networks face much
more complicated radio environments for instance the multi path loss, line of sight loss and
interferences from other base station or terminals. Therefore, some of the general pre
condition need to applied for the end users to get the peak user throughput in the field test
environment, for instance the user need to try to get a better SINR near the antenna; stay in a
stable position where the SINR would not fluctuate, as moving at a certain speed could cause
Doppler shift and the radio environment cannot keep at an optimal level; the base station
need to reject the admission from other subscribers as the time-frequency resources could not
be shared. Under certain circumstances, TD-LTE could achieve peak throughput of 60 Mbps
and 19 Mbps for downlink and uplink with 20 MHz spectrum and 2×2 MIMO in field
environment, which is comparable with LTE FDD performance of 60 Mbps and 23 Mbps
for downlink and uplink with 2×10 MHz and 2×2 MIMO in the field test environment [22].
In the live network environment, users are randomly distributed in the cells with different
radio environment. Depending on the scheduling mechanism of radio equipment, subscribers
24
will be allocated time-frequency resources accordingly, but generally each subscriber will get
much lower date rate than the peak throughput, therefore the cell capacity could also reflect
the users real network experiences. According to the test results presented in [22], the average
throughput of a TD-LTE cell could only achieve 25 Mbps and 11 Mbps with the similar
spectrum and antenna mode, when multi users are admitted and neighboring cells are
activated. For each end user, the experienced data rates are slightly different from the
measured data rate in the test environment. Since the telecommunication measurement
usually takes the physical layer data rate for comparison, while the end users’ experiences are
usually been described in the data rate of application layer. The overhead of physical layer,
MAC layer, RLC layer, PDCP layer, IP layer, TCP/UDP layer and application layer could
degrade the actual user data rate that a subscriber could utilize, however the data rate in
physical layer could still approximately define the user experience and is capable for
comparison between different access technologies.
Latency is another metric, which could affect the user experiences of end users. There are
several types of latency, and most commonly measured are random access latency, idle to
active latency, handover latency and user plane latency. The latency could be measured as
one-way latency or round trip time, a concept brought from data network. Random access
latency defines how fast a user could camp to the network after switching on its devices, but
it is of less importance, as many mobile users tend to keep connected to the network. The
idle to active latency defines how fast a UE could switch from idle mode to connected mode
and start with the data service. As those latencies are in the grade of tens of milliseconds, the
end users could hardly notice them. Handover latency defines how fast a UE could connect
to the target cell during handover process, and resume data service. Based on the handover
type, handover latency could usually be categorized as intra-site handover latency and inter-
site handover latency. If the handover fails, the random access process will be triggered for
UE to get connected to the target cells, and the latency will presumably higher, but still can
hardly be notices, due to the packet switched nature of TD-LTE network. However, the user
plan latency could be more easily noticed, especially when the subscribers are using real time
applications, like video streaming or conferencing. According to [24], ITU-R’s standard is 10
ms for both uplink and downlink one way latency and LTE have achieved 4 ms for LTE
FDD system and 4.9 ms for TD-LTE. This results are a little bit ideal in the real situation, as
from the test results demonstrated in [22], TD-LTE could only achieve average user plane
round trip latency for around 20 ms to 30 ms, which is slightly higher than LTE FDD system
with the deviation for around 2 ms to 7 ms. This could be easily understood, as TD-LTE
system divide the time resources to uplink and downlink, the transmission need to wait for
the right timeslot to be finished, resulting the increases in the average user plane latency. But
how much the increased latency could affect the user experiences is dependent to the
application that the subscriber is using.
The asymmetric character also affects the performance of TD-LTE network, as the
previous evaluations are based on the symmetric DL/UL configuration in order to be
compared with other symmetric systems, like LTE FDD. However, as the spectral efficiency
of uplink and downlink varies, the radio between TD-LTE subframe does not necessarily
reflect the rate between uplink and downlink peak throughput. Therefore, based on the
verified spectral efficiency in [25], which is 5.7bps/Hz for downlink and 2.7bps/Hz for
uplink for 20MHz bandwidth, 2×2 downlink MIMO, 1×2 uplink SIMO and 16QAM
modulation, it can estimate the downlink and uplink asymmetric peak data rates. The special
subframe configuration is assumed to be number 1 with normal CP.
25
 UL/DL Config. D S U DL Peak Thr.(Mbps) UL Peak Thr.(Mbps)
0 2 2 6 37.46 32.4
1 4 2 4 60.26 21.6
2 6 2 2 83.06 10.8
3 6 1 3 75.73 16.2
4 7 1 2 87.13 10.8
5 8 1 1 98.53 5.4
6 3 2 5 48.86 27
Table 4.1 TD-LTE asymmetric peak throughput estimation

The estimation could be reflected in the field test results in [22], as the average test result
shows 80 Mbps and 8 Mbps respectably for downlink and uplink in subframe configuration
number 2 with 20 MHz bandwidth. Based on the previous discussion, the uplink and
downlink spectral efficiency is slightly higher for LTE FDD, comparing to TD-LTE, but if
the user pattern contradicts with the real ratio between downlink and uplink throughput, the
imbalance would also waste time-frequency resources.

4.1.2 TD-LTE enabled MBB service performance

Based on the analysis of TD-LTE network performance, the throughput of TD-LTE
system is generally lower than LTE FDD system with similar configuration, especially in the
uplink, but the network capacity and user throughput have been largely increased comparing
to 3G systems. However, the difference in throughput between TD-LTE and LTE FDD is
not conspicuous as the disparity between WCDMA and TD-SCDMA from the initial
deployment phase, therefore it could hardly be noticed by the end users. The additional
latency brought by TD-LTE system subframe design could however influence several types
of MBB services and applications, especially real time application; therefore TD-LTE based
real time services and applications need to be specially designed to compensate the additional
latency.
In live network, subscribers will be scheduled by the eNodeB together with other users,
therefore depending on the radio environment and network loads; subscribers will be serviced
with different qualities. Hence, when the network configuration follows the user pattern of
serving areas, the subscriber will have better chance of getting satisfactory services. The
majority of today’s mobile broadband traffic is Internet traffic, and based on the data from
3G networks1, it could get a glimpse of the users behaviors on the standard mobile broadband
service. The downlink throughput is four to six times of the uplink throughput in a regular
3G commercial network. However, the ratio between downlink and uplink could vary in
different markets, for instance the downlink to uplink traffic radio in Hong Kong is
approximately four, while in several of the European countries, the ratio goes up to more
than five. Even though the 3G traffic models could not completely simulate the traffic model
of LTE network, but generally the downlink to uplink traffic radio is not constant, which
could varies with network performance, application consumption nature, tariff and also
geographic regions. According to 3G network pattern, the most used UL/DL subframe
configuration of TD-LTE network could be number 1 and number 3 with a mixture of even
heavier downlink biased configuration. Comparing with other systems like LTE FDD, TD-

1
Data from China Mobile Research Institute presentation on high performance TD-LTE network
2
Mobile broadband with HSPA and LTE – capacity and cost aspects, published in 2010.
26
LTE could achieve more network utilization rate, better and balanced individual MBB service
performance, with lower spectral efficiency.

4.2 TD-LTE enabled new services and

applications
Based on the flexibility on asymmetric data rate between uplink and downlink, TD-LTE
system could facilitate certain type of service or application and penetrate particular market
segments. The MNO could customize its offer based on the user behaviors and service type;
the character of TD-LTE system is also potential to bring up new types of applications. For
instance, certain consumer type services usually have strong preference on downlink, while
M2M type services in the other hand have strong bias on the uplink.

4.2.1 Downlink biased services

The downlink throughput of live network traffic is much higher than uplink throughput, due
to the contribution from downlink-biased services, which have been growing with the
increasing popularity of Internet services and specific browsing habits. Some of the examples
include video sharing sites YouTube and Vimeo, music and sound streaming sites Spotify and
SoundCloud, video on demands services Netflix, social network Facebook and Twitter, or
location services Google Maps. Hence, many of the most frequently used mobile applications
and services with significant data consumption require more for the downlink capacity.
However, applications and services will contribute differently to the imbalance between
uplink and downlink traffic. According to the estimation in [15], the difference between
downlink and uplink data rate could range from 50 kbps in text based social network Twitter
to more than 8 Mbps in video streaming services, and on average the gap between downlink
and uplink data rate are around 1-2 Mbps, which could mostly be observed in on-line
browsing or social network services. When the simultaneous user number accumulates, the
total downlink consumption could dominate the network traffic. For the downlink biased
services, both LTE FDD and TD-LTE could be qualified. For LTE FDD, even though the
allocated uplink and downlink spectrum are identical, due to different spectral efficiency, the
ratio between downlink and uplink throughput are around 3:1 in commercial networks. As for
TD-LTE, it could support around 20 times differences between downlink and uplink
throughput, binging more flexibility if the downlinks traffic would further grow.
Besides the consumer applications, TD-LTE could also facilitate dedicated and customized
services with downlink bias, for instance broadcast service for public sectors, mobile
advertisement and entertainment systems on vehicles. The nature of the traffic for those
services is unidirectional or nearly unidirectional; therefore it is not efficient to offer those
services with symmetric systems like LTE FDD, which could waste time-frequency resources
and drive up the service cost. However, TD-LTE could make those delicate services more
affordable. The network implementation strategies is to utilize small pieces of authorized
fragmented spectrum resources or cognitive radio, and configure the network with heavy
downlink preference, for instance the DL/UL subframe configuration number 4 or even
number 5. The dedicated broadcast network could be built upon the primary network,
therefore the service of primary network would not be affected and the dedicated network
service quality could also be guaranteed with minimum network construction and operational
cost.

27
4.2.2 Uplink biased services
Uplink biased services are usually needed by specific sectors or groups, therefore TD-LTE
system could help operators to step into some of the new markets with cost efficient services
and applications. For public sectors, the uplink-biased services that TD-LTE could offer are
video surveillance for security, and live traffic monitoring for traffic control and scheduling
center. Traditionally, those services are built upon fixed broadband service, therefore the
surveillance locations are restricted and dedicated traffic line implementation could be costly.
Medical sectors could benefit from the home medical monitoring and reporting systems,
enabled by TD-LTE system with uplink preference configuration. This mechanism could
reduce the medical expense for the sector, relieve the pressures of limited medical resources
in some regions and facilitate cross-region medical consolation and cooperation. However,
the uplink traffic of medical monitoring service could vary based on the reporting contents.
If pure data and text informatics are reported, the uplink traffic data rate would only be
around a few hundreds kbps, when the video monitoring is necessary, the uplink traffic could
be up to a few Mbps, based on the video quality required.
Another sector, which requires uplink-biased services, are broadcasters. TD-LTE could
reduce the expenses brought by delicate satellite and microwave links. Due to the broadcast
level video are of high quality, the downlink to uplink subframe configuration could be
configured as number 0 or number 6. TD-LTE could enable fast deployment of reporting
systems with affordable cost and guaranteed quality. However, the restriction is that certain
system could only be implemented cost efficiently in the regions, where decent mobile
infrastructure has already been deployed.
M2M applications are new types of uplink-biased system, which could be enabled by TD-
LTE system. Even though M2M applications have not been widely deployed, the cost
efficiency bought by TD-LTE system would bring new development opportunities for M2M
applications. In private sectors, M2M applications are future proof technologies, where
consumers would bring more of their traditional electronics like refrigerators or air
conditioner on line for remote and smart management, and positioning services of mobiles or
other electronics could also be in demands. In industry and commercial sectors, M2M
applications could be used to track logistics along the supply chain for large cooperates or
enable mobile payments possibilities for private business owners. The M2M applications is
usually scattered and large in qualities, therefore mobile systems like TD-LTE, which could be
configured with uplink preferences, became an optimal option, but the requirements for
simultaneous connection number could bring challenges for TD-LTE solutions.
Even though TD-LTE system is more cost efficient for uplink biased services and
applications, comparing to other technologies, the scale of certain applications is of most
importance for operators and other parties like service providers and developers.

4.2.3 Specific scenario and service packages

TD-LTE could be served as a primary network for standard consumer mobile broadband
network, but some operators also choose to position TD-LTE network as a complement
network, because TD-LTE system could be implemented with affordable spectrum,
construction and operational costs; it could also interoperate with other mobile systems,
adjust flexibly based on changing traffic patterns and efficiently control the interference level

28
with other systems. Hence, a few services could be made possible based on those characters
to target certain mobile broadband application scenarios or particular customer segments.
Enterprise access network services could help TD-LTE operators generate new revenues in
addition to its regular services. Depending on the business area of served companies, the data
consumption could vary, but the Internet has changed the traditional manners of working in
increasing numbers of market segments, causing the data consumption demands increase in a
fast pace. TD-LTE could offer enterprise customers decent traffic volume with low cost, high
availability and satisfactory security. As a complement network, TD-LTE could utilize
separated and unpaired spectrum resources, which could efficiently mitigate the interference
to the primary network. Beside of constructing TD-LTE network as an additional data layer
to primary mobile network, TD-LTE femto solution has drew increasing attention for
enterprise solutions.
TD-LTE operators could also consider providing mobile network services to replace the
fixed broadband services for home users, since TD-LTE is a data centric network. This
service could especially gain popularity in some of the remote areas, where it is too costly to
provide fiber access to each house in the region. As the TD-LTE network could share most
of the infrastructure from sites, backhaul to core structure with the primary mobile networks
like 3G and LTE FDD, the cost of construction and maintaining could be eliminated. With
adoptive network configuration, TD-LTE could easily adjust according to the specific user
pattern of the region. As the residential areas could be scattered, the network configuration
could be highly customized without interferences with each other. For the remote residential
areas, operators could also provide voice solution over TD-LTE, in addition to the mobile
broadband services.
Another TD-LTE enabled application could be the replacement network for radio
backhaul transmission or LTE-Advanced relay. For radio backhaul network solution, TD-
LTE system could replace the traditional microwave radio links and fiber networks to
transmit data between base station and core network with satisfactory data rates. As backhaul
leasing constitute a large sum of operators network operational cost, the vacant TD-LTE
frequency resources that operators have been left from other network implementations, could
help increase operator’s profit margin. In LTE relay scenarios, TD-LTE air interface could be
used to connect Donor eNodeB and Relay nodes, which could save the spectrum resources
of primary network, and prevent interference with different nodes by utilizing spectrum
recourse in a different band. Hence, the network performance could be further enhanced.
For some temporary scenarios, TD-LTE system could bring more flexibility to operators’
network operation. For instance, TD-LTE could adapt to variation of user pattern rapidly in
scenario of music festivals and sports events. The user patterns would experience dramatic
changes during the time of events, for instance the uplink traffic could presumably climb
faster than the downlink traffic. The same situation also applies to some of the tourist resorts,
where the user consumption and patterns would change completely between high seasons and
off seasons. Therefore, it is not efficient to dimension the network based on annual average,
and dynamic changes of network configuration could help operators to provide users with
satisfactory services.

4.2.4 TD-LTE market and operation strategies

TD-LTE enabled services could be categorized into standard mobile service, dedicated
services and customized services. Operators could position TD-LTE systems as the primary
29
or complement networks, enabling different types of services and applications, which could
influence the network implementation planning and market strategies of TD-LTE operators.
Table 4.2 provides a summary for the identified services and applications that TD-LTE could
enable based on its characters.
Video sharing
Music and sound streaming
Consumer Services
Video on demands
and Applications
Downlink Social networks
Biased
Services Location services
Broadcast service for public sectors
Dedicated Systems Mobile advertisement
Entertainment systems on vehicles
Security video surveillance
Live traffic monitoring
Dedicated Systems
Household medical monitoring and reporting systems
Uplink
Broadcaster reporting systems
Biased
Services Connecting home electronics service
Devices positioning service
M2M Applications
Logistic tracking
Mobile payments
Enterprise access network
Household fixed broadband replacement
Radio backhaul transmission replacement
Specific Scenario and Services
LTE-A relay
Festival and events broadband access
Tourist resort broadband access
Table 4.2 TD-LTE enabled services and applications

To enable such services and applications and differentiate with other radio access
technologies, TD-LTE system has several of its core competitiveness, including improved
network capacity and data rates, utilization of unpaired and fragmented spectrum resources,
efficient spectrum, construction and operational costs, adoptive downlink to uplink
asymmetric configuration, interoperability and leverage from other 3GPP systems.
Therefore, based on the competitiveness, several of strategies of TD-LTE network
operation could be identified.
1. Focus on the increasing data demands: TD-LTE is a data centric mobile broadband
solution, therefore its primary objective is to relieve the network pressure brought by

30
 mobile traffic tsunami, especially when it is positioned as the primary mobile
broadband network. However, through hierarchical network structure, TD-LTE
could also supplement end users in high demands locations with additional capacity as
the complement network.
2. Adjust network configuration according to user patterns: Due to the user patterns
could change with time and locations, the network configuration of TD-LTE system
could adjust according to user behaviors to achieve better network efficiency and user
experiences.
3. Obtain additional revenues though customized services: TD-LTE could enable
asymmetric services and applications for various segments from public sectors,
medical sectors to broadcasters. The cost efficiency and satisfactory performance are
the main reasons for TD-LTE to overtake traditional solutions.
4. Cultivate new markets on M2M applications and wearable devices: By using
fragmented spectrum resource and applying uplink prior configurations, TD-LTE
could drive the growth of M2M applications by providing scale and low cost
connection solutions for devices.
5. Explore new business segments like enterprise and household services: When
operator position TD-LTE as a complement network, it could utilize the unpaired
spectrum resources by implementing the network for different markets and exploring
new revenues. Beside TD-LTE’s mission to complement the data consumption in
some of the high demands locations, it could be served as enterprise network or
household fixed broadband replacement, since those areas are scattered and the
spectrum resources could be reused.
6. Reduce internal operational costs: For operators with unpaired spectrum resources,
TD-LTE could be used as a backhaul transmission solution, which could help
operators to reduce operational costs.
TD-LTE system could provide comparable mobile broadband service to help operators to
bridge the gaps between user demands and network capability. However, its deployment
strategies are more flexible than other technologies like LTE FDD. For operators, who are
intent to explore new services and markets, TD-LTE could offer more options in addition to
the traditional services and source of revenues.

4.3 The development of chipset and terminal

equipment
Proper chipset and terminal support is the key to the success of radio access system, and it
also applies to TD-LTE. End users’ decision among different systems are usually associated
to terminal products of more favorable, and different terminal products like handsets,
dongles, MIFI or mobile modems, act in different roles in the market; therefore the
competitiveness of those products against their competitors could influence the development
of the whole TD-LTE ecosystems. Therefore TD-LTE enabled chipsets and terminals in the
market could be discussed, and the roadmap of major chipset and terminal vendors could be
demonstrated. The discussion also includes the multi-mode multi-band products, which could
benefit operators with several systems other than TD-LTE and increase data roaming
possibilities among different technologies.

31
4.3.1 Introduction to TD-LTE terminal products
The developments of terminal product and infrastructure products are correlated in many
ways. The popularity of different types of terminal products influence the network
deployment focuses, especially when smart phones have revolutionized the mobile industry
and made data revenue gradually taken over the traditionally voice revenue. To meet end
users’ network expectations, both proper terminals and infrastructures need to be provided.
The terminal products available in the market could be grouped into several categories,
including smart phones, tables and notebooks, data cards or dongles, wireless routers and
CPE, M2M and other consumer electronics products, as well as modules.
Data cards and dongles are usually the pioneers of the terminal family. With the initial TD-
LTE system deployments, dongles are usually among the first to be handed to friendly users
and appear in the markets. The plug-and-play and high-speed characters of TD-LTE dongles
accords with users’ browsing habits. For early adopters, it is good chance to experience the
performance of TD-LTE systems, while for subscribers requiring both mobility and data
qualities, like business travellers and salespersons, TD-LTE dongles is a good solution for
them. However, with the popularity of the systems, the market of smart phones, tables and
notebook are presumably to grow and becomes a major battlefield for different access
technologies and terminal vendors. The target segments of wireless routers and CPE are
enterprise and household, therefore the capacity, stability and security of TD-LTE enabled
routers and CPE are of most concerns. In the areas of M2M applications and consumer
electronics, TD-LTE terminals could include sensing and measuring devices, surveillance
cameras, point of sale machine, vehicle entertainment systems, and smart household
electronics. Even though those applications have not been widely applied, their core
components like TD-LTE enabled modules have already been developed and commercialized
by terminal vendors with large range of selection, which lays a solid foundation for the
growth of diversified applications.
TD-LTE share similarity with LTE FDD and TD-SCDMA system, helping it leverage
from the development of LTE FDD and TD-SCDMA terminal products. To support legacy
systems and international roaming, TD-LTE terminals development has also show the trend
of multi frequency and multi mode support.

4.3.2 Chipset and terminal development

4.3.2.1 Chipset vendors and development
According to [27], global chipset vendors have been or planned to be involved in the
development of TD-LTE chipset proactively, including 17 vendors globally. Those vendors
step into TD-LTE development based on its traditional strength on TD-SCDMA, LTE FDD
and WiMAX, therefore the commercialized chipsets have compatibility with different
systems. Beside of TD-LTE single mode chipset, vendors have developed multimode chipsets
with selection from 2 to 6 modes.
Qualcomm releases the most comprehensive TD-LTE capable chipset, which support 6
modes, including the most used 3GPP standards from GSM, CDMA2000, UMTS, TD-
SCDMA to TD-LTE and LTE FDD. Other traditional FDD chipset vendors like Hisilicon,
traditional TD-SCDMA chipset vendors like Leadcore and Spreadtrum, and greenfield TD-
LTE dedicated vendor Innofidei have followed up by releasing 5 mode chipset, including
GSM, UMTS, TD-SCDMA to TD-LTE and LTE FDD. The 5 and 6 modes chipsets are
32
expected to become the main forces of global TD-LTE chipset markets. However, some
vendors have also provided chipsets customized for Chinese markets, which is presumably to
be the largest TD-LTE market, including 4 modes (GSM, TD-SCDMA, TD-LTE and LTE
FDD) and 3 modes (TD-SCDMA, TD-LTE and LTE FDD) chipsets.
By far, both 40 nm and 28nm chipset production technologies have been adopted. In the
initial phase, 40nm technology is the main stream, which is often adopted by wireless routers
and CPE. With the introduction and massive production of 28nm chipsets, TD-LTE
compatible multimode smart phones are expected to increase.
Taking Qualcomm as a example, as one of the largest chipset vendors in the world, it has
established the strategies of common platform of TD-LTE and LTE FDD, therefore based
on the similarity of the two system and leverage of its FDD chipset development experiences,
Qualcomm merge the development of the two systems. It also established two major
products line targeting dongles and wireless routers market with MDM 9200 and MDM 9600,
and targeting smart phones and tables market with MDM 8960, which adopt 28nm
technology.

4.3.2.2 Terminal vendors and development

The development of TD-LTE capable terminal has also been reached certain scale with
multiple selections for operators and end users. According to the statistics by April 2013 in
[2], there have been 166 TD-LTE compatible terminal devices released in the global market.
However, the majority of the terminal products are dongles and wireless routers,
overwhelming the terminal markets of TD-LTE. There are only 18 smart phones and 2 tables
products, out of the total 166 devices. Even though the TD-LTE compatible smart phones
include some of the most popular modules, the qualities need further growth. According to
[27], the operation frequency of TD-LTE terminal products are from 2.3G and 2.6G bands,
which aligns with TD-LTE infrastructure deployments. For lower bands like 700M, there is
no supported terminal product.
The development of TD-LTE terminal products would need to transfer the focuses from
dongles to smart phones. As dongles and wireless routers have became the major TD-LTE
products, smart phones could bring new development of TD-LTE ecosystem. Major vendors
have also target the smart phone market as their long-term strategies. Samsung have
commercialized its flagship smart phone products of Galaxy series with TD-LTE versions.
Sony, Huawei, LG, HTC and ZTE will also try to seize the market opportunity by releasing
their TD-LTE smart phone product lines. The procurement of global operators, like Softbank
in Japan and Bharti Airtel from India, has showed that the massive distribution of TD-LTE
smart phones is expected to come.
The TD-LTE terminal strategies of ZTE reflect the market development of TD-LTE
terminal products. ZTE have defines three stages for the development of TD-LTE products,
including early stage, growing stage and mature stage. In the early stage, the data only
terminals are the main focus, and during the growing stage, smart handsets are expected to
reach scale with enhancement in voice solution and real time services. Therefore, TD-LTE
terminal development is currently in the growing phase. While in the mature phase, TD-LTE
terminal products would be adopted in M2M applications and consumer electronics.

33
4.3.3 Trends and challenges of TD-LTE terminal products
Based on the current development of TD-LTE chipsets and terminals, the challenges and
trends of TD-LTE terminal products could be identified.
1. Multimode terminal products are essential to TD-LTE ecosystem, and will become
the mainstream of TD-LTE terminal products. As leverage from LTE FDD for
convergent network and international roaming, and merging TD-LTE with legacy 2G
and 3G systems becomes a common network deployment strategy for TD-LTE
operators, the abundance of multimode terminal products is key to the operators’
strategies. With the increasing supports of chipset vendors, the multi mode terminal
products would enjoy both high performance and low costs, becoming a natural
choice for global TD-LTE operators.
2. Multi band support is the trend of TD-LTE terminals. Currently, TD-LTE terminals
mainly support 2.3G and 2.6G bands, which aligns with infrastructure development.
However, the support for lower band like 700M and multiband would be one of the
main requests from TD-LTE operators, since operators usually implement lower band
networks for coverage. The supported band is an important factor in the development
of chipsets, hence operators need to define its network frequency planning at a early
stage and form global unity, which could facilitate the development for chipset and
terminal products.
3. Smart phones are the driving force of TD-LTE ecosystem development. A large part
of mobile broadband traffic and revenue are generated from smart phone subscribers.
Therefore, TD-LTE ecosystem needs to gain an edge by providing smart phone
products with low power consumption, high performance and competitive costs.
Large selections of smart phone product from low ends to high ends need to be
placed into the markets. Unlike the data only terminals, TD-LTE smart phones need
to support satisfactory voice solutions like CSFB or VoLTE. Only with the support
from end users, TD-LTE smart phones could then reach scale, reduce the production
cost and contribute to the ecosystem.
4. Terminal products need to customize for specific applications and segments. For
instance, wireless router and CPE for delicate enterprise solution and private business,
M2M devices for public sectors and smart home solutions. The diverse and
customized terminal products would bring new market and growth.

34
 5 Network Deployment Strategies
and Cost Analysis
The chapter discusses the practical network deployment opportunities and strategies when TD-LTE coexist
with other networks like GSM, UMTS and LTE FDD, including topics about voice solution,
interoperability, sharing possibilities and system positioning. A cost analysis is then carried out in the scenario
of TD-LTE implementation reusing HSPA network, and the scenario of TD-LTE femto implementation.
The corresponding network deployment strategies and cost control mechanism are then discussed based on the
calculation results.

5.1 Coexistence and interoperability between

TD-LTE and legacy networks
Along with the implementation of TD-LTE system, MNO will usually maintain their legacy
network at the same time. 2G and 3G networks will presumably coexist with TD-LTE in the
years ahead. Usually the service over 2G, 3G and TD-LTE network will be provided as a
bundle, since the different systems could complement each other by interoperability and
service division.
During the initial deployment phase, most TD-LTE operators have not deployed IMS
infrastructures; therefore it is not possible to provide voice solution based on TD-LTE
network. In order to provide users with both high-speed mobile broadband service and
traditional voice service, CSFB becomes the solution. The basic concept of CSFB is to use
TD-LTE for mobile data traffic and legacy networks as voice solution. The CSFB operation
includes several key procedures; including combined EPS/IMSI attach procedure, combined
TAU and LAU procedure, mobile originated call procedure and mobile terminated call
procedure. UE initiates the combined EPS/IMSI attach procedure after powers on, and
MME of TD-LTE network will notify the information to the core structure of 2G and 3G
networks. After UE is connected to TD-LTE network, it will also periodically initiated the
combined TAU and LAU procedure and let all the core structure of 2G, 3G and LTE
network could get informed about its current location information. Taking the CSFB
procedure between UMTS and TD-LTE as an example, when UE initiate an outgoing voice
call requests to TD-LTE network, MME will notify the eNodeB to start the CSFB procedure
and transit UE to UMTS network for the voice service. For the scenario of incoming voice
call, the paging information is transmitted from MSC in UMTS core to MME in TD-LTE
core and eventually passed to UE; then the UE send CSFB request to MME and the eNodeB
initiate the CSFB procedure after receiving the information from MME. The transition from
TD-LTE to UMTS network could be done by Inter-RAT handover or redirection. For Inter-
RAT handover, the UE will be transited directly to the target cell, while in redirection
procedure, UE have to initiate the voice call setup procedure to the UMTS cells with the best
signal strength. However, even though the CSFB solution could leverage the voice call
capability of legacy networks, terminals will still loss the high-speed connection that TD-LTE
system offered during the CSFB procedures.
35
 TD-LTE roll out are usually carried out in different phases, during the initial deployment
phases, TD-LTE systems would usually only be deployed in the most densely populated
areas. The network deployments would then expand into boarder areas gradually. Hence, the
interoperability between TD-LTE and legacy networks would help operators to provide
ubiquitous communication service to subscribers. TD-LTE could support optimized mobility
management towards other 3GPP legacy systems like GSM and UMTS, and for non-3GPP
operators like WiMAX, limited handover and roaming functionality is still protected. As a
high-speed mobile broadband solution, TD-LTE could be positioned to fulfill the increasing
data demands instead of providing full coverage. Therefore, operators could maintain the
legacy networks like GSM for guaranteed voice service and UMTS for basic data service, both
of which could provide a national level coverage, while TD-LTE network for high quality
mobile broadband for most needed locations and groups. In the locations, where both TD-
LTE and legacy network are presented, the TD-LTE could be set as the prior access network,
since its user experience and network capacity are considered to be better than legacy
networks.
Besides the positioning and interoperability in the network level, TD-LTE could also
explore the site sharing possibility with legacy systems, since most operators usually have
already deployed a certain amount of legacy network in the serving region. The sharing could
be executed in several levels, from site locations, backhaul facilities, to antenna systems.
However, most legacy networks like GSM and UMTS are operating in lower spectrum bands,
while TD-LTE operators usually operate TD-LTE system in higher spectrum bands like 2.6G
band, meaning that usually the coverage range of legacy networks are higher than TD-LTE
systems, due to the propagation character of different frequency. Therefore, TD-LTE sites
might not find proper legacy sites to reuse in some of the locations. However, the coverage of
2.1G UMTS systems is comparable with 2.6G TD-LTE systems, therefore the 3G operators
of 2.1G UMTS could enjoy high sharing efficiency when 2.6 TD-LTE system are selected.
The reuse between TD-LTE system and legacy network could help operators to reduce
production and operational costs of TD-LTE system, but it might also bring inflexibility in
the network planning, making that it could be difficulty to optimize both the TD-LTE and
legacy system at the same time.
Based on the diverse performance brought by different systems, operators could also
differentiate its service offer and tariff. It could position legacy network as a basic package
and TD-LTE as the preliminary service at the initial phase, when TD-LTE is being deployed.
However, the production cost per GB for TD-LTE system is presumably lower than legacy
network, operators could gradually transit its subscriber to TD-LTE systems, providing a
seamless and enhanced mobile broadband experiences.

5.2 Convergence of TD-LTE and LTE FDD

Among the global TD-LTE operators, many have also chosen LTE FDD systems to provide
a convergent network solution, exploring the benefits of both systems. TD-LTE and LTE
FDD shares many similarities in system design, but issues of the convergence networks
remain from aspects like network implementation to market strategies.
TD-LTE and LTE FDD could operate in multiple frequency bands, which could enhance
the overall system capacity and performance in high demands areas. For instance, operator
could implement one layer of LTE FDD system in lower bands for coverage, and
complement the network capacity with TD-LTE system and LTE FDD systems in some
36
higher bands. Since the operating spectrum of TD-LTE and LTE FDD are from different
bands, the interferences between the two systems could be presumably mitigated. As for the
mobility management, TD-LTE and LTE FDD system share the same signaling procedures,
and all mobility management functionality are supported within or between TD-LTE and
LTE FDD system, but the handover between TD-LTE and LTE FDD system bring
additional latency, which however could hardly be noticed by data traffic users. The
convergence network of TD-LTE and LTE FDD could help operators enable global roaming
to both of the technologies, as some oversea LTE operators might only implement the
network in single mode.
The sharing extent between TD-LTE system and LTE FDD system are rather large.
Besides the sharing of site location, backhaul and antenna systems, TD-LTE and LTE FDD
could also share radio equipment like base band processing units and core infrastructure, as
many equipment vendors are developing the two technologies under the same platform.
The technology maturity level and implementation experiences of TD-LTE and LTE FDD
convergence network grows over the time as more operators implement the solution, but it is
still needed to position the two networks based on the technological and market strategies. A
common solution is to position TD-LTE system as the complement network and LTE FDD
as the primary network. When the network loads of LTE FDD systems increase, users will be
directed towards TD-LTE network. Based on the counters of the present traffic of both
systems, UE will be transited between networks and new users will be admitted to different
networks according to load balancing principals. Operators could also position TD-LTE
network as the prior network, especially when operators possess decent amount of unpaired
spectrum resources. TD-LTE network could be deployed on the most densely populated
areas and LTE FDD network will provide coverage for high-speed data services. UE will be
admitted TD-LTE network in prior for the better user performance when both of the
network are presented, but when users travels to the outage area of TD-LTE networks, they
will be handover to the underlying network of LTE FDD. Due to the asymmetric character
of TD-LTE systems, operators could also define the admittance and balancing strategy based
on the service of subscribers. For instance, the subscribers using symmetric applications, like
voice services, will camp to LTE FDD network, while subscribers using highly asymmetric
data services will be accessed to TD-LTE system in prior.
The convergence network solution of TD-LTE and LTE FDD could be used to construct
a heterogeneous network structure. For instance, TD-LTE systems will focus on the indoor
solutions by implementing TD-LTE femtocells to selected locations and users with specific
requirement, while LTE FDD will be implemented through outdoor macrocells. The solution
could help operators to provide highly customized services and seamless communication
experiences to certain groups. For some of the highly demands locations, deploying TD-LTE
femtocells together with LTE FDD macro cells could also help operators to reduce the
network production cost.

5.3 Cost analysis

Concerning the cost efficiency of TD-LTE system deployment, the reuse and sharing among
technologies would be a promising solution. The reuse and sharing possibilities among
different systems could be explored in several different hierarchies. This could include sharing
infrastructures like site leasing, equipment integration, network management and planning;
and also could be sharing the capacity and coverage of legacy network and intergrade the new
37
system with the existing based on network dimensioning. The reuse could be between TD-
LTE system and the legacy systems, like 2G and 3G networks, or between FDD LTE system
and TD-LTE system when operators want to implement both FDD LTE and TD-LTE
system at same time. It might also be an encouraging solution to implement a Heterogeneous
Network, by positioning TD-LTE femtocell at selected indoor locations, where user demands
are excessive comparing to average. As most mobile traffic are generated indoor, the solution
could offload the traffic pressure of macro site, and presumably eliminate the network cost
for particular user scenarios. Based on the cost calculation of the mentioned scenarios,
corresponding network implementation strategies and cost control mechanism could be
discussed.

5.3.1 Cost analysis for reusing legacy network infrastructure and

capacity
Background and market description
The sample scenario under analysis is considered as a simulated area containing rural, urban
and capital geographic landform. It is based on a medium size European country with
population less than ten million, for instance Sweden or Denmark. The sample area is highly
developed in the telecom industry, with open market and encouraging market competition
among operators. Several operators hold presence nationwide, including an incumbent, and a
Greenfield operator who enters the market during the 3G surges. Comparing to its
predominant counterparts, the Greenfield operators are usually under more pressure in terms
of customer retaining through differentiated service and low cost. Therefore, those operators
with limited resource are considered for the model under analysis. All major 3GPP
specifications have been deployed or under deployment within the region, from GSM, UMTS
to TDD and FDD version of LTE. For simplification reasons, the Greenfield operator has
currently owned a nationwide HSPA network and is judging the possibility of building a TD-
LTE network on top of that, as a complement for the data pressure of 3G networks. Its GSM
service is leased from other operators, therefore not considered for the reuse cost analysis.
The spectrum resources possessed is sufficient for the deployment; therefore it is assumed
that 2 10MHz paired spectrum in 2.1GHz band have been allocated to UMTS/HSPA
deployment, while 20MHz unpaired spectrum in 2.6GHz band is exclusively reserved for TD-
LTE system deployment. As a newly established telecom operator, 20% of the market share is
assumed.

User Demand Assumption

The total population of the country is assumed to be 10 million, out of which 20% is rural
population, 70% is urban population and 10% of people live in the capital area. The
population density is 20 inhabitants per square kilometer for rural area, 1000 inhabitants per
square kilometer for urban area, and 5000 inhabitants per square kilometer for capital area.
Two types of user demands are estimated, including high demands and low demands. The
high demands case usually reflects the trends of date growth and new habits of surfing
seamlessly and wirelessly, for instance stand-alone mobile broadband subscription, in which
case the user tend to use the mobile data service as a complement for the standard fix
broadband service, and monthly data volume per subscription are considered as 8.4 Gbyte. As
for the low demands case, it often reflects the current or near past browsing intensity, for
instance the typical add-on mobile broadband service, which is usually binding together with
38
mobile call service and SMS. Considering the data consumption though mobile devices are
generally lower than the desktop devices, the monthly data volume per subscription is
considered to be 4.2 Gbyte. Since the market share of the operator is 20%, for simplification
reason, the penetration rate of both high and low demands are considered to be 20%. The
data consumptions are distributed during the 8 busy hours per day, therefore the network
capacity need to meet the user demands during the peak hours.

RAT Characteristics (Coverage and Capacity)

The operator has constructed a nationwide HSPA network. For the cell edge user,
considering slow fading and outdoor environment, the maximum propagation loss is assumed
to be 153 dB to provide a minimum data rate as 384kbps. The newly constructed TD-LTE
network configures the DL/UL sub frame ratio to be 2:2 with the total bandwidth to be
20MHz and implement 2 2 MIMO for the downlink. In order to offer cell edge user with
more than 500kbps rate, it is assumed that the maximum propagation loss to be 156 dB in
outdoor environment. Both HSPA and TD-LTE network are usually uplink limited in terms
of cell range, however it is assumed that the maximum propagation loss is the same for both
uplink and downlink.
Applying Cost231-Hata propagation model,
L = 46.3+ 33.9 lg f −13.82 lg hb − a(hm ) + (44.9 − 6.55lg hb )lg d + Cm

a(hm ) = (1.1lg f − 0.7)hm −1.56 lg f + 0.8

In order to apply for the case, a modification is made to Cm ,

"
$ 3dB captial area
$
Cm = # 0dB urban area
$
$ −15dB rural area
%

Where L stands for propagation loss; f stands for the frequency; hb stands for the height
of base station; hm stands for the height of mobile station; d stands for the distance between
base station and mobile station.
Therefore, the cell radius and coverage could be derived in Table 5.1, when three sector
site are implemented for both HSPA and TD-LTE network,
Capital Urban Rural
HSPA Cell Radius 0.58 km 0.97 km 6.9 km
TD-LTE Cell Radius 0.57 km 0.96 km 6.84 km
HSPA Coverage per site 0.86 km2 2.45 km2 123.55 km2
TD-LTE Coverage per site 0.84 km2 2.4 km2 121.38 km2
Table 5.1 Cell range and coverage of HSPA and TD-LTE

Assuming the existing network is implemented with R7 HSPA and carrier aggregation
technologies utilizing 10MHz bandwidth. The newly built TD-LTE network have downlink 2
2 MIMO and UL/DL Configuration 2:2 with special sub frame format 1. According to
industrial results featuring real cell throughputs, it is reasonable to consider 0.74bps/Hz as the
39
spectral efficiency of the legacy HSPA network and 1.46bps/Hz for TD-LTE network. Since
three sectors are implemented for each site, the capacity of each 3G HSPA site is 22.2Mbps.
As for TD-LTE site, the uplink and downlink timeslot in each sub frame is identical, however
DwPTS in the special sub frame could also be used to transfer downlink data and it takes
around 0.65ms in each half frame, therefore the efficient downlink bandwidth for TD-LTE
systems equals to 11.2 MHz, and as a result, the capacity of each TD-LTE site is 49.2Mbps.

Cost Assumptions
Due to the market and economic situation varies from country to country, and fluctuates over
time; a rough estimation is made based on NSN white paper2 and published dissertation by
Klas Johnsson 3 with own assumptions. All the cost assumption is made to reflect the
configuration of Macro HSPA site (2 carrier in band 2.1GHz), and Macro TD-LTE site
(20MHz in band 2.6GHz). The unit is presented in K .
Capital Urban Rural
Radio Equipment (HSPA/TD-LTE) 50 50 50
Upgrade Radio Equipment Cost 30 30 30
Site Cost 60 80 100
Site Lease (Annual) 10 10 8
Power and Backhaul (Annual) 5 5 5
O&M (Annual) 10% of total CAPEX
Table 5.2 Cost structure of HSPA/TD-LTE sites

The Radio Equipment cost is depreciated over 5 years, and site cost is depreciated over 10
years.

Cost Analysis and Deployment Strategies

For the defined case, coverage criterion becomes one of the driving forces of TD-LTE
network implementation in high demand, but HSPA implementation has shown even more
imbalances between coverage and demands, due to its capacity limitation. In low demand
case, the current HSPA network has shown sufficient capacity in most areas based on the
existing implementations, but coverage restrain becomes more conspicuous if TD-LTE is
about to be implemented nationwide. Due to nationwide coverage is a prerequisite for the
network operation, sufficient HSPA sites are deployed in excess of the basic demands of rural
area, there is usually no additional site expansion or upgrade need to be planned in low
demand case, as the existing HSPA network could provide satisfactory network capacity and
user experience. But this situation may change as the user demands are increasing over the
years. Especially when some operators want to use mobile broadband to replace fixed
broadband in far reaching areas, this could mean that the demands of those areas will
experience significant increase apart from their current patterns. Population density is one of
the crucial factors for the network dimensioning, thus on the contrary of rural deployment,
dense urban environment, for instance capital area in the defined case, are usually capacity

2
Mobile broadband with HSPA and LTE – capacity and cost aspects, published in 2010.
3
Cost Effective Deployment Strategies for Heterogeneous Wireless Networks

40
 limited. This is especially true as the mobile network have shown increasing pressure in data,
as customer are expecting more service from their mobile devices. When people flock in big
cities, the existing HSPA network with 10MHz bandwidth could hardly meet the demands. As
a consequence, network expansion is considered, for instance adding new sites or carriers, or
upgrading the network with more advanced technologies. As the intermediate territory, which
covers the vast majority of the nations’ population, urban deployment is dependent with the
demands level in the defined case. By increasing the user demands, the network will be
gradually transferred from coverage limited to capacity limited. Therefore, as a general trend,
future expansion should be considered for those areas.
Based on the given case, there are five deployment strategies with different depth of
reusing existing network and infrastructure, available to be compared in terms of cost and
potential for future expansion.
A. Expand the current network with HSPA technologies for demand increases.
B. Reuse the existing HSPA network for coverage and expand the network with TD-LTE system for
user demands. The newly built LTE network will reuse the HSPA network infrastructure when it is
possible.
C. Reuse the existing HSPA network for coverage and expand the network with TD-LTE system for
user demands. The newly built TD-LTE network is stand-alone.
D. Construct a nationwide LTE level service for both coverage and user demands, and reuse HSPA
network infrastructure.
E. Construct a nationwide stand-alone LTE level service for both coverage and user demands.
The reuse scale of each case could therefore be summarized in Table 5.3.
Reuse HSPA Infrastructure NOT Reuse HSPA Infrastructure
Reuse HSPA Coverage
and Capacity CASE B CASE C
NOT Reuse HSPA
Coverage and Capacity CASE D CASE E
Table 5.3 Reuse scale of Case B-E

The network cost mainly comes from the operational cost, as large quantities of sites are
deployed in rural area and urban area. Since the OPEX is proportional to the site number
maintaining, and it therefore drives up the network cost in total. However, as demands
fluctuate, the network implementation strategies will be adjusted accordingly. Both the
CAPEX and OPEX are very sensitive to the demands level.

41
 Figure 5.1 Annual network production cost

For the entire country, whenever TD-LTE system is applied, high demands case will
usually utilize more network capacity than low demands case and result in lower production
cost per Gigabyte, as in Figure 5.2, therefore if the demands level could be reflected linear
with ARPU, the profit will growth with the demands. While the increasing user demands
could become the bottleneck of HSPA based network, resulting in significant production cost
rise, as in Figure 5.1. Therefore, in order to address the increasing data demands, it is more
cost efficient to implement TD-LTE network instead of network expansion with the HSPA
network.

Figure 5.2 Production cost per GB

Based on the calculation, it is important for the operator to dimension the total user
demands and intergrade the new infrastructure with the existing network to provide a
42
seamless solution, which could deduct the network cost and meet the user demands at the
same time, which is especially true when the network demands stay at a lower level. But there
are two issues with this mechanism. Firstly, the multimode terminal needs to be generally
available in the market. Secondly, as the user demands increases, the expense gaps between
building a TD-LTE network integrating the coverage and capacity of HSPA network, or
building a nationwide TD-LTE service is diminished. However, reusing current HSPA
infrastructure and building the new TD-LTE network upon it as much as possible is more
appealing, comparing to expanding the network with existing HSPA network or stand alone
TD-LTE network, since both CAPEX and OPEX could be significantly deducted. This is
especially true as the user demands exceed the current network capacity, and more HSPA or
TD-LTE sites need to be in place to absorb the increasing demands. According to Figure 5.3,
the site infrastructure reuse is effective for cost reduction for all regions, including capital,
urban and rural areas. Due to the large number of site need to be upgraded in urban areas, the
cost reduction is most significant in those areas. The mechanism of reusing HSPA coverage
and capacity is also effective in cost reduction for most of the situations with one exemption,
i.e. CASE C and CASE E in capital area with high demands. This is because even though the
reuse of HSPA coverage and capacity could reduce the new TD-LTE site number, but the
OPEX of existing HSPA site is still high, making it cheaper to construct a stand alone TD-
LTE system in the area. Therefore, in the high demand case of capital area, the site reusing is
the only effective cost control mechanism. In the low demand case, the urban area and rural
area does not need to expand the capacity of the current networks, therefore it is not efficient
to construct TD-LTE system at the early stage, as in Figure 5.3. The TD-LTE construction in
those areas needs to wait until further growth in user demands.

Figure 5.3 Network implementation cost by areas

Besides, a few more network deployment strategies and cost control mechanisms could be
devised for TD-LTE operators, based on the cost comparison.
1. 2.6G bands TD-LTE system might not the best option for large-scale coverage with
guaranteed cell edge user performance. As for a green field operator with limited
penetration, it is better to choose other solutions for the sake of cost control, i.e.
lower frequency band LTE, like 700M or 900M bands. At an early stage, the cell edge
43
 performance should not be prioritized, and concentric cell solution could be the
ultimate solution, i.e. implementing 700M or 900M bands for coverage, and 2.6G
bands for additional demands.
2. TD-LTE in 2600 and HSPA in 2100 have rather similar coverage range, which means
that most newly build TD-LTE network could find correspond HSPA sites to reuse.
3. According to the data prediction in the cost analysis, the year 2013/14 is a good start
to implement and operated TD-LTE technology to meet the demand requirement for
a developed market, while for a developing market, the TD-LTE system might need
to wait for a few years to be implemented and bring profit, when the data demands
increase to a sufficient level.
4. Most Network expense is derived from operational cost, and the operational cost is
usually approximately proportion to the number of sites maintained. Therefore,
keeping the total number of sites in a reasonable amount is important for a Greenfield
operator. Reusing a site infrastructure and implementing multi layers with different
technologies, for instance HSPA and TD-LTE in case B and case D, is also important
to long-term profitability.
5. The TD-LTE system in the case has the UL/DL configuration 2:2 with special sub
frame format 1. Other configuration could be implemented to adjust the user
demands. But it is a zero-sum between downlink and uplink, therefore the user
behavior need to be analyzed to tune the network accordingly.
6. Network implementation and expansion could start from the high demands or
densely populated area to the low demands and sparse populated area. Firstly, the
dense area is more sensitive to network capacity, as the user demands could increase
fast. Secondly, the site utilization ratio of densely populated area is much higher when
TD-LTE is implemented, which means lower production costs for each subscriber,
and therefore the users could generate more revenue for the operators in those areas.

5.3.2 Cost analysis for HetNet deployment using TD-LTE and

LTE FDD system
Motivation
Introducing femtocell is mainly for the purpose of addressing user demands generated indoor.
According to user behavior, subscribers tend to consume more data traffic in stationary
indoor environment than outdoor moving environment. This situation could deepen, as
operators plan to step into the dedicated mobile broadband service against fixed broadband,
where users have accustomed with the “unlimited” data service and high data rate. This could
bring challenges for the current mobile network infrastructure, where network load are
carried on outdoors macrocells without sufficient dimensioning for indoor users’ quality of
service. Penetration loss brought by concrete wall and floor attenuation could deteriorate user
experience, where excellent network services are most needed, for instance transit center,
residential area, or central business district. Femto solution has however not achieved the
economics of scale in 2G and 3G eras, but as data service became even more vital to operator
business, Femto solution could possibly take off as LTE service are brought to the market.
Even though femtocell brings opportunities of offloading the data pressure from
macrocells, it also brings interferences to the entire network environment. Macrocells are
usually planned and maintained by operators, who constantly check network KPIs and adjust
radio equipment accordingly to mitigate interference and assure user experiences. But
44
femtocells are usually out of the control of operators, as they are large in quantity and
sometimes even owned by other parties. TD-LTE femtocell could relieve the concerns of
operators, since it could operate in spare and completely different frequency band, apart from
LTE FDD systems. Operators could then do not have to split the limited LTE FDD
spectrum resources between macro and femto cells, or implement complex interference
mitigation schemes. According to 3GPP specification, the allocated TD-LTE spectrums are
bands higher than 1800MHz in release 9, while LTE FDD bands could be configured in as
low as 700MHz bands. Therefore LTE FDD system with lower bands, which could resist
more penetration loss, can be used by macro solution as coverage layer. While TD-LTE
system with higher frequency bands, which suffer more propagation loss, could then be used
for Femto solution to reduce its interferences to macro cells and at the same time
complement additional indoor demands. Besides, femtocell usually covers limited users and
areas; it is ideal for service customization and brings additional values. For a small amount of
users, it is much easier to predict the customers’ behaviors and acquire user preferences. TD-
LTE femtocells could then adjust uplink and downlink sub frame ratio to accommodate user
requirement and differentiated customer services. Highly customization could be achieved by
targeting user groups precisely, without affecting services offered to others with different
traffic patterns, even though different user groups might geographically close to one another.

Scenario Description and User Demands

Femtocells, due to its characteristics, are usually implemented in areas where demands are
excessive and propagation loss are higher than average. Therefore, the industry is focused
more on scenarios like home and enterprise solution. The cost analysis is intended to be
carried out in a science park, where 70 000 employees are working in the area, and out of
which 20 000 employees are from ICT sectors. The population density of the science park is
10 000 inhabitants per square meter, and therefore it covers an area of seven square
kilometers.
The mobile operator intends to implement a network for the science park to satisfy
employees with necessary network quality and volume for working purpose. Due to the
nature of work in different sectors, data consumption might vary. It is assumed that the
average data usage for non-ICT employees is 1GB in a regular working day; while for ICT
employees the number goes up to 2GB per day. Distributing the total volume to 8 working
hours per day, the average date rate per user could be 278kbps and 556kbps respectively.

Macro and Femto Network Configurations

The operator plans to exclusively offer the mobile date service within the science park, but
every infrastructure is built from scratch. Two options are to be examined. Macro sites are
configured in LTE FDD technology with 2 2 MIMO for downlink and 20MHz bandwidth
in 900M band. Femtocells are configured in TD-LTE technology with 2 2 MIMO and
20MHz system bandwidth in 2.6G band. Each macro site has three sectors, and in order to
offer the indoor user with sufficient data rate, 25dB penetration loss is considered for the
network dimensioning. According to Cost231-Hata propagation model, it could assume that
the outdoor macro site has a coverage range up to 1.14 km, which means it could cover 3.37
km2 area. The result is quite surprising, meaning 900M LTE have quite good coverage
capacity even for the indoor environment. In order to cover the whole science park, there are
only 3 micro sites are needed.

45
 The spectral efficiency of outdoor macro site in the case could be assumed as 1.57 bps/Hz.
With three sectors per site setup, each LTE macro site has the downlink capacity of 94.2
Mbps. Therefore in order to meet the demands of each business sector, the site number in
Table 5.4 need to be implemented at least,
Non-ICT ICT Total
Macro Site Number for Demands 148 118 266
Table 5.4 Macro site number requirement

Comparing to the coverage requirement, the macro site is highly capacity limited in the
high demands region of the case, even though LTE technology is applied. For TD-LTE
system, femtocells share rather similar spectral efficiency with macro sites. In the defined case,
it is assumed that the typical UL/DL Configuration 2:2 with special sub frame format 1 is
applied, therefore for each femtocell with omni antenna, the capacity could be considered as
16.4 Mbps. The site number needed to meet the capacity demands can then be calculated as
847 for Non-ICT users, 678 for ICT users, and 1525 in total. However, according to real
industry specification, for instance NSN TD-LTE femtocells products, the average user
served simultaneously for each femtocell could only be 8 to 10. As a consequence, the
femtocells deployed need to be further increased, and the traffic for each femtocell is much
lower than its full capacity. The supported coverage for each femtocell is about 100 square
meters according to current industry products, which means coverage of femtocells is not the
bottleneck for femtocell deployment in office environments. In overall, the femtocell number
for the defined case could be assumed as in the Table 5.5,
Non-ICT ICT Total
Femtocell Numbers 6250 2500 8750
Table 5.5 Femtocell number requirement

Cost Structure and Analysis

For macro deployment, based on the previous assumption of 3G Macro sites, it is assumed
that the network cost of the LTE macro site in the defined case follows the cost structure in
the Table 5.6, and all the units are in K .
Radio Equipment (LTE FDD 20MHz) 60
Site Cost 60
Site Lease (Annual) 10
Power, Backhaul and License (Annual) 8
O&M (Annual) 10% of total CAPEX
Table 5.6 Cost structure of Macro LTE sites

For femtocell deployment, the expenses are mainly coming from the one-time cost of
femtocell device, network dimensioning and planning, site installation; and continuous costs
of backhaul access, power, customer support and maintenance. The cost structure are rather
similar to the existing WLAN deployment scenario, however TD-LTE femtocells have not
reached the production scale of WLAN infrastructure, and the production cost is presumably
higher than each WLAN access point in the current state. Therefore, as a rough assumption,
it is assumed that each femtocell has the average CAPEX of 2 000 Euro, and annual OPEX
46
of 800 Euro. It is assumed that both the macro and femto infrastructure are done within the
first year, and the cost analysis are made over the term of five years with the discount rate of
10%.
The cost analysis in Table 5.7 have shown that Femto solution has significant advantages in
terms of total network construction and maintenance cost in the defined case. The CAPEX
of Femto solution is much lower than Macro solution, and it is even more distinct for the
groups with higher traffic requests. For Non ICT sector with lower traffic consumption, the
OPEX of Macro solution is however slightly lower comparing to Femto solution; this might
be caused by the over provisioning of femtocell capacity due to the constrain of concurrent
user number per cell, which waste much spare capacity, especially for the user groups with
limited traffic demands. However as the construction cost of Macro solution are much higher,
making the Femto solution still gains a little edge in general. As ICT sector employees utilize
more femtocells’ capacity, the network deployment scheme has leaned to Femto solution, as
both the CAPEX and OPEX have show excellence over Macro solution.
Network Cost in K CAPEX OPEX NPV

Non ICT Macro 17 760 4 440 36 274.2

Sector Femto 12 500 5 000 33 349.33
ICT Sector Macro 14 160 3 540 28 921.33
Femto 5 000 2 000 13 339.73
Total Macro 31 920 7 980 65 195.53
Femto 17 500 7 000 46 689.06
Table 5.7 Network cost comparison

As in Figure 5.4, due to the relatively high construction cost of macro site in year one, the
deployment scheme might bring more threats to the cash flow of operators if a relatively
larger project is planned. While Femto solution cost is smoother over the years, which means
the network implementation could enjoy a higher pace, due to the balance of operators.

Figure 5.4 Macro and Femto solution production cost from year 1-5

47
 Generally, Femto solution has showed lower production cost per GB, besides increasing
traffic expectation challenges more on the production cost of Macro solution. Due to the
dimensioning of femtocells is based on the supported concurrent connection per cell, the
average production cost per user per month is identical for each business sector, but the
average production cost per user per month for Macro solution will grow with the traffic
expectation. If the same ARPU is assumed, Femto solution will definitely bring more profit
for operators.
Production Cost in Euro Year 1 Production Year 2-5 Average Production
Cost Production Cost Cost
Per GB Per User Per Per User Per GB Per User
Per GB Per Per
Month Month Month
Non ICT Macro 1.77 37 0.36 7.4 0.64 13.32
Sector Femto 1.4 29.17 0.4 8.33 0.6 12.5
ICT Sector Macro 1.77 73.75 0.35 14.75 0.64 26.55
Femto 0.7 29.17 0.2 8.33 0.3 12.5
Total Macro 1.77 47.5 0.35 9.5 0.64 17.1
Femto 1.09 29.17 0.31 8.33 0.47 12.5
Table 5.8 Production cost for different sectors

Based on the cost analysis, operator need to precisely check the traffic pattern of customers,
as it defines the preferable network deployment solution. Femto solution could only be
selected, in case the average traffic volume grows to sufficient level.

TD-LTE Femto Network Deployment Issues and Strategies

Femto solution is promising in cost for the defined case, but discussions could be carried out
additionally for the femtocell network deployment strategies.
1. When femtocells is deployed only for the purpose of complement fixed broadband, it
might not be an issue, but in order to provide a seamless connection to the end users,
multi-mode terminal is required for the indoor TD-LTE and outdoor LTE FDD
scenario. The potential cost of multi-mode terminal products might hinter the
customers from accepting this setup. Therefore worldwide scale could be a
prerequisite of femto solution in terms of both infrastructure and terminal, which is
an issue operaters need to take into consideration before network implementation.
2. A few technical issues which link to the user experiences and cost could be mentioned
for operters to define corresponding deployment strategies.
• Transmission: Backhaul cost consist a larger part of the operational cost of
femtocells. Besides, the transmission quality is also crucial for Femto solutions.
As traditional macro site will usually have its own transmission network,
femtocell could only reuse the existing Internet connection in some cases,
therefore the stability and security issues need to be considered for the
planning.
• Interference: Interference might be induced among femtocells, macrocells and
terminals. Therefore, it is recommended to use different frequency for macro
48
 and femto deployment as in the defined case. But even so, interferences could
still happen among femtocells. The wall and floor loss could address the issue
in most cases. But due to femtocells are sometimes user-maintained, it is hard
to plan and control the status and locations of each cells. Therefore proper
interference mitigation algorithm and mechanics should be planned and
implemented.
• Synchronization: TD-LTE network is strict on time synchronization, which
brings challenges to Femto solutions. GPS signal is usually satisfactory but
requires clear sky view and introduce additional cost, which is not ideal for
Femto deployment. 1588v2 based IP Clock could provide clock sources for
femtocells economically, but transmission network support is a prerequisite.
• Inter-RAT HO: Even though Inter-RAT handover is supported by 3GPP, it
might take a few more time for equipment and terminal vendors to provide a
satisfactory solution with minimum effect on the handover overhead.
However, mobile broadband user could hardly notice the difference if they
could be successfully redirected to outdoors network.
• SON: Femto solution has high requirement for self-organizing. To reduce the
CAPEX and OPEX, self configuration, optimization and management is
essential for large scale deployment. Plug and play need to be supported for
market expansion.
3. The bottleneck of femto solution is the limitation on the concurrent connections or
users. The network cost will have significant drop if a few more users could be
admitted to each cell.
4. The user experience of Femto is even, since there is no particular cell edge user
concept, and it is much easier to provide assured service to the end users, as the signal
strength could be guaranteed.
5. For TD-LTE femto operators, service customization is a promising strategy, which
could potentially increase the profit by attracting more customers. Due to each
femtocell serves limited amount of users, it is easier to adjust the DL/UL
configuration to provide an optimal network access experience according to users’
need. For more advanced features like self and dynamic configuration according to
real time data could also help TD-LTE femto solution gain more favor from end
users.
5. The femtocell production cost might decrease over the year as the solution reaches
sufficient scale, and the operational cost will approach WLAN AP. However, this
need to be further proved by the market in the years ahead.
6. The cost analysis is for downlink, but the uplink could follow the similar pattern as
the downlink and if the user group have shown exceptional uplink preference, Femto
solution could still show superiority due to the over provisioning of its capacity. It is
also worth to mention that the efficient bandwidth of TD-LTE in Femto solution is
half of the FDD bandwidth in Macro solution, since Femto solution is not capacity
limited and could reuse the spectrum over small areas, its spectrum resources have
better utilization rates.

49
 6 Case Study
The case study is carried out for the purpose of verification and comparison. Two MNO are picked, including
Hi3G Denmark from developed markets and China Mobile from emerging markets. As one of the major
momentum for TD-LTE ecosystem, China Mobile will be analyzed with more focuses, while Hi3G
Denmark will be examined on how the general conclusion of TD-LTE system investigation could apply to
Danish market and will also often be discussed more as a reference to China Mobile for comparison purpose.

6.1 China Mobile

China owns three major MNOs including China Mobile, China Unicom and China Telecom.
According to the latest report, Ministry of Industry and Information Technology (MIIT) of
China has announced to allocate all available 190MHz bandwidth in 2.6G bands to TD-LTE.
The analysis is carried out for China Mobile, since China Mobile, as the largest operator in
China and the world, have confirmed to implement the next generation mobile network in
TDD mode. The Chinese market for LTE implementation is yet to be clear, but deem to be
one of the most important one, which could shape the international LTE landscape.
In this section, CMCC’s choice of TD-LTE is discussed based on the discussion in Chapter
3, showing both the determinant and special causes in the selection. China Mobile’s current
development on TD-LTE services and application will also be discussed, together with the
strategies it could follow. Apart from discussion in Chapter 4 about the global trend and
challenges of TD-LTE terminal products, China Mobile’s terminal development and
strategies will be revealed in this section. Finally, the network deployment topics like
convergence with LTE FDD, integration with TD-SCDMA, site sharing and femto solution
of China Mobile will be mentioned.

6.1.1 CMCC’s choice of TD-LTE

When most operators need to choose from different mobile broadband technologies, they
usually investigate the technologies’ potential through issues like spectrum availability, cost of
implementation, network deployment difficulties, development of ecosystem, network
performance and applications, which have been discussed in early chapters. However, China
Mobile’s choice of TD-LTE is much easier, because it is a governmental decision. Due to
intellectual property reasons, China is one of the main pushing hands behind time division
duplex technology since 3G eras. This helps TD-LTE enjoy more government resources and
supports comparing to other technologies. As the telecom regulators, National Development
and Reform Commission (NDRC) and Ministry of Industry and Information Technology
(MIIT) coordinate the TD-LTE development by license distributing, spectrum allocation,
TD-LTE network trial supervision, TD industry and ecosystem supporting, and international
TD technologies promoting.
However, in the commercial perspective, TD-LTE is still a nature choice of China Mobile.
One of the main reasons is the availability of TD-LTE spectrums. According to 3GPP
TS36.101 specification, frequencies around 700M, 1.8G, 2.6G and 3.6G are allocated to TD-
LTE, which are much less than the frequency bands allocated to FDD. But due to the

50
government support from Chinese authority, TD-LTE system in China have been planned
with more spectrum, comparing to its counterparts in Europe, U.S. and Japan. Currently,
there are 190MHz bandwidth in 2.6G bands (2500-2690) allocated to TD-LTE deployment,
out of which 50MHz (2570-2620) have been used for the trial network in several of the major
cities across the country. This allocation aligns with the frequency planning of other major
TD-LTE markets, which helps reduce the research and production cost of equipment and
terminal vendors, stimulate the industry and facilitate future international roaming to other
countries.
Another benefit of China Mobile’s choice of TD-LTE is that TD-LTE could leverage from
other system, in this case TD-SCDMA. As the only major TD-SCDMA operator, China
Mobile have gained much experiences of operating a TDD network, and all the TD-SCDMA
frequencies in 1800M bands that China Mobile operates could be reused to upgrade to TD-
LTE system, according to the data in Appendix B, which could presumably save large amount
of network implementation cost for the corporation, especially in the initial phase.
In the network deployment perspective, TD-LTE is still a proper mobile broadband option
for China Mobile, since it could enable multiple network deployment possibilities in Chinese
telecom market. China Mobile have positioned TD-LTE network as its primary mobile
broadband solution with 2.6G bands and possibly 1.8G bands, both of which have limited
coverage. But China Mobile could implement 700M TD-LTE network as an under layer as a
cost efficient coverage solution, which is possible in Chinese market. However, 700M bands
is currently mainly allocated to national broadcast, which might cause competition between
operators and national broadcast authority, but it is said that the national broadcast would
release the 700M bands by 2015. Unlike other operators, who position 2.6G TD-LTE as a
complement network for data traffic, China Mobile could implement a different solution for
indoor traffic complement, for instance 3.6G TD-LTE.

6.1.2 CMCC’s strategies on TD-LTE services and applications

The traditional competitors of China Mobile are China Unicom and China Telecom, which
would possibly implement LTE FDD system as their solutions for the next generation mobile
broadband services. Therefore, it is important for China Mobile gain an edge against LTE
FDD operators through decent network performances, and innovative services and
applications, aiming to restore users’ confidence in its mobile broadband network. At the
initial phase, China Mobile has been tested TD-LTE network across the nation with a high
profile, in order to gain user attention and prepare for building a substantial market share for
LTE and mobile data service.
China Mobile could focus on differentiating its TD-LTE system as a network with high
availability and capacity after commercial launch. This is made possible due to the abundance
of TD-LTE spectrum resources, since the drawbacks of TD-LTE for its slightly lower
spectral efficiency than LTE FDD, could be compensated with increased bandwidth. Besides,
according to China Mobile’s network statistics, its carried network traffic is about 5:1 for
downlink and uplink 4 , which brings opportunities to increase China Mobile’s network
efficiency, since TD-LTE system is ideal for asymmetric traffic, therefore careful network
tuning according to user behaviors could help China Mobile win more support from
subscribers.

4
Interview with CMCC president XI Guohua at MWC 2013

51
 China Mobile could also explore revenues though particular services and applications,
which matches TD-LTE system’s enhanced network performance and asymmetric characters.
This could help China Mobile tackle the challenges not only from traditional rivals, but also
new competitors like Internet companies. The services and applications could be downlink-
biased, for instance video on demand and streaming services, public broadcast systems,
mobile advertisement, and entertainment system on vehicle; or uplink-biased services like
mobile video surveillance system, home medical monitoring system, on site broadcaster
reporting system, and M2M applications. China Mobile could also provide applications linked
to location-based service (LBS), mobile payment, and integrated voice solution. Cooperation
with content providers of publishing, music or video, whose applications require large data
consumptions, should also be one of China Mobile’s strategies on services and applications.
To provide a unified platform for above mentioned services and applications, China Mobile
has launched its own mobile market and been really successful. For specific scenarios, China
Mobile could also use its TD-LTE network to step into enterprise access network service, and
fixed broadband replacement, which is a traditional business area of China Telecom. The
dynamic adjusting of TD-LTE’s asymmetric based on the changing user pattern, as the
mobile broadband provider of temporary events, could also be a new territory for China
Mobile to explore. This scenario has been successfully demonstrated during the Shanghai
Expo.
According to the annual report of China Mobile Limited, China Mobile has defined its
mobile Internet strategies as ”smart pipe, open platform, featured businesses and friendly
interface”. Therefore, a qualified TD-LTE system is a priority to those strategies, and on top
of that China Mobile need to explore the value of the system by providing users with
customized and demanded services and applications.

6.1.3 CMCC TD-LTE terminal products

At the early stage of TD-SCDMA development, China Mobile had suffered from the lack of
satisfactory terminal products; hence to avoid similar situations happening to TD-LTE
system, China Mobile has put more focuses on their terminal strategies. At 2013 mobile world
congress, China Mobile showcased several TD-LTE terminal products at the initial phase,
including some of the most selling products like Huawei Ascend P2 and Samsung Galaxy S3.
Other TD-LTE compatible smart phones, MiFi, and tables from ZTE and LG have also been
presented. China Mobile believed those products will boost the users’ interesting of choosing
TD-LTE network from the initial phase.
China Mobile has declared its most demanded terminal products should support 5 different
RATs, including GSM, WCDMA, TD-SCDMA, TD and FDD version of LTE, and multiple
frequency bands, especially those most applied home and abroad. However, this is rather
challenging for terminal and chipset vendors, hence China Mobile initiate measures to
stimulate the industry, including releasing procurement contracts of one million terminals in
the year of 2013, accelerating the definition of network deployment timetable and strategies,
and developing terminal products with proprietary brands. China Mobile has also targets
different segments by trying to supply terminal products from high ends to low ends with
various types, including smart phones, tables, laptops, data cards, CPE, and MiFi.
Through trial network in Hangzhou and commercial network in Hong Kong, China
Mobile has tested users’ perception and acceptability of new TD-LTE terminal products. By
providing TD-LTE terminal products to friendly users in Hangzhou, China Mobile has

52
monitored rather active data usages. While in Hong Kong, even though most LTE products
supports only FDD mode, China Mobile decided to launch a serious of dual mode terminal
products supporting both TD-LTE and LTE FDD, by the second half of 2013 in Hong
Kong markets.

6.1.4 CMCC Network deployment progress and strategies

By the year of 2013, China Mobile has constructed TD-LTE trial networks in 13 mainland
cities, aiming to gain network deployment experiences in different scenarios before massive
national level rollout, and stimulate the industry chain for future commercialization.
Therefore, the selected equipment vendors are with various backgrounds, including
international vendors like Ericsson, NSN, Alcatel-Lucent, and domestic competitors like
Huawei, ZTE, and DTmobile. With the experiences of trial networks, China Mobile has
planned to construct around 200 thousands sites covering over 100 cities in the year of 2013,
becoming the largest LTE operator in the world.
As a pioneer of China Mobile’s mobile broadband service, China Mobile Hong Kong has
put its first commercial TD-LTE and LTE FDD convergent network on air, setting a good
example for the convergence solution. The seamless and load balancing solution has attracted
increasing new subscribers joining the data plan. The network provider have positioned TD-
LTE network in 2.3G as the high-speed data layer together 2.6G LTE FDD system, and
FDD 1.8G network are being used for coverage. TD-LTE network have served as a prior
network due to its high capacity and performance, since users will be connected to TD-LTE
network in prior to other networks as long as TD-LTE service is within reach, while in TD-
LTE outrage areas, user will then be directed to LTE FDD networks.
The convergent network solution have been proved as a success in Hong Kong, and it
could also be extended and replicated in the rest of the country, helping China Mobile to
realize the targets of providing seamless communication experiences with large range of
selections for the end users. China Mobile could reduce its network construction cost by
implementing 900M LTE FDD networks for coverage, since it currently operates GSM
network in such bands, possible to be reused, as in Appendix B. However, Chinese GSM
network are still loaded, occupying the spectrum resources for the construction of a FDD
based network to complement TD-LTE systems. Therefore, the practical strategies for China
Mobile now is to prioritize TD-LTE systems to construct a demand-based multiband TD-
LTE network from the beginning, releasing the increasing data pressure in mainland.
Fortunately, the relatively abundance of spectrum resources, satisfactory spectral efficiency of
TD-LTE system and its compatibility with China Mobile’s 3G systems could help China
Mobile to address its user demands smoothly.
China Mobile has defined three major network deployment strategies, including
constructing new TD-LTE sites in 1.8G and 2.6G band or upgrading current TD-SCDMA
network to TD-LTE system with working frequencies in 1.8G bands. For the solution of
rollout new 1.8G TD-LTE sites, site sharing is possible between new TD-LTE network and
legacy network TD-SCDMA, but it also has drawbacks like interferences with TD-SCDMA
and limited spectrum resources. Besides, the most accepted global TD-LTE frequency is 2.6G
bands, meaning that the ecosystem of 1.8G TD-LTE is less mature than 2.6G TD-LTE.
Unlike other TD-LTE operators, as the only major TD-SCDMA, China Mobile have the
possibility to upgrade its TD-SCDMA system to TD-LTE system, enjoying the benefits of
minimized network modification, reduced network construction cost and accelerated network

53
construction pace, which is important for China Mobile’s LTE market share in Chinese
market by providing LTE service ahead of other operators. But it also brings inflexibility to
the network planning and difficulties in the network optimization. Due to time slot
coordination and RRU power sharing, the TD-LTE network performance of upgrade
solution could not be optimized. For 2.6G solutions, the site coverage is less than 1.8G
solutions, but in urban areas, China Mobile has over dimensioned the 3G networks for
capacity reason, due to the relatively high population density in Chinese cities. Hence, 2.6G
sites could find sufficient TD-SCDMA sites to reuse in most urban locations. However, the
major benefits of 2.6G solution is the 190M bandwidth allocated by MIIT for TD-LTE
deployment. Besides, 2.6G bands are far away from the existing 2G and 3G bands as in
Appendix B, which means less interference would be caused. Currently, many international
markets are using 2.6G bands for their TD-LTE network, therefore if China Mobile select
2.6G as its main working band, international roaming could be made convenient and users
might also enjoy larger selection of terminal products due to the development of global
ecosystem. Therefore, in the current stage, 2.6G solution seems gain an edge to other
alternatives, and should be especially focused.
China Mobile should also consider heterogeneous network construction as one of their
network deployment strategies. According to the statistic from CMCC, its mobile data traffic
grows 187.6% during the year 2012 5 , and 90% of data traffic in CMCC’s network are
generated indoors6, meaning that indoor mobile network capacity became more and more
important for China Mobile. Based on the facts, China Mobile could start to explore the
possibility of implementing indoor Femto solution in high frequency bands like 3.6G on
selected locations. In such case, service customization and multi band terminals is key to the
development.

6.2 Denmark Hi3G

Hi3G Denmark is the youngest brand in Danish telecom operator business, other than Hi3G
there are three other competitors operating in Denmark nationwide including Telia, Telenor
and TDC. Concerning their development in mobile broadband, three of them have operated
in LTE by 2012. Telia and TDC offer LTE network in FDD mode, while Hi3G constructs
the network in both FDD and TDD mode. Danish market is highly competitive, and Hi3G as
the latest challenger into the market, has tried to get market share through the breakthrough
on decent data service and affordable price in 3G services. Therefore, mobile broadband have
been defined as one of its core competiveness.
In this section, the general conclusions of research questions from early chapters are being
used to analyze the situation of Hi3G under the framework of Danish telecom market. Firstly,
the reason of why Hi3G should choose TD-LTE is revealed. Then, other market
opportunities about services, applications and terminal development that Hi3G could seize
will be discussed. The network deployment strategies about how to positioning TD-LTE
system in different phases, and how to integrated its LTE system with legacy UMTS network
will also be analyzed.

5
China Mobile Limited Annual Report 2012
6
Interview with State Radio Regulation (SRRC) bureau chief Xie Feibo at 2011 4G WORLD Conference in Beijing.

54
Hi3G’s choice of TD-LTE
Hi3G’s choice of TD-LTE mainly resulted from its positioning as a low cost mobile
broadband operator and spectrum resources limitation. In Appendix C, comparing to other
major operators, Hi3G lacks sufficient paired spectrum resource in LTE service, which might
cause relatively high network deployment costs, limited capacity and poor network
performance. This could be extremely dangerous to a low cost and mobile broadband
centered operator, like Hi3G. In 1.8G bands and 2.6G bands, Hi3G only owns nearly half of
the paired spectrum, comparing to all other operators. Spectrum is crucial for LTE
development, thus if only LTE FDD technology is implemented, Hi3G could hardly
implement a competitive LTE network with the similar capacity and performance as other
major operators. The auction of 800M bands makes the situation worse, besides Hi3G, all
other operators have won a bid successfully in 800M bands, which provide the potential for
them to expand the LTE networks and provide end users with higher capacity and coverage,
especially when 800M bands have better penetration than higher frequency bands. Hence, it is
a nature choice for Hi3G to implement a TD-LTE system with its 25MHz unpaired
spectrum, which could relieve the situation of lacking LTE FDD spectrum resources.

Strategies on services and applications

Due to the limitation of spectrum resources, Hi3G need to precisely target selected customer
segments or promote services and applications based on its network statues to mitigate its
weakness on spectrum. Since Hi3G might not to carry out a nationwide TD-LTE roll out, it
could therefore provide some of the high demands enterprises or residential areas with
exclusively customized TD-LTE mobile broadband service in the outage areas. Hi3G could
also develop featured service packages based on TD-LTE’s asymmetry character, for instance
video surveillance for public sectors, conferencing systems for business sectors, and
guaranteed high-speed Internet service for primary users. Besides, Hi3G could continue to
attract its traditional target groups like young generations by promoting asymmetric services
and applications like video sharing, music streaming, location services and social network,
which have already been popular among the users.

Development on terminals
However, as a TD-LTE and LTE FDD convergent network provider, Hi3G also face the
issues with proper multi-mode terminal products. Hi3G has launched several dongle and MiFi
products based on LTE FDD, but as the network implementations proceed, more multi-
mode products like smart phones and tablets need to be selected and promoted. Hi3G still
need to wait for the further growth of global TD-LTE terminal industry, which could be
boosted by the launch of China Mobile TD-LTE network. As China Mobile will operate TD-
LTE service in 2.6G bands, which is in line with Hi3G, Hi3G could benefit from the scale of
the multi-mode terminals products in these bands.

Network deployment strategies

Based on the statistics from Danish telecom authority Erhvervsstyrelsen, during the second
half of 2012, Hi3G’s mobile broadband network carries 22.6% of the total data traffic with
only 11.2% of the total data subscription7, meaning that Hi3G subscribers consume around
twice of the data traffic than national average. Therefore, Hi3G could construct a convergent

7
Telestatistik: Mobil Baggrundsark andet halvår 2012

55
network based on both TD-LTE and LTE FDD to satisfy its data user’s appetite with
relatively limited spectrum resources. The most possible solution for Hi3G is to position LTE
FDD as its primary mobile broadband network, and TD-LTE system as complement network
to increase network capacity. Hi3G has launched its LTE FDD service at September 2012. It
also constructed and tested its TD-LTE system in Copenhagen area by 2012, but the launch
date of TD-LTE system is not yet clear. At the initial phase, most of Hi3G’s sites are
operated in 1.8M bands for the purpose of coverage and timing of launch. Since the LTE
FDD commercial launch in 15 cities, Hi3G has expanded its LTE network to 69 cities8 by the
first half of 2013. As LTE FDD network reaches most of the Danish population, Hi3G needs
to transfer its network implementation focuses to increasing the network capacity for its LTE
systems, which then is the time to implement the TD-LTE services. This operation could
start from the most densely populated and high demands areas like Copenhagen, Aarhus and
Odense, where most of its data traffic are generated. As a low cost operator, Hi3G could
decide TD-LTE’s presences in other locations, based on the monitoring of user patterns in
those specific regions. As a load-balancing network, TD-LTE could help Hi3G to provide
comparable mobile data service to its data-centered subscribers.
Hi3G could also reuse its UMTS sites infrastructure for cost mitigation. Hi3G has already
operated a nationwide UMTS network in 2.1G bands, whose coverage ability is rather similar
to the LTE network in 2.6G bands, meaning most of its LTE sites could find existing UMTS
sites to reuse. For 1.8G bands LTE sites, the coverage range are even higher, meaning that
reusing existing UMTS sites is also not a big issue for those 1.8G band sites. Beside of
reusing sites, some radio equipment could be reused as well; by introducing radio combiners,
TD-LTE systems and LTE FDD systems could reuse the existing antenna and feeder systems
of UMTS network, which also helps Hi3G bring down the product cost. Its UMTS network
could also be served as the voice solution complement for its convergent TD-LTE and LTE
FDD network through CSFB.

8
Hi3G Demark Publication: Sådan får du 4G/LTE hos 3

56
 7 Conclusion and Future Work
Based on the discussion in the thesis, the research questions raised at the beginning will be answered in this
chapter. The major findings of the project will be listed, aiming to provide a more comprehensive understanding
to the system with an overall view. The possible improvements for the project will be mentioned, providing
directions for future work.

7.1 Research question discussion

Why choose TD-LTE, comparing to other mobile access technologies?
There are good reasons to choose TD-LTE, but also good reasons to abandon it. TD-LTE
operators could choose it for its economic and abundant spectrum resources, comparable
coverage, asymmetric characters that could accord with the current traffic patterns and
potentially enable new services and applications, channel reciprocity, burgeoning ecosystem,
varied network deployment alternatives, and good potentials for future expansion. However,
TD-LTE also has the limitations like lack of spectrum in low bands, complex synchronization
solution, lag at terminal development, and lower spectral efficiency than LTE FDD.
Since there are long lists for both of the pros and cons of TD-LTE, it is more important
for operators to analyze whether the TD-LTE’ advantages are truly beneficial for themselves
under their own situations. According to the case study, among all factors, spectrum
availability seems to be rather conclusive for operators’ choice of TD-LTE, especially when
the operator could obtain abundant paired spectrum resources with reasonable price.
Is the user experience of TD-LTE different from other systems in terms of standard services or applications?
What particular services or application can be facilitated by TD-LTE system?
Theoretical calculation, lab test and field test results have shown that TD-LTE has
comparable peak throughput with LTE FDD, but with longer latency. However due the
improved system design, the peak throughput and capacity of TD-LTE have both been
largely improved from its predecessors. For the standard mobile broadband services and
applications, the user experience of TD-LTE is comparable with LTE FDD, beside of the
latency sensitive real time applications. But comparing to other 3G technologies, the user
experiences of TD-LTE have been significantly improved in all aspects.
With the characters of asymmetry and high capacity, TD-LTE could enable varied
particular services and applications. With downlink heavy configurations, TD-LTE could
improve the user experiences of consumer type services and applications like video on
demand, media streaming, and social networks; it could also enable dedicated downlink biased
services like public broadcasting, mobile advertisement and entertainment system on vehicles.
With uplink heave configurations, TD-LTE could be used to explore the markets of specific
sectors with service and applications like video surveillance system, home medical monitoring
and reporting systems, or on site broadcaster reporting system. The cost efficiency of TD-
LTE uplink transmission resources could also enable M2M applications. Besides, TD-LTE
could be used in other specific scenarios, like fixed broadband replacement services for
enterprise and residential customers, or backhaul complement applications. TD-LTE is highly

57
capable for such services and applications, due to the advantages like high capacity, cost
efficiency and less interference. The flexibility of uplink to downlink capacity adjusting of
TD-LTE facilitates the mobile broadband services in specific occasions or locations, where
traffic patterns are constantly changing, like special events or tourist resorts.
What is the market prospect for TD-LTE enabled chipsets or terminals?
According to the market investigation of TD-LTE chipsets and terminals, the ecosystem is
wholesome and inspiring in general. In the chipset development, reasonable amount of
vendors including some market leaders have been involved in the development of multi mode
multi band chipsets for both handsets and routers. As a positive trend, TD-LTE chipset
development has also been merged with LTE FDD under the common platform in
Qualcomm, gaining the full leverage of more mature LTE FDD development. The market
prospect of terminals seems bright, as TD-LTE has been widely supported by global vendors.
The TD-LTE products category has covered phone, tablet, dongle, router, CPE, M2M and
module, however the smart phone product is still the bottleneck of TD-LTE terminal
development at the current stage. As the TD-LTE network construction of Japan, India and
China continues, the ecosystem of TD-LTE terminal would experience further growth.
How to integrate TD-LTE with self-owned FDD LTE, 3G or 2G systems?
The integration possibility of TD-LTE with legacy networks like GSM or UMTS could be
explored through voice solutions like CSFB, where GSM and UMTS could be used to
provide the complement voice service for TD-LTE users. TD-LTE and legacy networks
could also be integrated to provide a ubiquitous communication solution, serving to each
other as coverage and data capacity complements. However, the mobility management among
systems, and multi mode terminals are important for the adoption of this solution. The
implementation of TD-LTE could also reuse the site location of legacy systems for cost
efficiency, and a typical example is the site sharing between 2.1G UMTS and 2.6G TD-LTE
as their coverage range is comparable. In the perspective of service differentiation, the data
service could be provided with different hierarchies based on the different network
performance of TD-LTE and legacy networks.
TD-LTE could also be used to construct a convergent network with LTE FDD, both of
which are date centric systems. Due to the similarity in the system design, the mobility
management and roaming are fully supported but with additional latency. The lower band
could be used for coverage, and high band could be used for capacity enhancement,
regardless of TDD or FDD versions. The network element of LTE FDD and TD-LTE could
be deeply shared, from site location, backhaul, and antenna to radio equipment and core
infrastructures. Based on the network strategies of specific operators, TD-LTE could be
positioned as either primary or complement network through admittance strategies and
coverage areas. The convergence of TD-LTE femto and LTE FDD macro is also a potential
structure for heterogeneous network.
What deployment strategies and cost control mechanisms could be applied for TD-LTE implementation?
Site Infrastructure Sharing: Reuse the site infrastructure of legacy network for TD-LTE system
implementation has been proved as an effective cost control mechanism, regardless user
demands or geographic areas. The network CAPEX, OPEX and production cost per GB will
all drop for some degrees compared to stand alone solutions. The benefits of site
infrastructure sharing increase with data demands.

58
Reuse Coverage and Capacity of Legacy Network: To reuse the coverage and capacity of legacy
network like UMTS is cost efficient, but not always. However for a Greenfield operator, who
have limited market share, it still might be a good strategy. The deployment of TD-LTE
network should usually start with areas with higher population density and user demands, as
the break point of a more cost efficient full replacement of legacy network with TD-LTE in
data service happens sooner in those areas. For areas with less traffic volume, the
implementation of TD-LTE should follow a slower pace, as the reuse strategy of coverage
and capacity of legacy network might be still more cost efficient in those areas. Concerning
the effect of user demands on this strategy, the benefits of reusing legacy network’s coverage
and capacity will diminish with the increasing of demands. Another prerequisite for this
strategy is that the multi mode terminals products are generally available among users.
Femto solution: The solution of TD-LTE femto could be cost efficient for concentrated and
indoor scenarios than LTE FDD macro solution. The CAPEX of TD-LTE Femto solution is
usually lower than Macro solution, but OPEX in the other way might surpass the Macro
solution for sectors with low traffic demands. Therefore, the investment of TD-LTE Femto
solution is smoother than Macro solution over the years. In such concentrated and indoor
environment like science parks, transition centers or central business districts, TD-LTE
Femto solution generally enjoys lower production cost per GB and per user. Due to the high
capacity of Femto solution, the cost benefits will increase with user demands.

7.2 Conclusions
To provide a better picture of TD-LTE to operators, the thesis tries to discuss the system
though the perspectives of both market and network implementations. The market analysis
helps discover the profit potentials and growth prospect, while the network implementation
strategies answer the questions of how to implement a system with decent performance and
reasonable cost.
TD-LTE might not be as widely adopted as LTE FDD, but for TD-LTE operators, the
most efficient marketing of TD-LTE is differentiation strategy. In this sense, TD-LTE could
be used to explore the market of innovative services and applications for varied new sectors
and scenarios, in addition to the mainstream consumer mobile broadband service. Based on
the characters of high capacity, cost efficiency and flexible asymmetry, TD-LTE operator
could tailor the mobile broadband service according to sector and scenario specific traffic
patterns. However, even though TD-LTE could facilitate customized mobile broadband
services and applications, market cultivation is still an important challenge for TD-LTE
operators. Fortunately, a relief for TD-LTE operators is that the development of TD-LTE
terminal ecosystem seems wholesome, and multi mode multi band products would dominate
the market in the long run. From the analysis, TD-LTE smart phones would be a key market,
which still need additional focuses and further development.
In the strategies of network implementation, there are many ways to explore the benefits
of TD-LTE system, but for TD-LTE operators, system positioning is of great importance. A
common solution is to construct a convergent network with both TD-LTE and LTE FDD,
and TD-LTE system would usually be positioned as the data layer. However, stand-alone TD-
LTE network is still a feasible solution, especially for those operators with abundant unpaired
spectrum.

59
 Different cost control mechanisms would usually have their own conditions and
restrictions. For national rollout, site infrastructure sharing between TD-LTE and UMTS is
promising, especially when the traffic demands increase. However, reusing the coverage and
capacity of UMTS network is only cost efficient during the initial deployment of TD-LTE,
and as the demand increase, a whole network replacement seems to be the ultimate solution.
Another strategy is that TD-LTE deployment should start from the high demands areas, and
the national rollout pace should adjust according to the traffic demands prediction. For
compact and indoor scenarios, TD-LTE femto could be a cost efficient solution for mobile
broadband services. However, the bottleneck of the Femto solution is the limitation on the
number of concurrent connections for each cell. The network production cost of Femto
solution could be further reduced as the scale is reached.

7.3 Discussion and future work

From the analysis, the availability of unpaired spectrum resource plays an important part for
the adoption of TD-LTE system. Therefore, a further analysis could be done about the
possibilities and difficulties of getting sufficient spectrum resource for the deployment. Beside
of the analysis of licensed spectrum resource, the technological and financial analysis of using
cognitive radio to utilize spectrum hole for TD-LTE deployment could be investigated.
In the services and applications part, the possible new market opportunities of TD-LTE
system have been discussed. However, the market volume and prospect of such services and
applications could be further discussed. The market strategies about how to cultivate the
markets and obtain user acceptance could also be interesting topics to be further analyzed.
Beside, quantitative cost efficiency calculations of the mentioned service and application
models could be conducted to improve the feasibility study.
The cost modeling of TD-LTE deployment is simplified. The actual deployment
possibilities are more than the case could describe. Besides, the cost calculation is mainly
conducted for the radio access network, thus additional implementation costs could be
included for the modeling of a complete TD-LTE system, for instance the cost of core
infrastructure and cost of spectrum resource. As the convergent network solution between
TD-LTE and LTE FDD becomes more recognized, the cost analysis could also be done for
such solution, showing operators whether it is more cost efficient than the stand-alone TD-
LTE or LTE FDD solution.
The thesis could be further enhanced in many ways, as the TD-LTE community continues
to grow, even more issues and topics would presumably emerge.

60
References
[1] Cisco, “Cisco Visual Networking Index: Global Mobile Data Traffic Forecast Update,
2012–2017,” Feb. 2013.
[2] GTI Secretariat, “TD-LTE Industry Briefing.” 15-Apr-2013.
[3] B. Ayvazian, “LTE TDD Operator Business Case & Adoption Forecast,” Heavy Reading
White Paper, 2011.
[4] R. Susitaival, H. Wiemann, J. Östergaard, and A. Larmo, “Internet Access Performance in
LTE TDD,” in Vehicular Technology Conference (VTC 2010-Spring), 2010 IEEE 71st, 2010,
pp. 1–5.
[5] D. Astely, E. Dahlman, A. Furuskar, and S. Parkvall, “TD-LTE The radio-access solution
for IMT-Advanced/TDD,” in 2010 5th International ICST Conference on Communications and
Networking in China (CHINACOM), 2010, pp. 1–5.
[6] M. S. K. Manikandan, E. Shanthi, and S. J. Thiruvengadam, “Network Performance
Analysis of TCP-Based 3GPP LTE Time Division Duplex Systems,” Wireless Pers
Commun, vol. 65, no. 4, pp. 863–873, Aug. 2012.
[7] H. Jinling, “TD-SCDMA/TD-LTE evolution - Go Green,” in 2010 IEEE International
Conference on Communication Systems (ICCS), 2010, pp. 301–305.
[8] J. Markendahl, “Business innovation strategies to reduce the revenue gap for wireless
broadband services,” Communications & Strategies, no. 75, pp. 35–55, 2009.
[9] J. Werding, “Revenue requirements for mobile operators with ultra-high mobile
broadband data traffic growth,” 2010.
[10] V. Krizanovic, “Mobile broadband access networks planning and evaluation using
techno-economic criteria,” in Information Technology Interfaces (ITI), Proceedings of the ITI 2012
34th International Conference on, 20120600, pp. 281–286.
[11] Goldman Sachs, “TD-LTE: gearing up to cover 2.7bn people in Asia by 2013.” 30-Jun-
2011.
[12] Nokia Siemens Networks, “TD-LTE whitepaper.” 201011.
[13] Motorola, “Motorola LTE Solutions: The TD-LTE Advantage.” 2010.
[14] Angel Ivanov, “TD-LTE and FDD-LTE A Basic Comparison.” Ascom, 12-Jan-2012.
[15] Motorola, “TD‑LTE: Enabling New Possibilities and Revenues for Operators.” 2010.
[16] H. Chesbrough and R. S. Rosenbloom, “The role of the business model in capturing
value from innovation: evidence from Xerox Corporation’s technology spin‐off
companies,” ICC, vol. 11, no. 3, pp. 529–555, Jun. 2002.
[17] J. Markendahl, “Mobile Network Operators and Cooperation: A Tele-Economic Study of
Infrastructure sharing and Mobile Payment Services,” KTH, 2011.
[18] “3GPP TR 25.913 V9.0.0 Requirements for Evolved UTRA (E-UTRA) and Evolved
UTRAN (E-UTRAN) (Release 9),” Dec. 2009.

61
[19] J. Wannstrom, “LTE-Advanced,” May 2012.
[20] “3GPP TS 36.211 V9.1.0 Physical Channels and Modulation (Release 9),” Mar. 2010.
[21] Motorola, “TD-LTE: Exciting Alternative, Global Momentum,” 2010.
[22] TD Industry Alliance, “TD-LTE Industry Development Report (2012),” Jan. 2013.
[23] TD Industry Alliance, “TDD Global Spectrum Report (2012),” Jan. 2013.
[24] E. Dahlman, S. Parkvall, and J. Skold, 4G: LTE/LTE-Advanced for Mobile Broadband:
LTE/LTE-Advanced for Mobile Broadband. Academic Press, 2011.
[25] R. Julius, “Update from the LTE/SAE Trial Initiative,” presented at the Informa LTE
World Summit, London, 19-Nov-2008.
[26] R. Julius, “Latest Results from the LSTI, Feb 2009,” 25-Feb-2009.
[27] Global TD-LTE Initiative, “TD-LTE Industry and Market development,” Dec. 2012.

62
Appendix
Appendix A: Cost analysis results for reusing legacy network infrastructure and
capacity (Case A to Case E)

CASE A Capital Urban Rural Total

High CAPEX (Total) in K 51 480 265 590 88 800 405 870
Deman
CAPEX (Annual) in K 7 488 36 774 11 840 56 102
ds
OPEX (Annual) in K 18 226 137 340 39 256 194 822

Production Cost per GB in 1.28 1.23 1.27 1.24

Annual Network Cost in M 25.71 174.11 51.1 250.92

Low CAPEX (Total) in K 12 980 0 0 12 980

Deman
CAPEX (Annual) in K 1 888 0 0 1 888
ds
OPEX (Annual) in K 9 126 80 136 22 680 111 942

Production Cost per GB in 1.09 1.14 1.13 1.13

Annual Network Cost in M 11.01 80.14 22.68 113.83

Table A.1 Cost analysis of CASE A

CASE B Capital Urban Rural Total

High CAPEX (Total) in K 6 360 27 660 8 010 42 030
Deman
CAPEX (Annual) in K 1 272 5 532 1 602 8 406
ds
OPEX (Annual) in K 6 694 82 902 23 481 113 077

Production Cost per GB in 0.4 0.63 0.62 0.6

Annual Network Cost in M 7.97 88.43 25.08 121.48

Low CAPEX (Total) in K 1 590 0 0 1 590

Deman
CAPEX (Annual) in K 318 0 0 318
ds
OPEX (Annual) in K 6 217 80 136 22 680 109 033

Production Cost per GB in 0.65 1.14 1.13 1.08

Annual Network Cost in M 6.54 80.14 22.68 109.35

Table A.2 Cost analysis of CASE B

63
CASE C Capital Urban Rural Total
High CAPEX (Total) in K 23 320 119 860 40 050 183 230
Deman
CAPEX (Annual) in K 3 392 16 596 5 340 25 328
ds
OPEX (Annual) in K 11 570 105 952 30 156 147 678

Production Cost per GB in 0.74 0.87 0.88 0.86

Annual Network Cost in M 14.96 122.55 35.5 173

Low CAPEX (Total) in K 5 830 0 0 5 830

Deman
CAPEX (Annual) in K 848 0 0 848
ds
OPEX (Annual) in K 7 436 80 136 22 680 110 252

Production Cost per GB in 0.82 1.14 1.13 1.1

Annual Network Cost in M 8.28 80.14 22.68 111.1

Table A.3 Cost analysis of CASE C

CASE D Capital Urban Rural Total

High CAPEX (Total) in K 16 230 92 490 26 400 135 120
Deman
CAPEX (Annual) in K 2 742 18 090 5 140 25 972
ds
OPEX (Annual) in K 8 941 90 150 25 502 124 593

Production Cost per GB in 0.58 0.77 0.76 0.75

Annual Network Cost in M 11.68 108.24 30.64 150.57

Low CAPEX (Total) in K 7 430 92 490 26 400 126 320

Deman
CAPEX (Annual) in K 1 462 18 090 5 140 24 692
ds
OPEX (Annual) in K 6 861 90 150 25 502 122 513

Production Cost per GB in 0.83 1.53 1.52 1.46

Annual Network Cost in M 8.32 108.24 30.64 147.21

Table A.4 Cost analysis of CASE D

64
CASE E Capital Urban Rural Total
High CAPEX (Total) in K 34 870 378 690 123 600 537 160
Deman
CAPEX (Annual) in K 5 072 52 434 16 480 73 986
ds
OPEX (Annual) in K 8 242 81 564 23 072 112 878

Production Cost per GB in 0.66 0.95 0.98 0.93

Annual Network Cost in M 13.31 134 39.55 186.86

Low CAPEX (Total) in K 26 070 378 690 123 600 528 360
Deman
CAPEX (Annual) in K 3 792 52 434 16 480 72 706
ds
OPEX (Annual) in K 6 162 81 564 23 072 110 798

Production Cost per GB in 0.99 1.9 1.96 1.82

Annual Network Cost in M 9.95 134 39.55 183.5

Table A.5 Cost analysis of CASE E

65
Appendix B: Spectrum Resources of Major Chinese Operators 9

2G 3G LTE
China Mobile GSM 900: TD-SCDMA: Not Available now, but
TD-LTE is tested by
2×9MHz Paired Spectrum 20MHz Unpaired
China Mobile in several
(890-909/935-954) Spectrum (1880-1900)
of the major cities across
EGSM 900: 15MHz Unpaired the nation.
Spectrum (2010-2025)
2×5MHz Paired Spectrum
(885-890/930-935)
GSM 1800:

2×25MHz Paired Spectrum

(1710-1725/1805-1820)

China Unicom GSM 900: WCDMA: Not Available

2×6 MHz Paired Spectrum" 2×15 MHz Paired
(909-915/954-960)" Spectrum"
GSM 1800: (1940-1955/2130-2145)
2×10 MHz Paired Spectrum
(1745-1755/1840 1850)
China Telecom 2G not available CDMA: Not Available

2×15MHz Paired
Spectrum (825-840/870-
885)

2×15MHz Paired
Spectrum (1920-
1935/2110-2125)

9
Resources from Chinese spectrum resource allocation table and State Radio Regulation of China website, until mid 2013

66
Appendix C: Spectrum Resources of Major Danish Operators 10

2G GSM 900 3G WCDMA LTE 1800 LTE 2600 LTE 800

2100
Hi3G Denmark Not Available 2×15MHz 2×10MHz 2×10MHz Paired Not
Paired Paired Spectrum (2520- Available
Spectrum Spectrum 2530/2640-2650)
(1920- (1710.1-
25MHz Unpaired
1935/2110- 1720.1/1805.1
Spectrum (2570-
2125) -1815.1)
2595)
Telia/Telenor Telia 2×11.8MHz 2×15MHz 2×23.6 MHz 2×20 MHz Paired Not
joint venture: Paired Paired Paired Spectrum (2530- Available
Spectrum Spectrum Spectrum 2550
TT-Netvaerket 2×10 MHz
(880.1-
(1950-1965/ (1761.3- 2650-2670) Paired
891.9/925.1-
2140-2155) 1784.9/1856.3 Spectrum
936.9)
-1879.9) (791-801
MHz and
Telenor 2×9MHz 2×15MHz Not Available Not Available
832-842
Paired Paired
MHz)
Spectrum Spectrum
(905.9- 2×19.4 MHz 2×20 MHz Paired
(1965-
914.9/950.9- Paired Spectrum (2550-
1980/2155-
959.9) Spectrum 2570/2670-2690)
2170)
(1741.9-
1761.3/1836.9
-1856.3)

TDC 2×9MHz 2×15MHz Not Available 2×20 MHz Paired Not

Paired Paired Spectrum (2500- Available
Spectrum Spectrum 2520/2620-2640)
(896.9- (1935- 2×21.8
905.9/941.9- 1950/2125- 2×20 MHz
MHz Paired
950.9) 2140) Paired
Spectrum
Spectrum
(1720.1- (801-821
1741.9/1815.1 MHz and
-1836.9) 842-862
MHz)

10
Resources from GSMA 900 MHz and 1800 MHz band refarming case study Denmark and Danish Business Authority website, until mid
2013
67
 TRITA-ICT-EX-2014:7

www.kth.se

Powered by TCPDF (www.tcpdf.org)