HSPA+ Evolution to Release 12: Performance and Optimization

Foreword

Our industry has undergone massive change in the last few years and it feels like only yesterday that T-Mobile USA, and the industry as a whole, was offering voice centric services and struggling to decide what to do about data. Once the smartphone revolution started and consumption of new services literally exploded, wireless operators had to quickly overcome many new challenges. This new era of growth in the wireless industry continues to create great opportunity coupled with many new challenges for the wireless operators.

In my role of Chief Technical Officer at T-Mobile USA, I have enjoyed leading our company through a profound technology transformation over a very short period of time. We have rapidly evolved our network from a focus on voice and text services, to providing support to a customer base with over 70% smartphones and a volume of carried data that is growing over 100% year on year. At the time of writing, we offer one of the best wireless data experiences in the USA which is one of the fastest growing wireless data markets in the world. We provide these services through a combination of both HSPA+, and more recently, LTE technologies.

Many wireless operators today have to quickly address how they evolve their network. And with consumer demand for data services skyrocketing, the decision and choice of technology path are critical. Some operators may be tempted to cease investments in their legacy HSPA networks and jump straight to LTE. In our case, we bet heavily on the HSPA+ technology first, and this has proved instrumental to our success and subsequent rollout of LTE. As you will discover throughout this book, there are many common elements and similarities between HSPA+ and LTE, and the investment of both time and money into HSPA will more than pay off when upgrading to LTE. It certainly did for us.

Furthermore, industry forecasts indicate that by the end of the decade HSPA+ will replace GSM as the main reference global wireless technology and HSPA+ will be used extensively to support global voice and data roaming. This broad-based growth of HSPA+ will see continued development of existing economies of scale for both infrastructure and devices.

This book provides a great combination of theoretical principles, device design, and practical aspects derived from field experience, which will help not only tune and grow your HSPA+ network, but also LTE when the time comes. The book has been written by 26 experts from multiple companies around the world, including network infrastructure and chip set vendors, mobile operators, and consultancy companies.

As worldwide renowned experts, Harri and Antti bring a wealth of knowledge to explain all the details of the technology. They have helped create and develop the UMTS and HSPA technologies through their work at NSN, and more recently pushed the boundaries of HSPA+ from Release 8 onward.

Pablo has been a key player to our HSPA+ success at T-Mobile, working at many levels to drive this technology forward: from business case analysis and standardization, to leading technology trials through design and optimization activities. His practical experience provides a compelling perspective on this technology and is a great complement to the theoretical aspects explored in this book.

I hope you will find this book as enjoyable as I have, and trust that it will help advance your understanding of the great potential of this key and developing technology.

Neville Ray

CTO

T-Mobile USA

Preface

HSPA turned out to be a revolutionary technology, through making high-speed wide-area data connections possible. HSPA is by far the most global mobile broadband technology and is deployed by over 500 operators. Data volumes today are substantially higher than voice volumes and mobile networks have turned from being voice dominated to data dominated. The fast increase in data traffic and customer expectation for higher data rates require further evolution of HSPA technology. This book explains HSPA evolution, also called HSPA+. The book is structured as follows. Chapter 1 presents an introduction. Chapter 2 describes the basic HSDPA and HSUPA solution. Chapter 3 presents multicarrier and multiantenna evolution for higher efficiency and higher data rates. Chapter 4 explains continuous packet connectivity and high speed common channels. Multiflow functionality is described in Chapter 5, voice evolution in Chapter 6, and heterogeneous networks in Chapter 7. Advanced receiver algorithms are discussed in Chapter 8. ITU performance requirements for IMT-Advanced and HSPA+ simulation results are compared in Chapter 9. Chapters 10 to 13 present the practical network deployment and optimization: HSPA+ field measurements in Chapter 10, network planning in Chapter 11, network optimization in Chapter 12, and smartphone optimization in Chapter 13. Terminal design aspects are presented in Chapter 14. The inter-working between LTE and HSPA is discussed in Chapter 15 and finally the outlook for further HSPA evolution in Chapter 16. The content of the book is summarized here in Figure P.1.

Figure P.1 Contents of the book.

Acknowledgments

The editors would like to acknowledge the hard work of the contributors from Nokia, from T-Mobile USA, from Videotron Canada, from Teliasonera, from Renesas Mobile and from Signals Research Group: Mika Aalto, Luigi Dicapua, Ryszard Dokuczal, Mikael Guenais, Timo Halonen, Matthias Hesse, Thomas Höhne, Maciej Januszewski, Jean-Marc Lemenager, Thierry Meslet, Laurent Noel, Brian Olsen, Hisashi Onozawa, Hannu Raassina, Karri Ranta-aho, Jussi Reunanen, Fernando Sanchez Moya, Alexander Sayenko, Jeff Smith, Mike Thelander, Jeroen Wigard, Victor Wilkerson, and Carl Williams.

We also would like to thank the following colleagues for their valuable comments: Erkka Ala-Tauriala, Amar Algungdi, Amaanat Ali, Vincent Belaiche, Grant Castle, Costel Dragomir, Karol Drazynski, Magdalena Duniewicz, Mika Forssell, Amitava Ghosh, Jukka Hongisto, Jie Hui, Shane Jordan, Mika Laasonen, M. Franck Laigle, Brandon Le, Henrik Liljeström, Mark McDiarmid, Peter Merz, Randy Meyerson, Harinder Nehra, Jouni Parviainen, Krystian Pawlak, Marco Principato, Declan Quinn, Claudio Rosa, Marcin Rybakowski, David A. Sánchez-Hernández, Shubhankar Saha, Yannick Sauvat, Mikko Simanainen, Dario Tonesi, Mika Vuori, Dan Wellington, Changbo Wen, and Taylor Wolfe.

The editors appreciate the fast and smooth editing process provided by Wiley and especially by Sandra Grayson, Liz Wingett, and Mark Hammond.

We are grateful to our families, as well as the families of all the authors, for their patience during the late night writing and weekend editing sessions.

The editors and authors welcome any comments and suggestions for improvements or changes that could be implemented in forthcoming editions of this book. The feedback is welcome to the editors' email addresses [email protected], [email protected], and [email protected].

Abbreviations

3GPP

Third Generation Partnership Project

ACK

Acknowledgment

ACL

Antenna Center Line

ACLR

Adjacent Channel Leakage Ratio

ACP

Automatic Cell Planning

ADC

Analog Digital Conversion

AICH

Acquisition Indicator Channel

AL

Absorption Losses

ALCAP

Access Link Control Application Part

AM

Acknowledged Mode

AMR

Adaptive Multirate

ANDSF

Access Network Discovery and Selection Function

ANQP

Access Network Query Protocol

ANR

Automatic Neighbor Relations

APE

Application Engine

APNS

Apple Push Notification Service

APT

Average Power Tracking

aSRVCC

Alerting SRVCC

AWS

Advanced Wireless Services

BBIC

Baseband Integrated Circuit

BCH

Broadcast Channel

BH

Busy Hour

BiCMOS

Bipolar CMOS

BLER

Block Error Rate

BOM

Bill of Material

BPF

Band Pass Filter

BSC

Base Station Controller

BT

Bluetooth

BTS

Base Station

CA

Carrier Aggregation

CAPEX

Capital Expenses

CCCH

Common Control Channel

CDMA

Code Division Multiple Access

CIO

Cell Individual Offset

CL

Closed Loop

CL-BFTD

Closed Loop Beamforming Transmit Diversity

CM

Configuration Management

CMOS

Complementary Metal Oxide Semiconductor

CoMP

Cooperative Multipoint

CPC

Continuous Packet Connectivity

C-PICH

Common Pilot Channel

CPU

Central Processing Unit

CQI

Channel Quality Information

CRC

Cyclic Redundancy Check

CS

Circuit-Switched

CSFB

CS Fallback

CSG

Closed Subscriber Group

CTIA

Cellular Telecommunications and Internet Association

DAC

Digital Analog Conversion

DAS

Distributed Antenna System

DASH

Dynamic Adaptive Streaming over HTTP

DC

Direct Current

DC

Dual-Carrier

DCA

Direct Conversion Architecture

DCH

Dedicated Channel

DC-HSDPA

Dual-Cell HSDPA

DC-HSPA

Dual-Cell HSPA

DDR

Double Data Rate

DF

Dual Frequency

DFCA

Dynamic Frequency and Channel Allocation

DL

Downlink

DM

Device Management

DMIPS

Dhrystone Mega Instructions Per Second

DPCCH

Dedicated Physical Control Channel

DRAM

Dynamic Random Access Memory

DRX

Discontinuous Reception

DS

Deep Sleep

DSL

Digital Subscriber Line

DSP

Digital Signal Processing

DTX

Discontinuous Transmission

E-AGCH

Enhanced Absolute Grant Channel

EAI

Extended Acquisition Indicator

Ec/No

Energy per Chip over Interference and Noise

E-DCH

Enhanced Dedicated Channel

EDGE

Enhanced Data rates for GSM Evolution

E-DPCH

Enhanced Dedicated Physical Channel

eF-DPCH

Enhanced Fractional Dedicated Physical Channel

EGPRS Enhanced GPRS E-HICH

E-DCH Hybrid ARQ Indicator Channel

eICIC

Enhanced Inter-Cell Interference Cancellation

EIRP

Equivalent Isotropical Radiated Power

EMI

Electro-Magnetic Interference

EPA

Extended Pedestrian A

EPC

Evolved Packet Core

EPS

Evolved Packet System

E-RGCH

Enhanced Relative Grant Channel

eSRVCC

Enhanced SRVCC

ET

Envelope Tracking

EU

European Union

E-UTRA

Enhanced Universal Terrestrial Radio Access

EVM

Error Vector Magnitude

FACH

Forward Access Channel

FBI

Feedback Information

FDD

Frequency Division Duplex

F-DPCH

Fractional Dedicated Physical Channel

FE

Front End

FE-FACH

Further Enhanced Forward Access Channel

feICIC

Further Enhanced Inter-Cell Interference Coordination

FEM

Front End Module

FGA

Fast Gain Acquisition

FIR

Finite Impulse Response

FM

Frequency Modulation

FOM

Figure of Merit

FS

Free Space

FTP

File Transfer Protocol

GAS

Generic Advertisement Service

GCM

Google Cloud Messaging

GGSN

Gateway GPRS Support Node

GoS

Grade of Service

GPEH

General Performance Event Handling

GPRS

General Packet Radio Service

GPS

Global Positioning System

GPU

Graphical Processing Unit

GS

Gain Switching

GSM

Global System for Mobile Communications

GSMA

GSM Association

GTP

GPRS Tunneling Protocol

HARQ

Hybrid Automatic Repeat-reQuest

HD

High Definition

HDMI

High Definition Multimedia Interface

HDR

High Dynamic Range

HEPA

High Efficiency PA

HLS

HTTP Live Streaming

HO

Handover

HPF

High Pass Filter

H-RNTI

HS-DSCH Radio Network Temporary Identifier

HSDPA

High Speed Downlink Packet Access

HS-DPCCH

High Speed Downlink Physical Control Channel

HS-DSCH

High Speed Downlink Shared Channel

HS-FACH

High Speed Forward Access Channel

HSPA

High Speed Packet Access

HS-RACH

High Speed Random Access Channel

HS-SCCH

High Speed Shared Control Channel

HSUPA

High Speed Uplink Packet Access

HTTP

Hypertext Transfer Protocol

IC

Integrated Circuit

IEEE

Institute of Electrical and Electronics Engineers

IIP

Input Intercept Point

IIR

Infinite Impulse Response

ILPC

Inner Loop Power Control

IMEISV

International Mobile Station Equipment Identity and Software Version

IMSI

International Mobile Subscriber Identity

IMT

International Mobile Telephony

IO

Input–Output

IP

Intellectual Property

IP

Internet Protocol

IQ

In-phase/quadrature

IRAT

Inter Radio Access Technology

ISD

Inter-Site Distance

ISMP

Inter-System Mobility Policy

ISRP

Inter-System Routing Policy

ITU-R

International Telegraphic Union Radiocommunications sector

JIT

Just In Time

KPI

Key Performance Indicator

LA Location Area LAC

Location Area Code

LAU

Location Area Update

LCD

Liquid Crystal Display

LDO

Low Drop Out

LNA

Low Noise Amplifier

LO

Local Oscillator

LS

Light Sleep

LTE

Long Term Evolution

MAC

Medium Access Control

MAPL

Maximum Allowable Pathloss

MBR

Maximum Bitrate

MCL

Minimum Coupling Loss

MCS

Modulation and Coding Scheme

MDT

Minimization of Drive Tests

MGW

Media Gateway

MIMO

Multiple Input Multiple Output

MIPI

Mobile Industry Processor Interface

ML

Mismatch Loss

MLD

Maximum Likelihood Detection

MME

Mobility Management Entity

MMMB

Multimode Multiband

MMSE

Minimum Mean Square Error

MO

Mobile Originated

MOS

Mean Opinion Score

MP

Multiprocessing

MPNS

Microsoft Push Notification Service

MSC

Mobile Switching Center

MSC-S

MSC-Server

MSS

Microsoft's Smooth Streaming

MSS

MSC-Server

MT

Mobile Terminated

NA

North America

NACK

Negative Acknowledgement

NAIC

Network Assisted Interference Cancellation

NAS

Non-Access Stratum

NB

Narrowband

NBAP

NodeB Application Part

NGMN

Next Generation Mobile Network

OEM

Original Equipment Manufacturer

OL

Open Loop

OLPC

Open Loop Power Control

OMA

Open Mobile Alliance

OPEX

Operational Expenses

OS

Operating System

OSC

Orthogonal Subchannelization

OTA

Over The Air

PA

Power Amplifier

PAE

Power Added Efficiency

PAM

Power Amplifier Module

PAPR

Peak to Average Power Ratio

PCB

Printed Circuit Board

PCH

Paging Channel

PDCP

Packet Data Convergence Protocol

PDP

Packet Data Protocol

PDU

Payload Data Unit

PIC

Parallel Interference Cancellation (PIC)

PICH

Paging Indicator Channel

PLL

Phase Locked Loop

PLMN

Public Land Mobile Network

PM

Performance Management

PRACH

Physical Random Access Channel

PS

Packet Switched

P-SCH

Primary Synchronization Channel

PSD

Power Spectral Density

QAM

Quadrature Amplitude Modulation

QBE

Quadband EGPRS

QoE

Quality of Experience

QoS

Quality of Service

QPSK

Quadrature Phase Shift Keying

QXDM

Qualcomm eXtensible Diagnostic Monitor

R99

Release 99

RAB

Radio Access Bearer

RAC

Routing Area Code

RACH

Random Access Channel

RAN

Radio Access Network

RANAP

Radio Access Network Application Part

RAT

Radio Access Technology

RB

Radio Bearer

RET

Remote Electrical Tilt

RF

Radio Frequency

RFFE

RF Front End

RFIC

RF Integrated Circuit

RLC

Radio Link Control

RNC

Radio Network Controller

ROHC

Robust Header Compression

RoT

Rice over Thermal

RRC

Radio Resource Control

RRM

Radio Resource Management

RRSS

Receiver Radio Signal Strength

RSCP

Received Signal Code Power

RSRP

Reference Signal Received Power

rSRVCC

Reverse SRVCC

RSSI

Received Signal Strength Indicator

RTP

Real Time Protocol

RTT

Round Trip Time

RUM

Real User Measurements

RX

Receive

SAW

Surface Acoustic Wave

S-CCPCH

Secondary Common Control Physical Channel

SCRI

Signaling Connection Release Indication

SD

Secure Digital

SD

Sphere Decoding

S-DPCCH

Secondary Dedicated Physical Control Channel

S-E-DPCCH

Secondary Dedicated Physical Control Channel for E-DCH

S-E-DPDCH

Secondary Dedicated Physical Data Channel for E-DCH

SF

Single Frequency

SF

Spreading Factor

SFN

System Frame Number

SGSN

Serving GPRS Support Node

SHO

Soft Handover

SI

Status Indication

SIB

System Information Block

SIC

Successive Interference Canceller

SIM

Subscriber Identity Module

SINR

Signal to Interference and Noise Ratio

SIR

Signal to Interference Ratio

SMSB

Single Mode Single Band

SNR

Signal to Noise Ratio

SoC

System on Chip

SON

Self-Organizing Network

SRVCC

Single Radio Voice Call Continuity

S-SCH

Secondary Synchronization Channel

SSID

Service Set Identifier

STTD

Space Time Transmit Diversity

SW

Software

TA

Tracking Area

TAC

Tracking Area Code

TAU

Tracking Area Update

TDD

Time Division Duplex

TD-SCDMA

Time Division Synchronous Code Division Multiple Access

TFCI

Transport Format Control Indicator

TIS

Total Isotropic Sensitivity

TM

Transparent Mode

TRP

Total Radiated Power

TRX

Transceiver

TTI

Transmission Time Interval

TVM

Traffic Volume Measurement

TX

Transmit

UARFCN

UTRAN Absolute Radio Frequency Channel Number

UDP

User Datagram Protocol

UE

User Equipment

UHF

Ultrahigh Frequency

UI

User Interaction

UL

Uplink

UM

Unacknowledged Mode

UMI

UTRAN Mobility Information

UMTS

Universal Mobile Telecommunications System

URL

Uniform Resource Locator

USB

Universal Serial Bus

U-SIM

UMTS SIM

UTRAN

Universal Terrestrial Radio Access Network

VAM

Virtual Antenna Mapping

VCO

Voltage Controlled Oscillator

VIO

Input Offset Voltage

VoIP

Voice over IP

VoLTE

Voice over LTE

vSRVCC

Video SRVCC

VST

Video Start Time

VSWR

Voltage Standing Wave Ratio

WB

Wideband

WCDMA

Wideband Code Division Multiple Access

Wi-Fi

Wireless Fidelity

WiMAX

Worldwide Interoperability for Microwave Access

WLAN

Wireless Local Area Network

WPA

Wi-Fi Protected Access

WW

World Wide

XPOL

Cross-Polarized

ZIF

Zero Insertion Force

1

Introduction

Harri Holma

1.1 Introduction

GSM allowed voice to go wireless with more than 4.5 billion subscribers globally. HSPA allowed data to go wireless with 1.5 billion subscribers globally. The number of mobile broadband subscribers is shown in Figure 1.1. At the same time the amount of data consumed by each subscriber has increased rapidly leading to a fast increase in the mobile data traffic: the traffic growth has been 100% per year in many markets. More than 90% of bits in mobile networks are caused by data connections and less than 10% by voice calls. The annual growth of mobile data traffic is expected to be 50–100% in many markets over the next few years. Mobile networks have turned from voice networks into data networks. Mobile operators need to enhance network capabilities to carry more data traffic with better performance. Smartphone users expect higher data rates, more extensive coverage, better voice quality, and longer battery life.

Figure 1.1 Number of subscribers with mobile broadband technologies

Most of the current mobile data traffic is carried by HSPA networks. HSPA+ is expected to be the dominant mobile broadband technology for many years to come due to attractive data rates and high system efficiency combined with low cost devices and simple upgrade on top of WCDMA and HSPA networks. This book presents HSPA evolution solutions to enhance the network performance and capacity. 3GPP specifications, optimizations, field performance, and terminals aspects are considered.

1.2 HSPA Global Deployments

More than 550 operators have deployed the HSPA network in more than 200 countries by 2014. All WCDMA networks have been upgraded to support HSPA and many networks also support HSPA+ with 21 Mbps and 42 Mbps. HSPA technology has become the main mobile broadband solution globally. Long Term Evolution (LTE) will be the mainstream solution in the long term for GSM and HSPA operators and also for CDMA and WiMAX operators. HSPA traffic will continue to grow for many years and HSPA networks will remain in parallel to LTE for a long time. The technology evolution is shown in Figure 1.2.

Figure 1.2 Radio technology evolution

HSPA has been deployed on five bands globally, see Figure 1.3. The most widely used frequency band is 2100 MHz 3GPP Band 1 which is used in many countries in Europe, Asia, the Middle East, and Africa. The other high-band options are Band 4 at 2100/1700 MHz and Band 2 at 1900 MHz. In many cases operators have deployed HSPA on two bands: high band for capacity and low band for coverage. The two low-band options are 900 MHz and 850 MHz. The low-band rollouts have turned out to be successful in improving the network coverage, which gives better customer performance and better network quality. Low bands have traditionally been used by GSM. Using the same bands for HSPA is called refarming. GSM and HSPA can co-exist smoothly on the same band.

Figure 1.3 HSPA main frequency bands

1.3 Mobile Devices

The very first 3G devices ten years ago suffered from high power consumption, high cost, poor quality, and lack of applications. Battery power consumption during a voice call has dropped from more than 400 mA to below 100 mA in the latest devices, 3G smartphone prices have dropped below 50 EUR, and a huge number of applications can be downloaded to the devices. The attractive price points for 3G smartphones are enabled by the low-cost chip sets available from multiple vendors. The major improvement in power consumption shows the importance of RF and baseband technology evolution. The large HSPA market size brings high volume production and enables low cost devices. The attractive price points are needed to make the devices available to billions of subscribers. HSPA devices are available for many operating systems including Android, iOS, and Windows Phones. Those smartphone platforms offer hundreds of thousands of applications for the users. Not only has the data capability been improved but also the voice quality has been enhanced with the introduction of High Definition (HD) voice based on Adaptive Multirate Wideband (AMR-WB) codec.

Since the number of global HSPA frequency variants is just five, it is possible to make a global HSPA device that can be used in any HSPA network worldwide. That makes the logistics simpler. The LTE frequency options are more complex and, therefore, multiple variants of the same devices are required for the global market.

An example low end 3G device is shown in Figure 1.4: HSPA Release 7 capability with 7.2 Mbps data rate, dual band support, 2.4′′ screen, 90 g weight, and up to 500 hours of standby time. Such devices can act as the first access to the Internet for millions of people.

Figure 1.4 Low end HSPA phone – Nokia 208. Source: Nokia. Reproduced by permission of Nokia

1.4 Traffic Growth

The new smartphones and their applications are increasing traffic volumes rapidly in mobile networks. The combined data volume growth of a few major operators over a two-year period is shown in Figure 1.5: the growth rate has been 100% per year. The fast data growth sets challenges for network capacity upgrades. Higher spectral efficiency, more spectrum, and more base stations will be needed. If the traffic were to grow by 100% per year for a 10-year period, the total traffic growth would be 1000 times higher.

Figure 1.5 HSDPA data volume growth of a few major operators

Not only the data volume but also the signaling volumes are increasing in the network. Smartphone applications typically create relatively small packet sizes that are transmitted frequently, causing continuous channel allocations and releases. Figure 1.6 shows the average data volume for each High Speed Downlink Shared Channel (HS-DSCH) allocation. Each bar represents a different mobile operator. Those operators with a low data volume per allocation (below 100 kB) have high penetration of smartphones while those operators with a large data volume per allocation (above 200 kB) have relatively higher penetration of USB modems and laptop connectivity. If the average data volume per allocation is low, then more allocations and more related signaling will be required. If we assume 50–100 kB per allocation and the data volume 1 GB/sub/month, this corresponds to 330–660 allocations per subscriber per day. Assuming that the busiest hour takes approximately 6% of the daily allocations, we have then on average 20–40 allocations per busy hour. That means the network allocates resources on average every 2 minutes for every smartphone subscriber during the busy hour. Such frequent channel allocations require high signaling capacity. The high signaling also causes more uplink interference because of the control channel overhead. Therefore, solutions for uplink interference control solutions are needed in the smartphone dominated networks.

Figure 1.6 Average data volume for HS-DSCH channel allocation

1.5 HSPA Technology Evolution

The first version of HSDPA was defined in 3GPP Release 5 during 2002, the backwards compatibility started in 2004, and the first HSDPA network was launched in 2005. Release 5 allowed a maximum data rate of 14 Mbps in downlink, although the first networks supported only 1.8 and 3.6 Mbps. Since then, data rates have been increasing in 3GPP releases both in downlink and in uplink. The peak rate in Release 11 is up to 336 Mbps in downlink and 35 Mbps in uplink. The evolution is shown in Figure 1.7. The maximum data rates in the commercial networks currently (mid 2014) are 42 Mbps in downlink and 11 Mbps in uplink, which are based on Release 8.

Figure 1.7 Peak data rate evolution in downlink and in uplink

The latency evolution helps end user performance since many interactive applications benefit from the low latency. The typical latency in WCDMA Release 99 networks was 150–200 ms while HSDPA Release 5 enabled sub-80 ms latency, and the combination of HSDPA and HSUPA in Release 6 gave even sub-30 ms latency. Latency here refers to round-trip time, which is the two-way latency through the mobile and the core network. The HSPA latency improvements are shown in Figure 1.8. Latency consists of the transmit delay caused by the air interface frame sizes and by the processing delays in UE and by the delays in the network elements and transport network. The channel allocation delay has also improved considerably from typically 1 second in the first WCDMA networks to below 0.4 seconds in the latest HSPA networks.

Figure 1.8 Round-trip time evolution

1.6 HSPA Optimization Areas

A number of further optimization steps are required in HSPA to support even more users and higher capacity. The optimization areas include, for example, installation of additional carriers, minimization of signaling load, optimization of terminal power consumption, low band refarming at 850 and 900 MHz, control of uplink interference, and introduction of small cells in heterogeneous networks. Many of these topics do not call for major press releases or advanced terminals features but are quite important in improving end user satisfaction while the traffic is increasing.

The most straightforward solution for adding capacity is to add more carriers to the existing sites. This solution is cost efficient because existing sites, antennas, and backhaul solutions can be reused. The load balancing algorithms can distribute the users equally between the frequencies.

The limiting factor in network capacity can also be the signaling traffic and uplink interference. These issues may be found in smartphone dominated networks where the packet sizes are small and frequent signaling is required to allocate and release channels. The signaling traffic can increase requirements for RNC dimensioning and can also increase uplink interference due to the continuous transmissions of the uplink control channel. Uplink interference management is important, especially in mass arenas like sports stadiums. A number of efficient solutions are available today to minimize uplink interference.

The frequent channel allocations and connection setups also increase the power consumption in the terminal modem section. Those solutions minimizing the transmitted interference in the network tend to also give benefits in terms of power consumption: when the terminal shuts down its transmitter, the power consumption is minimized and the interference is also minimized.

Low-band HSPA refarming improves network coverage and quality. The challenge is to support GSM traffic and HSPA traffic on the same band with good quality. There are attractive solutions available today to carry GSM traffic in fewer spectra and to squeeze down HSPA spectrum requirements.

Busy areas in mobile networks may have such high traffic that it is not possible to provide the required capacity by adding several carriers at the macro site. One optimization step is to split the congested sector into two, which means essentially the introduction of a six-sector solution to the macro site. Another solution is the rollout of small-cell solutions in the form of a micro or pico base station. The small cells can share the frequency with the macro cell in co-channel deployment, which requires solutions to minimize the interference between the different cell layers.

1.7 Summary

HSPA technology has allowed data connections to go wireless all over the world. HSPA subscribers have already exceeded 1 billion and traffic volumes are growing rapidly. This book presents the HSPA technology evolution as defined in 3GPP but also illustrates practical field performance and discusses many of the network optimization and terminal implementation topics.

2

HSDPA and HSUPA in Release 5 and 6

Antti Toskala

2.1 Introduction

This chapter presents the Release 5 based High Speed Downlink Packet Access (HSDPA) and Release 6 based High Speed Uplink Packet Access (HSUPA). This chapter looks first at the 3GPP activity in creating the standard and then presents the key technology components introduced in HSDPA and HSUPA that enabled the breakthrough of mobile data. Basically, as of today, all networks have introduced as a minimum HSDPA support, and nearly all also HSUPA. Respectively, all new chip sets in the marketplace and all new devices support HSDPA and HSUPA as the basic feature set. This chapter also addresses the network architecture evolution from Release 99 to Release 7, including relevant core network developments.

2.2 3GPP Standardization of HSDPA and HSUPA

3GPP started working on better packet data capabilities immediately after the first version of the 3G standard was ready, which in theory offered 2 Mbps. The practical experience was, however, that Release 99 was not too well-suited for more than 384 kbps data connections, and even those were not provided with the highest possible efficiency. One could have configured a user to receive 2 Mbps in the Dedicated Channel (DCH), but in such a case the whole downlink cell capacity would have been reserved to a single user until the channel data rate downgraded with RRC reconfiguration. The first studies started in 2000 [1], with the first specification in 2002 for HSDPA and then 2004 for HSUPA, as shown in Figure 2.1.

Figure 2.1 3GPP HSDPA and HSUPA specification timeline

While the Release 99 standard in theory allows up to 2 Mbps, the devices in the field were implemented only up to 384 kbps. With HSDPA in Release 5 the specifications have been followed such that even the highest rates have been implemented in the marketplace, reaching up to 14 Mbps, and beyond this in later Releases with different enhancements, as covered in the later chapters of this book. The first HSDPA networks were opened at the end of 2005 when the first data products appeared in the marketplace, some 18 months after the 3GPP specification freeze.

With HSUPA, commercial implementations appeared in the marketplace in mid-2007, some 15 months after the specification freeze of early 2006, as shown in Figure 2.1. With smart phones it took generally a bit longer for the technology to be integrated compared to a USB dongle-style product.

The Release 6 uplink with HSUPA including up to 5.8 Mbps capability has also been widely implemented, while the uplink data rate has also been enhanced beyond the Release 6 capability in later 3GPP Releases.

2.3 HSDPA Technology Key Characteristics

The Release 99 WCDMA was basically very much circuit switch oriented in terms of resource reservation, which restricted the practical use of very high data rates while amply providing the CS services (voice and video) that did not require a too high bit rate [2]. The Release 5 HSDPA included several fundamental improvements to the WCDMA standard, with the key items being:

BTS based scheduling;

physical layer retransmissions;

higher order modulation;

link adaptation;

dynamic code resource utilization.

BTS-based scheduling moved the scheduling of the common channels from the RNC to the BTS, closer to the radio with actual knowledge of the channel and interference conditions and resource availability. While the scheduling for HSDPA was moved to the BTS, lots of other intelligence remained in the RNC, and further HSDPA functionalities were in most cases additional functionalities, not replacing RNC operation. The Release 99 functional split for the radio resource management is shown in Figure 2.2.

Figure 2.2 Release 99 radio resource management functional split

From the protocol stack point of view, the new functionality is added to the MAC layer, with the new MAC-hs added in the BTS. This is illustrated in Figure 2.3, with the remaining MAC-d in the RNC responsible for possible MAC layer service multiplexing.

Figure 2.3 HSDPA user plane protocol stack

BTS-based link adaptation allowed reaction to the momentary downlink link quality with all the information of the radio link situation available in the BTS via the new physical layer Channel Quality Information (CQI) feedback. When a user comes closer to the BTS, the received power level is better than needed. This is due to the limited downlink power control dynamic range, since signals received by the users cannot have too large a power difference. This means that a user close to the BTS has a symbol power level tens of dBs greater than necessary. This excessive symbol power level is utilized with the link adaptation. For a user in a good link quality situation, higher order modulation and less channel coding can be used. Respectively, for a user with lower link quality, more robust modulation and more channel encoding redundancy can be used for increased protection. This is illustrated in Figure 2.4.

Figure 2.4 HSDPA link adaptation

The link adaptation thus ensures successful transmission even at the cell edge. When data first arrives at the buffer in the base station, the base station scheduler will, however, first prefer to check that the UE has good link condition. Normally, the link condition changes all the time, even for a stationary user. By placing the transmissions in such time instants, based in the CQI, the resource usage is minimized and the system capacity is maximized. The time window considered in the base station varies depending on the service requirements. The detailed scheduler implementation is always left for the network implementation and is not part of the 3GPP specifications.

Transmission of a packet is anyway not always successful. In Release 99 a packet error in the decoding would need a retransmission all the way from the RNC, assuming the Acknowledged Mode (AM) of RLC is being used. Release 5 introduced physical layer retransmission enabling rapid recovery from an error in the physical layer packet decoding.

The operation procedure for the physical layer retransmission is as follows:

After transmission a packet is still kept in the BTS memory and removed only after positive acknowledgement (Ack) of the reception is received.

In case the packet is not correctly received, a negative acknowledgement (NAck) is received from the UE via the uplink physical layer feedback. This initiates retransmission, with the base station scheduler selecting a suitable moment to schedule a retransmission for the UE.

The UE receiver has kept the earlier packet in the receiver soft buffer and once the retransmission arrives, the receiver combines the original transmission and retransmission. This allows utilization of the energy of both transmissions for the turbo decoder process.

The Hybrid Adaptive Repeat and reQuest (HARQ) operation is of stop and wait type. Once a packet decoding has failed, the receiver stops to wait for retransmission. For this reason one needs multiple HARQ processes to enable continuous operation. With HSDPA the number of processes being configurable, but with the UE memory being limited, the maximum data rate is achieved with the use of six processes which seems to be adapted widely in practical HSDPA network implementations.

The HSDPA HARQ principle is shown in Figure 2.5. From the network side, the extra functionality is to store the packets in the memory after first transmission and then be able to decode the necessary feedback from the UE and retransmit. Retransmission could be identical, or the transmitter could alter which bits are being punctured after Turbo encoding. Such a way of operating is called incremental redundancy. There are also smaller optimizations, such as varying the way the bits are mapped to the 16QAM constellation points between retransmissions. The modulation aspects were further enhanced in Release 7 when 64QAM was added. While the use of 16QAM modulation was originally a separate UE capability, today basically all new UEs support 16QAM, many even 64QAM reception as well.

Figure 2.5 HSDPA HARQ operation

To address the practicality of handling high bit rates, dynamic code resource sharing was introduced. Release 99 UE was allocated a specific downlink channelization code(s) dimensioned to carry the maximum data rate possible for the connection. When the data rate was smaller, the transmission was then discontinuous to reduce interference generated to the network. Thus, with the lower data rates, the code was then used only part of the time, as shown in Figure 2.6. It is not possible to share such a resource with any other user, which causes the orthogonal code resource to run out quickly, especially with the case of variable rate packet data connections. Had one used 2 Mbps in the field, it would have basically enabled only a single user to have been configured in a cell at the time.

Figure 2.6 Code resource usage with Release 99 DPCH

The dynamic code resource utilization with HSDPA means that for each packet there is separate control information being transmitted (on the High Speed Shared Control Channel, HS-SCCH), which identifies which UE is to receive the transmission and which codes are being used. When a UE does not have downlink data to receive on HSDPA, the codes are used by another user or users. There is a pool of 15 codes with a spreading factor of 16 available for HSDPA use, assuming that there is not too much other traffic, such as Release 99 based voice, requiring codes for DCH usage. One cell can have up to four parallel HS-SCCHs which UE could monitor, but for maximum performance it is usually better to send to fewer users at a time in order to minimize the overhead needed for the signaling. From a network point of view, one could configure even more than four HS-SCCHs for one cell, giving different UEs a different set of the channels, but there is no real benefit from such a configuration. The principle of the use of HS-SCCH to point out to the UE which High Speed Physical Shared Channel (HS-PDSCH) codes to receive is illustrated in Figure 2.7. As seen also in Figure 2.6 there is a timing offset between the control and data channel, which allows the UE to determine before the 2 ms Transmission Time Internal (TTI) on HS-PDSCH which codes are intended for the UE in question. This needs to be known beforehand, especially if the UE receiver is able to receive fewer codes than the maximum of 15.

Figure 2.7 HSDPA physical channels for data and control

The HS-SCCH is divided into two parts, with the first part carrying information on the codes received and also which modulation is being used, while the second part carries information on the HARQ process being used, transport block size, redundancy version or whether the transmission is new packet, or whether it should be combined with existing data in the buffer for the particular HARQ process. The spreading factor for HS-SCCH is always fixed to be 128, and UE will check from the first part whether the control data is intended for it or not (UE specific masking prevents successful decoding of another HS-SCCH than the one intended for the UE).

The structure for the HS-PDSCH is simple, as there are only symbols carrying user data, not control fields or pilot symbols. The number of channel bits fitting on a single code on HS-PDSCH is only impacted by the modulation being used, with the alternatives being QPSK and 16QAM from Release 5 or additionally the use of 64QAM as added in Release 7. Up to 15 codes may be used in total during a 2-ms TTI, as shown in Figure 2.8. The channel coding applied on the HS-PDSCH is turbo coding.

Figure 2.8 HS-SCCH structure

Besides downlink control information, there is also the need to transmit uplink physical layer control information for HSDPA operation. For that purpose a new physical channel was added in the uplink direction as well, the High Speed Dedicated Physical Control Channel (HS-DPCCH). The HS-DPCCH carries the physical layer uplink control information as follows:

Feedback for the HARQ operation, information on whether a packet was correctly decoded or not, thus carrying positive and negative acknowledgements. In later phases, different post/preambles have also been added to avoid NodeB having to decide between the ACK/NACK and DTX state, thus improving the detection reliability.

Channel Quality Information (CQI) helps the base station scheduler to set correct link adaptation parameters and decide when is a good moment to schedule the transmission for a particular user. The CQI basically gives from the UE an estimate of what kind of data rate could be received reliably in the current conditions the UE is experiencing. Besides the external factors, such as interference from other cells and relative power of the own cell (serving HSDPA cell), the CQI also covers UE internal implementation aspects, such as whether there is an advanced receiver or whether the UE has receiver antenna diversity.

The network can parameterize the HS-DPCCH to also use different power offsets for the different fields carrying CQI and ACK/NACK feedback, as shown in Figure 2.9. It is also worth noting that the HS-DPCCH is never transmitted on its own, but always in connection with the uplink DPCCH. Uplink DPCCH contains pilot symbols and thus provides the necessary phase reference for the NodeB receiver. Besides pilot symbols, the uplink DPCCH carries the downlink power control commands and the Transport Format Combination Indication (TFCI) to inform on the data rate being used on the uplink Dedicated Physical Data Channel (DPDCH). If only HSDPA is operated without HSUPA, all uplink user data is then carried on DPDCH.

Figure 2.9 HS-DPCCH carrying ACK/NACK and CQI feedback

The CQI information does not map directly to a data rate but actually provides an indication of the transport block size, number of codes, and level of modulation which the terminal expects it could receive correctly. This in turn could be mapped to a data rate for the assumed power available for HSDPA use. Depending on the type of UE, the CQI table covers only up to such modulation which the UE can support, after that an offset value is used. For this reason there are multiple CQI tables in [2]. In the first phase of HSDPA introduction the UEs on the market supported only QPSK modulation, but currently the generally supported modulations also include 16QAM, and in many cases also 64QAM. The values indicated by the CQI are not in any way binding for the NodeB scheduler; it does not have to use the parameter combination indicated by the CQI feedback. There are many reasons for deviation from the UE feedback, including sending to multiple UEs simultaneously or having actually different amounts of power available for HSDPA than indicated in the broadcasted value.

2.4 HSDPA Mobility

The mobility with HSDPA is handled differently from Release 99 based DCH. Since the scheduling decisions are done independently in the NodeB scheduler to allow fast reaction to the momentary channel conditions, it would be difficult to follow the Release 99 soft handover (macro diversity) principle with combining time aligned identical transmission from multiple BTS sites. Thus the approach chosen is such that HSDPA data is only provided from a single NodeB or rather by a single cell only, called the serving HSDPA cell. While the UE may still have the soft handover operation for DCH, the HSDPA transmission will take place from only a single NodeB, as shown in Figure 2.10.

Figure 2.10 Operation of HSDPA from a single NodeB only with DCH soft handover

When operating within the active set, the UE shall signal the change of the strongest cell and then the serving HSDPA cell can be changed to another NodeB. If the strongest cell is not part of the active set, then the new cell needs to be added to the active set before HSDPA transmission can take place from that cell. When changing the serving HSDPA cell, the packets in the source NodeB buffer are thrown away once the serving cell change is complete and the RNC will forward the packets not received to the new serving HSDPA cell NodeB, in case Acknowledged Mode (AM) operation of RLC is being used.

2.5 HSDPA UE Capability

In Release 5 a total of 12 UE categories were defined supporting up to 14 Mbps downlink peak data rate. Some of the UE categories were based on the possibility of having the UE receiving data not during consecutive TTIs, but in reality that possibility has not been utilized. In the first phase the products in the market had only QPSK modulation supported, enabling only 1.8 Mbps physical layer peak data rate, but soon implementations with 16QAM also became available enabling first 3.6 Mbps and then later up to 7.2 Mbps or even up to 10 or 14 Mbps. The relevant UE categories, which actually have been introduced to the market in wider scale, are shown in Table 2.1, with all categories given in [3]. The RLC data rates are calculated with 40 bit RLC packet size. In Release 7 3GPP introduced a flexible RLC packet size which allows reduction of the RLC header overhead.

Table 2.1 Release 5 HSDPA UE categories and their L1/RLC peak data rates