2 wcdma hsupa principle

Confidential Information of Huawei. No Spreading Without Permission

www.huawei.com

Copyright © 2008 Huawei Technologies Co., Ltd. All rights reserved.

WCDMA HSUPA

Principles



Foreword

� HSUPA: High Speed Uplink Packet Access

� HSUPA, as one of important feature from Huawei RAN6, has

been taken as an important enhancement to improve the

network performance

� This introduces an important feature of 3GPP R6, High Speed Uplink Packet Access

(HSUPA). As an uplink (UL) high speed data transmission solution, HSUPA provides a theoretical maximum UL rate of 5.76 Mbps on the Uu interface.

� It improves user experience with significantly higher data rate, lower delay, and faster connection setup, which allows operators to offer new services and attract new users.



Objectives

� Upon completion of this course, you will be able to:

� Outline the protocol architecture of HSUPA

� Know the key technologies of HSUPA



Contents

1. Introduction of HSUPA

2. HSUPA Concepts

3. Physical Layer Channels and Processing

4. MAC Protocols and Procedure



High Speed Uplink Packet Access

� Driver force for HSUPA

� Data Rate – demand for higher peak data rates in uplink

� Qos – lower latency

� Capacity – better uplink throughput

� Coverage – better uplink coverage for higher data rate

� Data Services are expected to grow significantly within the next few years. Current 2.5G

and 3G operators are already reporting that a significant proportion of usage is now devoted to data, implying an increasing demand for high-data-rate, content-rich

multimedia services. Although current Release 99 WCDMA systems offer a maximum practical data rate in Uplink of 384 kbps, the 3rd Generation Partnership Project (3GPP) has included in Release 6 of the specifications a new high-speed, low-delay feature called

High Speed Uplink Packet Access (HSUPA).

� HSUPA provides significant enhancements to the Uplink compared to WCDMA Release 99

in terms of peak data rate, cell throughput, and latency. This is achieved through the implementation of a fast resource control and allocation mechanism that employs such

features as Adaptive Coding, fast Hybrid Automatic Repeat Request (HARQ) and Shorter Physical Layer frames. With the addition of HSUPA, a better balance between Downlink HSDPA and Uplink traffic performance is also achieved.

� The High Speed Uplink Packet Access (HSUPA) is a 3GPP Release 6 feature, also called Enhanced Uplink (EUL) or Enhanced DCH (E-DCH).



UMTS Data Rate Evolution

GSM GPRS

EDGE

WCDMA

R99

HSDPA

R5

HSUPA

R6

10Mbps1.4Mbps/5.76MbpsHSUPA Release 6

10Mbps384KbpsHSDPA Release 5

384Kbps384KbpsWCDMA Release 99

120Kbps60KbpsEDGE

40Kbps20KbpsGPRS

9.6Kbps9.6KbpsGSM

Downlink Peak Data RateUplink Peak Data RateMobile Network

� Mobile network data rate evolution

� WCDMA data transmission evolved from GSM/GPRS, inheriting much of the upper layer functionality directly from those systems. The first commercial deployments of

WCDMA are based on a version of the standards called Release 99, with HSDPA introduced in Release 5 to offer higher speed Downlink data services.

� Enhanced Data rates for GSM Evolution (EDGE) is another system in the GSM/GPRS

family that some operators have deployed as an intermediate step before deploying WCDMA.

� Release 6 introduces the High Speed Uplink Packet Access (HSUPA) to provide faster data services for the Uplink.

� For HSUPA (Uplink) the theoretical maximum achievable peak data rate is 5.76 Mbps, while for HSDPA (Downlink) it is 14.4 Mbps.



Release 99 Uplink Packet Data

� DCH (Dedicated Channel)

� Variable spreading factor

� Closed loop power control

� Macro diversity (soft handover)

� RACH

� Common spreading code

� Fixed spreading factor

� No closed loop power control

� No soft handover

� Release 99 Uplink Packet Data

� There are two different techniques defined in the Release 99 specification to enable Uplink packet data. Most commonly, data transmission is supported using

either the Dedicated Channel (DCH) or the Random Access Channel (RACH).

� The DCH is the primary means of supporting packet data services. Each UE uses an Orthogonal Variable Spreading Factor (OVSF) code, dependent on the required

data rate. Fast closed loop Power Control is employed to ensure that a target Signal-to-Interference Ratio (SIR) is maintained in order to control the block error

rate (BLER). Macro Diversity is supported using soft handover.

� Data transfer can also be supported on the RACH. This common channel employs

an OVSF code, with a spreading factor between 32 and 256, as negotiated with UTRAN during the Access procedure. Because it needs to be shared among all UEs, higher data rates are generally not supported. Macro Diversity is also not

supported and the channel operates with a fixed (or slow changing) power allocation



Release 99 Uplink Limitation

� Large scheduling delay

� Radio resource is controlled from RNC

� Uplink DCCC (Dynamic channel configuration control)

� Large latency

� Transmission time interval duration of 10/20/40/80ms

� RNC based retransmission in case of errors (RLC layer)

� Limited uplink data rate

� Deployed peak data rate is 384kbps with limited subscriber

number

� Release 99 Uplink Limitations

� Among the available options for Uplink data transmissions in Release 99, the Common Channel (RACH) only allows for a small amount of data and a limited

duration of the transmission. Thus, from a practical point of view, the Dedicated Channel (DCH) is the way to accommodate packet services in a Release 99 network. However, significant limitations must also be faced when using the Uplink DCH:

� Large Scheduling Delay: In Release 99, the scheduling of resources is done in the serving RNC and involves Layer 3 signaling messages to and from the UE, which

causes the mechanism to be relatively slow in assigning or reconfiguring the resources assigned to a particular UE.

� Large Latency: The transmission time interval can vary from 80 ms down to 10 ms as best case, posing an intrinsic boundary to the latency values. In addition to that, the only available mechanism for retransmissions of erroneous packets is located in

RNC, thus significantly contributing to the latency figures

� Limited Uplink Data Rate: Though the standard allows for high data rate on the

Release 99 Uplink, typical values of maximum data rate observed in deployed

networks range from 64 kbps up to 384 kbps, while using a spreading factor of 4.

In order to achieve higher peak data rates, lower coding rates and multi-code

transmission shaould be used, but these are not available in R99 systems.



High Speed Uplink Packet Access

� E-DCH channel has been introduced

� Interference is shared by multiple users

� NodeB controls all UEs data rate with fast scheduling

E-DCH

E-DCH

E-DCHE-DCH

� HSUPA is realized by introducing the Enhanced Dedicated Channel (E-DCH)

� In HSUPA, the Node B allows several UEs to transmit at a certain power level at the same time. These grants are assigned to users by using a fast scheduling algorithm that allocates

the resources on a short-term basis (every 10ms or 2ms).

� The rapid scheduling of HSUPA is well suited to the burst feature of packet service. During periods of high activity, a given user may get a larger percentage of the available resources,

while getting little or no bandwidth during periods of low activity.



Improved Characters by HSUPA

� Higher peak data rate in uplink

� Reduced latency

� Faster retransmission to improve throughput

� Fast scheduling

� Optimize the resource allocation to maximize the total throughput

� Quality of Service support

� Improve QoS control and resource utilization

� Enhancement methods in HSUPA

� To overcome the Release 99 limitations previously mentioned, HSUPA has been introduced in Release 6

� The use of shorter TTI, fast resource scheduling, and fast retransmissions at the physical layer improves uplink data services, while addressing the release 99 limitations in terms of latency, peak data rate, coverage, and capacity. Additionally,

improved quality of service support helps to optimize resource utilization and guarantee the promised quality



HSUPA Key Technology Overview

� HSUPA key technologies

2ms TTI

Fast scheduling

Lower SF

New Channels

Fast L1 HARQ

Improved

Cell Capacity

Higher Peak Data Rate

Lower Latency

Improved QoS

Support

Fast Resource Scheduling

� Compared to R99 UL DCH, the enhance DCH specified for HSUPA in Release 6 offers the

following features:

� Shorter TTI of 2ms: which can reduce the latency and can be scheduled faster

� Lower SF: which can increase physical channel symbol rate , higher peak data rate is available

� Uplink L1 HARQ throughput: improve physical layer performance with fast

retransmissions

� New transport and physical channels

� Fast resource control: with new MAC entities in NodeB, radio resource can be scheduled faster to optimize the total throughput



Application Scenarios for HSUPA

� VoIP (Voice over IP)

� Low latency, Quality of Service control and improved uplink

capacity

� Game

� Lowe latency, fast resource allocation

� Personal blog update

� Upload personal essay, video, music and picture



Contents


2. HSUPA Concepts





HSUPA vs. HSDPA

Fast NodeB scheduler

Rise-over-Thermal (ROT)

Fast NodeB scheduler

Shared NodeB power and code

Fast power control

Soft handover

Rate/modulation adaptation

Single serving cell

HARQ with fast retransmission at layer 1

Dedicated channel with enhanced

capabilities

New high-speed shared channel

HSUPAHSDPA

� This slide lists some important aspects for comparing HSDPA and HSUPA to help

understand HSUPA principles and operation.

� The HSDPA concept is based on high speed shared channels with fast L1 HARQ

retransmission and rate and modulation adaptation to adjust to channel conditions. The fast scheduler is located in the Node B and assigns the available resources (power and codes) to several users. This enables cell power to be directed to a single user (or to a

small group of users) for a short period of time, during which other users do not get any data. In this way, one Node B transmitter can be shared among many UE receivers.

� For HSUPA, the channel remains a dedicated channel, but with enhanced capabilities such as fast scheduling and L1 HARQ retransmissions. Power control and soft handover are still

used to adapt to radio channel conditions. Because each UE has an independent transmitter with its own power and code availability, the HSUPA scheduler can accommodate many users to be received by the same Node B, where the Rise-over-

Thermal Noise level indicates the uplink loading of the system.



Rise-over-Thermal Noise

� In order to decode received data correctly, the uplink

interference shall be controlled.

� Rise-over-Thermal is a measure of the uplink load.

NodeB monitors uplink interference and tells UE how much power can be used to transmit uplink data.

� The Rise-over-Thermal noise level is a measure of the uplink load at the NodeB receiver.

� By increasing the number of UEs transmitting on the uplink and their transmit power, the overall level of interference in the uplink band also increases.

� The NodeB receiver perceives this level as noise, and it directly affects the decoding performance of uplink data transmissions.

� The NodeB controls the interference level by adjusting the UE grant assignments

according to the current interference level.

� When the UE receives a new grant, it uses it in combination with available UE

transmit power and amount of data in the buffer to determine the data rate and the corresponding transmit power.



NodeB Scheduler for HSUPA

� The HSUPA scheduler considers the trade-off between the

following two points:

� Several users those want to transmit at high data rate all the time

� Satisfying all requested grants while preventing overloading and

maximizing resource utilization

� Similar to HSDPA, HSUPA implements fast resource allocation and control with a scheduler

in the NodeB. While the HSDPA scheduler accommodates a common resource to several users, the HSUPA scheduler has a different task: it coordinates the reception of data

transmitted from several UEs to a single NodeB. This can be regarded as a very fast resource allocation of a dedicated channel (E-DCH), rather than a sharing of a common channel (HS-DSCH).

� On one side, each UE will tend to transmit as much as possible based on channel conditions, the amount of data in the buffer, and the power available. On the other side,

the scheduler will try to satisfy each connected UE while preventing overloading and maximizing resource utilization and cell throughput.



HSUPA Channel Operation

� The UE sends a transmission request to the

NodeB for getting resources.

� The NodeB responds to the UE with a grant

assignment, allocating uplink band to the

UE.

� The UE uses the grant to select the

appropriate transport format for the Data

transmission to the NodeB.

� The NodeB attempts to decode the received

data and send ACK/NACK to the UE. In case

of NACK, data may be retransmitted.

1. R

EQUEST

3. D

ATA

2. G

RANT

4. A

CK/N

ACK

� This slide illustrates HSUPA operation :

� 1. The UE asks the NodeB for a grant to transmit data on uplink.

� 2. If the Node B allows the UE to send data, it indicates the grant in terms of

Traffic-to-Pilot (T/P) ratio. The grant is valid until a new grant is provided.

� 3. After receiving the grant, the UE can transmit data starting at any TTI and may include further requests. Data are transmitted according to the selected transport

format, which is also signaled to the NodeB.

� 4. After the Node B decodes the data, it sends an ACK or NAK back to the UE. If

the NodeB sends a NAK, the UE may send the data again with a retransmission.



HSUPA Channel Operation (continued)

� 1. Transmission Request

� The UE request data transmission

by the scheduling information (SI),

which is determined according to

the UE power and buffer data

availability.

� The scheduling information is sent

to the NodeB.

UE

UE Buffer UE Power

Scheduling Information (SI)

� This slide illustrates a data transmission request from the UE through scheduling

information (SI), by which the UE asks the Node B for a grant to transmit data on Uplink E-DCH.

� UE power availability and UE buffer status are combined to determine the scheduling information.




� 2. Grant Assignment

� The Node B determines the UE

grant by monitoring uplink

interference (RoT at he receiver),

and by considering the UE

transmission requests and level of

satisfaction.

� The grant is signaled to the UE by

new grant channels.

NodeB

RoT SI

GRANT

Satisfaction

� This slide illustrates an HSUPA absolute grant assignment upon request from the UE. The

grant is determined based on uplink interference situation (Rise-over-Thermal noise) at the NodeB receiver and on the UE transmission requests and level of satisfaction.

� The Node B indicates the Traffic-to-Pilot (T/P) grant by downlink grant channel. The grant is valid until a new grant is given or modified.




� 3. Data Transmission

� The UE uses the received grant and,

based on its power and data

availability, selects the E-DCH

transport format and the

corresponding transmit power.

� Data are transmitted by the UE on

together with the related control

information.

UE

GRANTUE Power

Data and relatedcontrol information

UE Buffer

� This slide illustrates an HSUPA Data Transmission for scheduled grants.

� After receiving the grant, the UE can transmit data starting at any TTI and may include additional scheduling information. The transport format is first selected based on the

received grant, on the available power and on the data in the buffer.

� Data are transmitted on a set of E-DPDCH channels, and transport format Information is signaled to the Node B on the corresponding E-DPCCH. The Happy Bit (Happy Bit indicates

the UE’s level of satisfaction. ) is also included.




� 4. Data Acknowledgment

� The NodeB attempts to decode the

received data and indicates to the

UE with ACK/NACK is successful.

� If no ACK is received by he UE, the

data may be retransmitted.NodeB

ACK/NACK

Data and relatedcontrol information

� This slide illustrates the acknowledgment of data at the NodeB and HARQ retransmission.

� After the NodeB attempts to decode the data, it sends an ACK or NACK to the UE. If the NodeB sends a NACK, the UE may send the data again with a fast retransmission.



HSUPA Protocol Stack

SM (Session Management)

GMM (GPRS Mobility Management)

RRC (Radio Resource Control)

MAC-es and MAC-d (Medium Access Control)

RLC (Radio Link Control)

MAC-e

Physical LayerIub Interface

Iu Interface

UE NodeB RNC SGSN

AS

NAS

� In a Release 99 PS network, the NAS layer protocols terminate at the SGSN. The RRC, RLC,

and MAC protocols terminate at the RNC. The Physical Layer protocol terminates at the NodeB.

� The Release 5 specifications define a new sub-layer of MAC for the downlink called MAC-hs, which implements the MAC protocols and procedures for HSDPA. This sub-layer operates at the NodeB and the UE. The location of MAC-hs in the Node B has an

important implication for HSDPA operation.

� Similarly, the Release 6 specifications define a new sub-layer of MAC for the uplink called

MAC-e/es, which implements the MAC protocols and procedures for HSUPA. This sub-layer operates at the NodeB (MAC-e), at the RNC (MAC-es), and the UE (MAC-e/es).

� The location of MAC-e in the NodeB has an important implication for HSUPA operation, allowing for fast retransmissions at the physical Layer. The MAC-es, which is responsible for reordering of the data packets, is located in the RNC for HSUPA because a UE may be

in soft handover with multiple Node Bs. Transport channel frames are constructed by the MAC sublayer in the UE and sent over the air interface to each NodeB with which the UE is

in soft handover. The RNC receives identical transport channel frames from each NodeBover the Iub interfaces and performs reordering.



New HSUPA Uplink Channels

� Enhanced Uplink Dedicated Channel (E-DCH)

� Uplink Transport Channel

� E-DCH Dedicated Physical Data Channel (E-DPDCH)

� Uplink Physical Channel

� E-DCH Dedicated Physical Control Channel (E-DPCCH)

� Uplink Control Channel

� HSUPA introduces one new uplink transport channel and two new uplink physical

channels.

� Enhanced Uplink Dedicated Channel (E-DCH) – An uplink transport channel. The E-DCH

operates on a 2 or 10 ms Transmission Time Interval (TTI) and carries a single transport block per TTI. The channels is mapped on one or more (up to 4) E-DCH Dedicated Physical Data Channels (E-DPDCHs) and has an associated E-DCH Dedicated Physical Control

Channel (E-DPCCH).

� E-DCH Dedicated Physical Data Channel (E-DPDCH) – An uplink physical channel used to

carry uplink data for the E-DCH transport channel. It supports BPSK modulation with I and Q branches and it is allocated every TTI. Up to 4 channels can be used to carry the E-DCH

transport channel in a multi-code transmission scheme.

� E-DCH Dedicated Physical Control Channel (E-DPCCH) – An uplink physical channel for control information associated with E-DPDCH. It carries information about the transport

format used on E-DCH and the HARQ retransmission sequence number; it includes one bit to support scheduling decisions at the NodeB.



New HSUPA Downlink Channels

� E-DCH Hybrid ARQ Indicator Channel (E-HICH)

� Downlink Physical Channel

� E-DCH Absolute Grant Channel (E-AGCH)


� E-DCH Relative Grant Channel (E-RGCH)


� HSUPA introduces three new downlink physical channels:

� E-DCH Hybrid ARQ Indicator Channel (E-HICH) – A downlink physical channel that carries feedback (ACK/NAK) from the Node B on the previous data transmission, to support HARQ

retransmission. Since soft handover is supported for HSUPA, each cell belonging to the E-DCH Active Set transmits the E-HICH.

� E-DCH Absolute Grant Channel (E-AGCH) – A downlink physical channel that carries

scheduler grant information from the E-DCH serving cell. The absolute grant indicates directly to the UE the Traffic-to-Pilot ratio that shall be used for scheduled transmissions.

� E-DCH Relative Grant Channel (E-RGCH) – A downlink physical channel that carries scheduler grant information from cells belonging to the serving NodeB as well as to non-

serving cells in the E-DCH active set. The relative grant tells the UE to increase, decrease, or maintain the current Traffic-to-Pilot ratio.



HSUPA Channel Mapping

DCCH DTCH

E-DCH

E-DPCCH

E-DPDCH

E-HICH

E-AGCH

E-RGCH

� DCCH and DTCH can be mapped to E-DCH.

� A UE using HSUPA can also have additional Release 99 DCH and/or HSDPA channels, although the standard specifies restrictions for the possible combinations. Because power

control and soft handover are supported for E-DCH, the channel can be seen as an extension of the Release 99 DCH.

� The E-DPCCH, E-HICH, E-AGCH, and E-RGCH are physical layer (control) channels. They

carry no upper layer information, and therefore have no logical or transport channel mapping.



Uplink Channels

� E-DPDCH

� Carries the payload.

� May include a scheduling request from UE

to NodeB.

� E-DPCCH

� Carries control information required to

decode the payload carried by E-DPDCH.

� Carries an indication from UE to indicate

NodeB whether the assigned resources are

adequate.

TTI

SIPayloadHD

TTI

Resource

StatusControl

Information

� When the UE is operating in HSUPA mode, it uses the E-DPDCH to transmit data on the

uplink. Scheduling information also can be included as in-band signaling.

� MAC-e PDU Header (HD) – Indicates the composition of the MAC-e PDU payload in

terms of data descriptor and number of MAC-es PDU included.

� Payload – Includes the uplink data as MAC-es PDU of the data flow multiplexed in the transmitted MAC-e PDU.

� Scheduling Information (SI) – SI may be included. SI informs the NodeB about UE buffers, data flow priority, and UE power availability, in order to receive a proper

transmission grant.

� The E-DPCCH tells the NodeB whether the transmitted block is a new transmission or a

retransmission, and which transport format is selected for the E-DCH data channel. An additional bit is also included as feedback on the current resource status.

� Control Information – Includes a sequence number used to identify retransmitted

transport blocks and an indicator of the transport format combination used on the EDPDCH.

� Resource Status – The E-DPCCH also carries one bit (Happy Bit), which the UE uses to tell the NodeB that the granted data rate is not satisfactory.



Downlink Channels

� E-AGCH

� The absolute grant carries maximum allowed E-

DPDCH/DPCCH power ratio.

� Carries information that controls HARQ process.

� E-RGCH

� The relative grant carries a simple command to

increase (UP), decrease (DOWN) or keep (HOLD)

the current grant.

� E-HICH

� Gives feedback to the UE about previous data

transmission, carrying acknowledge (ACK) or

not acknowledge (NACK).

TTI

HARQ

ControlT/P Grant

TTI

Up/Down/Hold

TTI

ACK/NACK

� E-AGCH: The Serving HSUPA cell uses the absolute grant channel to tell the UE what

power level can be used for data transmission on E-DCH. It consists of the following:

� Traffic-to-Pilot Grant – Indicates the ratio between the E-DPDCH power and the

DPCCH power that the UE can use for E-DCH transmissions. The grant is a value ranging from 0 to 31 and coded in 5 bits. For each value, the standard specifies a Traffic-to-Pilot ratio.

� HARQ Control – An additional bit indicates the scope of the grant, which can be related only to the current HARQ process, implicitly identified by timing, or can

affect all running HARQ processes. In case of 10 ms TTI, the scope can only be “All Processes.”

� E-RGCH and E-HICH：Each cell belonging to the E-DCH active set can use the relative grant channel to tell the UE to increase, decrease, or keep the current grant. Each of those cells can transmit the hybrid ARQ indicator channel to tell the UE about the success (ACK)

or not (NAK) of the previous data transmission. The channel structure is the same for both channels. Up to 40 channels (20 E-RGCH and 20 E-HICH) can be multiplexed on a single

downlink code channel.



HSUPA Features

� Shorter TTI of 2ms

� In HSUPA both 10ms TTI and 2ms TTI are supported.

� A shorter TTI allows reduction of the latency and increasing the

average and peak cell throughput.

� Higher Peak Data Rate

� For a 10-ms TTI UE, peak data rate is limited to 2 Mbps.

� Higher peak data rates can be achieved with a 2ms TTI UE

� 5.76Mbps is the maximum peak data rate for HSUPA.



HSUPA Features (continued)

� Hybrid-ARQ

� N-channel stop-and-wait protocol, with 4 HARQ processes for

10ms TTI and 8 HARQ processes for 2ms TTI

� Synchronous retransmission

� Separate HARQ feedback is provided per radio link.

� Hybrid ARQ – The hybrid ARQ for HSUPA consists of an N-Channels stop-and-wait

protocol. The number of HARQ processes is 4 for a 10 ms TTI and 8 for a 2 ms TTI configuration. The retransmission is synchronous, with separate feedback provided for

each radio link. After requesting and receiving a grant for data transmissions:

� The UE transmits the data of the corresponding HARQ process to all NodeBs for which a radio link exists.

� Each Node B connected to the UE sends ACK/NAK back to the UE.

� The transmission is successfully completed if an acknowledge (ACK) is received.



E-DCH Active Set and Mobility Support

� There are three different types of radio

links in the UE E-DCH active set:

� Serving E-DCH Cell: The cell from which

UE receives AGCH.

� Serving E-DCH RLS: Set of cells that

contain at least the serving cell and

from which the UE can receive RGCH

� No-Serving RL: Cell that belongs to the

E-DCH active set but not belong to the

serving RLS and from which the UE can

receive a RGCH.

Serving

E-DCH cell

Serving E-DCH

Radio Link Set

(RLS)

Non-Serving E-DCH Radio

Link (RL)

� The E-DCH Active Set is limited to 4 cells, one of which is the E-DCH serving cell.

� The radio links that are in softer handover with the E-DCH serving cell (i.e., connected to the same NodeB) constitute the Serving E-DCH Radio Link Set (RLS).

� All other links in the E-DCH active set, which are connected to other NodeBs, are non-serving radio links.



Theoretical HSUPA Maximum Data Rate

� How to get 5.76Mbps:

� Lower channel coding gain

� Effective code rate = 1

� Requires very good channel conditions to decode

� Lower spreading factor

� UE uses SF 2

� Multi-code transmission

� UE uses 4 codes, 2 with SF2 and 2 with SF4

� 2ms TTI

� The following assumptions are needed to achieve the theoretical maximum data rate of

5.76 Mbps:

� Lower channel coding gain – Using an effective code rate of 1 increases the data

rate, but the channel conditions must be very good for the NodeB to correctly decode every data block on the first transmission.

� Lower spreading factor – UE must use SF 2.

� Multi-code transmission – Four codes (2 codes with SF2 and 2 codes with SF4) are used by E-DPDCH.

� Shorter TTI – 2ms TTI is needed. Because the maximum transport block size is 20000 bits with 10ms TTI, the maximum data rate for 10ms TTI is 2Mbps.

� In a practical scenario, the practical maximum data rate will be less than 5.76 Mbps, due to less than ideal channel conditions, the need for retransmission, and the need to share the UE power with other channels.



E-DPDCH with SF4 and Puncturing

� Maximum payload for spreading factor of 4, TTI of 2 ms and coding

rate of 1 is 1920 bits and the corresponding data rate is 960kbps.

1920 bits payload

1920 parity

1920 symbols

1920 modulation

symbols

1920 systematic 1920 parity

7690 chips

R = 1/3

Turbo Coding

SF=4

BPSK Modulation

Puncturing

2ms

� The examples presented so far have assumed a turbo code rate of 1/3 and BPSK

modulation. If we assume a single E-DPDCH and a transport block containing 640 data bits, rate 1/3 turbo coding produces 1920 symbols. BPSK modulation maps one symbol

onto one modulation symbol, which is then spread by the OVSF of length 4. This results in 7680 chips sent every 2ms, corresponding to the fundamental WCDMA chip rate of 3.84 Mcps.

� If the transport block is not exactly 640 data bits, the rate matching step adjusts the number of symbols after turbo coding to produce 1920 symbols.

� By increasing the coding rate, more data bits can be transmitted in a 2 ms TTI, thus increasing the data rate. Using a coding rate of 1, the data rate becomes 960 kbps,

because 1920 bits can be transmitted in 1920 modulation symbols. This corresponds to puncturing all the parity bits and transmitting only the systematic bits.



Lower Spreading Factor SF2

� Maximum payload for spreading factor of 2, TTI of 2 ms and coding

rate of 1 is 3840 bits and the corresponding data rate is 1920kbps.

3840 bits payload

3840 parity

3840 symbols

3840 modulation

symbols

3840 systematic 3840 parity

7690 chips

R = 1/3

Turbo Coding

SF=2

BPSK Modulation

Puncturing

2ms

� By changing the spreading factor from 4 to 2, the number of bits that can be transmitted

in a single TTI doubles from 640 to 1280, because now 7680/2 = 3840 symbols can be mapped onto 7680 chips. Again, a coding rate of 1 is assumed.



Multi-code Transmission

� For one UE in HSUPA operation, up to 4 E-DPDCH can be used

simultaneously, two using SF4 and two using SF2.

� Use of 4 codes transmission 2*SF2 + 2*SF4:

� (2*1920kbps) + (2*960kbps) = 5760kbps



HSUPA UE Capabilities

5742kbps2000kbps2&10msSF2 4Category 6

--2000kbps10msSF22Category 5

2886kbps2000kbps2&10 msSF22Category 4


1448kbps1448kbps2&10 msSF42Category 2


Peak rate

for TTI =

2ms

Peak rate

for TTI =

10MS

Supported

TTI

Minimum

SF

Max number

of E-DPDCH

channels

E-DCH

category

� What determines the maximum data rate supported by different categories of UE? It is a

combination of the maximum number of E-DPDCH channels, the spreading factor, and maximum bits in one TTI.

� For 10 ms TTI, a maximum of 2 Mbps peak data rate can be achieved, corresponding to a maximum transport block size of 20000 bits. To achieve higher rates, a TTI of 2 ms shall be used.

� With a single E-DPDCH channel, a spreading factor from 256 to 4 is allowed. For multi-code transmissions, only SF4 and SF2 are allowed, in the following combinations: (2 x SF4)

or (2 x SF2) or (2 x SF4 + 2 x SF2). Note that SF=2 is not permitted on a single code transmission.



Contents


2. HSUPA Concepts





New Channels for HSUPA

� Uplink Transport Channel

� E-DCH: Carries high speed uplink data

� Uplink Physical Channels

� E-DPDCH: Carries E-DCH

� E-DPCCH: Carries control signal for E-DPDCH

� Downlink Physical Channels

� E-HICH: Carries HARQ ACK/NACK indicator for E-DCH

� E-RGCH: Carries relative grant determined by the scheduler

� E-AGCH: Carries absolute grant determined by the scheduler

� The main introduction in Release 6 is the new data channel, Enhanced Dedicated Channel

or E-DCH, which carries the uplink high speed data. New physical channels are introduced to support E-DCH.

� On the uplink, two new physical channels are introduced: E-DPDCH (Dedicated Physical Data Channel for E-DCH) and E-DPCCH (Dedicated Physical Control Channel for E-DCH). The E-DCH can be mapped to one to four uplink E-DPDCHs (Dedicated Physical Data

Channels for E-DCH), with improved coding and modulation design. The physical layer control information, E-TFCI etc., is carried on one E-DPCCH (Dedicated Physical Control

Channel for E-DCH).

� On the Downlink, three new physical channels are introduced to support E-DCH. The

downlink physical channels E-HICH (HARQ Indicator Channel for E-DCH) and E-RGCH (Relative Grant Channel for E-DCH) are dedicated channels and they share a single channelization code assigned by the higher layer to the UE. The UE increases or decreases

its E-DCH data rate based on the relative grant indicator on E-RGCH. The downlink channel E-AGCH (Absolute Grant Channel for E-DCH) is a common channel shared by all the users

in the cell. The addressing on E-AGCH is realized by masking CRC bits with E-RNTI (RNTI for E-DCH).



New Channels in HSUPA Operation

� 1. The UE sends a request for resources. The

request includes status of its data buffers and is

sent on E-DPDCH.

� 2. Based on the request from the UE, the Node B

allocates a resource grant to the UE. The grant is

sent on the E-AGCH channel.

� 3. This grant can be modified by the Node B

every TTI using the E-RGCH channel.

� 4. The UE transmits data on E-DPDCH. Control

information needed to decode the data is sent

on E-DPCCH.

� 5. The Node B decodes the received packet and

informs the UE whether it could decode the data

successfully or not on the E-HICH channel.

E-DPDCH

E-DPCCH

E-AGCH

E-RGCH

E-HICH

1

4

3 521

4



E-DCH

� E-DCH is mapped to one or more E-DPDCHs.

� Control information for E-DCH is sent to E-DPCCH.

� One transport block (TB) is transferred in one TTI.

� Transmission time interval (TTI) can be 10ms or 2ms.

� Support for 10ms is mandatory in the UE.

� Support for 2ms is mandatory for UE with E-DCH peak capability

above.

� The standard allows up to four E-DPDCH channels to be configured for a single E-DCH.

However, the actual maximum number of E-DPDCH channels to be used in a connection depends on the UE capability as well as the UTRAN configuration.

� For E-DPCCH, only one E-DPCCH can be configured per UE.

� For HSUPA operation, the Node B negotiates the transmission time interval (TTI) on E-DCH for each UE. The decision could be based on the UE capability as well as on the negotiated

peak data rate.



E-DCH Channel Coding

� CRC

� A 24 bit CRC is attached to the

transport block.

� Channel Coding

� Turbo coding with 1/3 coding

ratio

Transport block

from MAC

Add CRC

attachment

Code block

segmentation

Channel coding

Physical layer HARQ/

rate matching

Physical channel

segmentation

Interleaving & physical

Channel mapping

Physical channel(s)

� Code block segmentation: Here, large transport blocks (with CRC bits) are chopped into

several smaller code blocks since the turbo encoder accepts only blocks with fixed length.

� Channel Coding: Only turbo coding is used and the channel coding ratio is 1/3.



E-DCH HARQ Rate Matching

� Hybrid HARQ/Rate Matching

� Hybrid ARQ match the number of bits at the turbo

coder to the total number of bits available in the

E-DPDCH(s).

� Redundancy Version (RV) controls rate matching.

Transport block

from MAC

Add CRC

attachment

Code block

segmentation

Channel coding


rate matching

Physical channel

segmentation


Channel mapping

Physical channel(s)

Bit

separation

RM_S

Bit

collectionRM_P1

RM_P2

Systematicbits

Parity 1

bits

Parity 2

bits

� The redundancy version (RV), with the parameters s and r, is used to determine the rate

matching parameters.

� s determines whether systematic bits or redundancy bits are prioritized.

� r determines which redundancy version is used.

103

112

001

010

rsE-DCH RV index



E-DCH Interleaving

� Physical Channel Segmentation

� To distribute bits among multiple E-DPDCH

when more than one E-DPDCH is used.

� Interleaving

� The same as UL DCH interleaving

� Channel Mapping

� If more than one E-DPDCH is used, the bits

should be mapped to different E-DPDCHs.

Transport block

from MAC

Add CRC

attachment

Code block

segmentation

Channel coding


rate matching

Physical channel

segmentation


Channel mapping

Physical channel(s)

� The rate matching function determines the number of E-DPDCH channels used to carry the

transport block in one TTI. The following rules are used to determine the number of E-DPDCH channels and the corresponding spreading factors:

� Consider the UE capability and the network limitations.

� Try first to use one E-DPDCH with a smallest spreading factor instead of two E-DPDCH channels with a larger spreading factor.



E-DPDCH Spreading Code

Cch,4,2 if SF = 4

Cch,2,1 if SF = 2E-DPDCH2

Cch,SF,SF/2E-DPDCH11

Cch,4,1E-DPDCH3

E-DPDCH4

Cch,4,1 if SF = 4Cch,2,1 if SF = 2

E-DPDCH2

Cch,SF,SF/4 if SF >= 4

Cch,2,1 if SF = 2E-DPDCH10

Spreading CodeE-DPDCHKNmax-dpdch

� The spreading codes used for a given TTI depends on the number of E-DPDCH channels

for the TTI. The table in the slide shows the code mapping for E-DPDCH. Nmax-dpdch is the maximum number of uplink DPDCHs.



E-DPDCH I/Q Channel Mapping

E-DPDCH1

E-DPDCHK

E-DPCCH

. . .

Channelization

codeGain factor

∑I + jQ

Scrambling

code

IQk

QE-DPDCH2

IE-DPDCH1YES1

IE-DPDCH2

QE-DPDCH1NO1

QE-DPDCH4

IE-DPDCH3

QE-DPDCH2

IE-DPDCH1NO/YES0

IQkE-DPDCHKHS-DSCH

configuredNmax-

dpdch

� E-DPDCHk is mapped to I brand or Q

brand according to IQk.

� E-DPCCH is always mapped to I

branch.

� When E-DCH is configured, at most one uplink DPDCH can be configured. When E-DCH is

not configured, the standard allows up to six uplink DPDCH to be used simultaneously.

� When no uplink DPDCH is configured, the standard allows up to four simultaneous E-

DPDCHs. When one uplink DPDCH is configured, only up to two simultaneous E-DPDCHsare allowed by the standard. When uplink DPDCH is configured, the two E-DPDCHschannel mapping is affected by whether HS-DSCH (hence HS-DPCCH) is configured or not.

� Gain factor means the power of the channel.



E-DPCCH

� E-DPCCH is always transmitted on uplink with E-DPDCH.

� Always transmitted with E-DPDCH simultaneously.

� E-DPCCH includes:

� RSN: Uplink HARQ transmission number

� E-TFCI: E-DCH transport format combination indicator

� Happy Bit: for support of scheduling

� Channelization code for E-DPCCH is Cch,256,1

� Always mapped to I branch



E-DPCCH Coding

� Data in one E-DPCCH subframe

� RSN: 2 bits

� E-TFCI: 7 bits

� Happy bit: 1 bit

� For 10ms TTI, the same coded bit

sequence is transmitted in 5 sub-

frames.

Multiplexing

Channel Coding

Physical channel mapping

one E-DPCCH subframe

RSN E-TFCI Happy bit

10 bits

30 bits



E-DPCCH Coding (continued)

� RSN bits in E-DPCCH are used to indicate the type of redundancy

version (RV) of each HARQ transmission and to aid in soft buffer

management at the NodeB.

� RSN = 0: First transmission

� RSN = 1: Second transmission

� RSN = 2: Third transmission

� RSN = 3: Additional transmission

� RV selection rules:

� UTRAN can configure the UE to use RV = 0 for all transmissions.

� Or UTRAN can configure the UE to use RSN to change RV.

� Retransmission sequence number (RSN) bits in the E-DPCCH indicate the type of

redundancy version (RV) of each HARQ transmission, and aid in soft combination at the

NodeB.

� The following table shows the exact relationship between RSN and RV index.

� Nsys is the number of systematic bits for transmission per TTI.

� Ned is the total number of bits for transmission per TTI.

Nsys/Ned >= 1/2Nsys/Ned < 1/2

Depends on CFN and

subframe number

Depends on CFN and

subframe number

3Nth

2023rd

3212nd

0001st

E-DCH RV indexE-DCH RV index

RSN

value

TX



E-DPCCH & E-DPDCH Frame Format

Slot 0 Slot 1 Slot 3 Slot i Slot 14

10 bits

Data, Ndata bits

1 subframe = 2ms

1 frame = 10ms

E-DPDCH

E-DPCCH

2560 chips

2560 chips, Ndata = 10*2kbits (k = 0…7)

� E-DPDCH and E-DPCCH and both frame-aligned with the uplink DPCCH.



E-AGCH

� E-AGCH is a common downlink channel.

� Fixed data rate: 30kbps

� QPSK modulation

� Spreading factor: 256

� E-AGCH carries absolute grant for E-DCH for all UEs in the cell.

� Transmission on E-AGCH can be 2ms or 10ms.

� 2ms if E-DCH TTI is 2ms

� 10ms if E-DCH TTI is 10ms

� UE listens to the E-AGCH from the serving cell only.

� The E-AGCH carries 5-bit absolute grant and 1-bit scope information. Absolute grant scope

determines whether the grant applies to current HARQ process only or to all HARQ

processes:

� Xags = 1: Per HARQ process

� Xags = 0: All HARQ processes

� The actual grant is the granted T/P ratio as given in the table below.

10(38/15)^221(134/15)^2

0INACTIVE11(42/15)^222(150/15)^2

1ZERO_GRANT12(47/15)^223(168/15)^2

2(7/15)^213(53/15)^224(95/15)^2*4

3(11/15)^214(60/15)^225(150/15)^2*2

4(15/15)^215(67/15)^226(119/15)^2*45(19/15)^216(75/15)^227(134/15)^2*4

6(24/15)^217(84/15)^228(150/15)^2*4

7(27/15)^218(95/15)^229(168/15)^2*4

8(30/15)^219(150/15)^230(150/15)^2*6

9(34/15)^220(119/15)^231(168/15)^2*6

IndexAbsolute grant

valueIndex

Absolute grant value

IndexAbsolute

grant value



E-AGCH Coding� Multiplexing

� 5 bits for the absolute grant values

� 1 bit (Xags) for the scope of the grant

� CRC

� 16 bits CRC is masked with E-RNTI

� E-RNTI is used to address UE.

� Channel Coding

� Rate 1/3 convolutional coding

� Rate Matching

� Puncturing to get 60 bits from 90 bits generated after

channel coding

� Physical Channel Mapping

� 60 bits mapped to one subframe (20 bits per slot)

� For 10ms TTI, same bits get repeated for all 5 subframe

Multiplexing

ID specific

CRC attachment

Channel coding

Rate matching

Physical channel

mapping

5 bits grant 1 bit scope

One E-AGCH subframe

6 bits

22 bits

90 bits

60 bits

� The 6 information bits plus 16 CRC bits, along with 8 tail bits, result in (6+16+8) = 30 bits

at the input of channel coding encoder. The 1/3 convolution encoder output 30*3 = 90

coded bits. Through rate matching, 30 bits are punctured, resulting in 60 coded bits to

form an E-AGCH subframe.



E-AGCH Frame Format

Slot 0 Slot 1 Slot 3 Slot i Slot 14

20 bits

1 subframe = 2ms

1 frame = 10ms

E-AGCH

2560 chips

� E-DPDCH and E-DPCCH and both frame-aligned with the uplink DPCCH.



E-HICH

� E-HICH is a dedicated downlink channel that carries HARQ ACK/NACK.

� QPSK modulation

� Spreading factor is 128 and the channelization code for E-HICH is same

with E-RGCH.

� Transmitted from all cells in the E-DCH active set.

� ACK/NACK is indicated using a binary indicator.

� ACK is +1.

� NACK from cells in serving E-DCH radio link set is -1.

� NACK from cells not in serving E-DCH radio link set is 0 (DTX).



E-RGCH & E-HICH Coding� Same channelization code Cch,128,k

� Different signature sequences, Css,40,m(i) and Css,40,n(i) for slot i

S

/p

Q

PS

K

∑∑∑∑1/0/-1(UP/HOLD/DOWN)

Css,40,m(i)

40 bits/slot

j

Cch,128,k

Scrambling GRGCH

S

/p

Q

P

SK

∑∑∑∑1/(-1 or 0)(ACK/NACK)

Css,40,n(i)

40 bits/slot

j

Cch,128,k

Scrambling GHICH

∑∑∑∑

� E-RGCH and E-HICH share the same channelization code, which is assigned by RRC

signaling. On E-HICH, the HARQ ACK/NACK is multiplied by a 40-bit signature sequence

Css,40,n(i), resulting in a 40-bit sequence to be transmitted. On E-RGCH, the relative grant

UP/HOLD/DOWN is multiplied by a different 40-bit signature sequence Css,40,m(i), resulting in

a different 40-bit sequence. The signature sequence is one of the orthogonal signature

sequences.



Channel Configuration

� E-DCH can be established in combination with the following

downlink configurations:

� Downlink DCH only

� HS-DSCH only

� Both DCH and HS-DSCH

� The following uplink configuration are possible:

� Uplink DCH only

� E-DCH only

� Both uplink DCH and E-DCH

� Downlink and uplink configurations can be combined independently.



Uplink Channel Configuration with

HSUPA

141-E-DCH only

1211DCH + E-DCH

--16DCH only

E-DPCCHE-DPDCHHS-DPCCHDPDCHConfigurations

� The maximum number of each type of channels for each possible

uplink channel configuration except for DPCCH

� Uplink DPCCH is always transmitted, even when DPDCH is not configured but E-DPDCH

and E-DPCCH are configured.

� E-DPCCH is always transmitted with E-DPDCH except when E-DPDCH is in DTX.

� The maximum number of E-DPDCH determines whether DCH is configured or not.

� The existence of HS-DPCCH does not affect the maximum number of E-DPDCH channels,

but it can affect the channelization mapping of E-DPDCH.



Contents


2. HSUPA Concepts





MAC Operation for HSUPA

1. Scheduling information

received from UE

2. Serving grant

determination 5. HARQ

6. De-multiplexing

7. Re-ordering

3. Multiplexing & transport

block size determination

4. HARQServing grant (T/P)

Generate request for

scheduling

UplinkInterference

determination

Transmit E-DCH transport

block, may also include

request for scheduling

Transmit

ACK/NACKTransmit grants (T/P ratio) to

the UE

Extracted

scheduling

information

DTCH DTCH

DTCH DTCH

MAC at

UTRAN

MAC at

UE

Deliver transport block

� MAC operation for HSUPA can be described as follows:

� The NodeB receives scheduling information from the UE. The scheduling

information from UE includes status of data buffers at the UE. It also includes an

indication of whether the UE thinks it has adequate resources.

� The NodeB determines a grant to allocate to the UE based on the uplink

interference measurement and scheduling request from the UE.

� Based on the allocated grant by the NodeB, the UE generate a transport block.

Data from multiple flows can be multiplexed during this step.

� The transport block is delivered to the UE HARQ process for transmissions.

� The HARQ process at the NodeB may request for retransmission.

� Demultiplexing of the transport block takes place on separate flow.

� A reordering function is performed in RNC to reorder PDUs received out of order,

and data is finally delivered to RLC.



HSUPA Protocol Stack

MAC-d

MAC-es

E-DCH FP

TNL

MAC-e

TNLPHY

E-DCH FP

MAC-d

MAC-es/MAC-e

PHY

DCCHDTCH

UE

DCCHDTCH

NodeBUu Iub SRNCIur

TNL TNL

DRNC

TNL: Transport Network Layer

DRNC: Drift RNC SRNC: Serving RNC

FP: Frame Protocol

DCCH: Dedicated Control ChannelDTCH: Dedicated Traffic Channel

� As part of HSUPA, MAC-e and MAC-es are added to the protocol stack in Release 6:

� The MAC-e, terminated in the UE and in the NodeB, to provide fast retransmission

by HARQ mechanism

� The MAC-es, terminated in the UE and in the serving RNC, for reordering function

� On the UE side, MAC-e/MAC-es is implemented in the terminal; however, on the UTRAN

side MAC functionality is split between NodeB (MAC-e) and RNC (MAC-es). Such an

architecture is necessary because soft handover is allowed for HSUPA.



UE MAC-e/es Architecture for HSUPA

HARQ

From MAC-d

Grants on

E-AGCH/E-RGCH(s) ACK/NACK

on E-HICH

E-DCH

Transport Block RSN, E-TFC for

E-DCH on E-DPCCH

MAC-e/es

E-TFC

SelectionMultiplexing and TSN setting

Control

Data

� HARQ: HARQ is responsible for storing MAC-e and performing retransmission functionality. RRC

configures the HARQ profile, which consists of power offset and the number of maximum

retransmissions. The HARQ entity provides the E-TFC, the Retransmission Sequence Number (RSN),

and the power offset to be used by Layer 1 (L1).

� Multiplexing and TSN setting: MAC-e/es architecture allows multiplexing of multiple MAC-d flows

into a single MAC-e PDU. This functionality is performed by Multiplexing and the Transmission

Sequence Number (TSN) setting entity. Multiple MAC-d PDUs from a single logical channel are

concatenated into MAC-es PDUs; one or multiple MAC_es PDUs can be concatenated into a single

MAC-e PDU. The MAC-e PDU thus created is transmitted in the next TTI, as instructed by the E-TFC

selection function. It is also responsible for managing and setting the TSN per logical channel for

each MAC-es PDU.

� E-TFC selection: This entity is responsible for E-TFC selection according to the scheduling

information (Relative Grants and Absolute Grants) received from UTRAN via L1, and for arbitration

among the different flows mapped on the E-DCH. RRC provides the detailed configuration of the E-

TFC entity, over the MAC-Control Service Access Point (SAP). The E-TFC selection function controls

the multiplexing function.



UTRAN MAC-e Architecture for HSUPA

MAC-d flow

Transmission of Grants on

E-AGCH and E-RGCH

ACK/NACKon E-HICH

E-DCH

Transport BlockRSN, E-TFC for

EDCH on E-DPCCH

MAC-e

De-multiplexing

Control

Data

HARQ

E-DCH

Scheduling

E-DCH

Control

� UTRAN MAC-e entity resides in NodeB

� At the UTRAN, MAC is split into MAC-e (in Node B) and MAC-es (in SRNC) entities. There

is a single MAC-e entity per UE in every Node B.

� E-DCH Scheduling: This entity manages E-DCH cell resources between UEs. Based

on scheduling requests received from UE and QoS for various applications,

Scheduling Grants are determined and transmitted. Transmission of grants takes

place on E-AGCH/E-RGCH(s).

� E-DCH Control: This entity is responsible for reception of scheduling requests and

transmission of Scheduling Grants.

� De-multiplexing: This entity provides de-multiplexing of MAC-e PDUs, which may

have been multiplexed into a single MAC-e PDU in the UE for transmissions. MAC-

es PDUs are forwarded to the associated MAC-d flow.

� HARQ: One HARQ entity can support multiple instances (HARQ processes) of stop

and wait HARQ protocols. Each process is responsible for generating ACKs or

NAKs indicating delivery status of E-DCH transmissions. ACK/NAKs are transmitted

on EHICH.



UTRAN MAC-es Architecture for HSUPA� UTRAN MAC-es entity resides in RNC

MAC-d flow

FromMAC-e in NodeB #1

MAC-es

Disassembly

Reordering/

Combining

Disassembly

Reordering/

Combining

Reordering Queue

Distribution

MAC-d flow #1 MAC-d flow #n

Disassembly

FromMAC-e in NodeB #n

Reordering Queue

Distribution

Reordering/

Combining

� At the UTRAN, MAC is split into MAC-e (in Node B) and MAC-es (in SRNC) entities. There

is a single MAC-es entity per UE in RNC.

� Reordering Queue Distribution: Multiple logical channels may be multiplexed into

a single MAC-d flow. The reordering queue distribution function routes the MAC-

es PDUs belonging to a single logical channel to the appropriate reordering buffer.

� Reordering/Combining: MAC-es PDUs received at the reordering entity are already

separated based on the MAC-d flow and the logical channel using DDI. These

received MAC-es PDUs are reordered according to the received TSN. MAC-es PDUs

with consecutive TSNs are delivered to the disassembly function upon reception.

Mechanisms for reordering MAC-es PDUs received out-of-order are left up to the

implementation. There is one Re-ordering Process per logical channel. In the case

of soft handover, MAC-es PDUs may arrive from multiple NodeBs, which are part

of the E-DCH Active Set. The combining function is performed in the MAC-es.

� Disassembly: This function extracts MAC-d PDUs from MAC-es PDU and delivers

those PDUs to the MAC-d flow.



Hybrid ARQ (HARQ) Operation

� N-channel Stop-and-Wait (SAW)

protocol, with 4 processes for 10

ms TTI and 8 processes for 2 ms

TTI for continuous transmission.

� Synchronous retransmission, fixed

timing relation means no need

for process ID.

� Separate HARQ feedback is

provided per radio link.

NodeBNodeB

E-DCH active set

UE

� The Hybrid ARQ for HSUPA consists of an N-channel stop-and-wait protocol.

� The number of HARQ processes is 4 for 10ms TTI and 8 for 2ms TTI respectively. The

number of processes required depends on HARQ round trip time (RTT).

� The retransmission is synchronous with separate feedback provided for each radio link.

� The procedure is as follows:

� The UE transmits the data of the corresponding HARQ process to all NodeB for

which a radio link exists.

� If an acknowledge (ACK) is received ,the transmission is successful.



HARQ Operation and RSN

� HARQ feedback is transmitted on E-HICH.

� E-DPCCH transmits 2-bit retransmission sequence number (RSN).

� RSN identifies HARQ transmission sequence number.

� Incremented at every retransmission and stop at 3.

� RSN of 0 indicates a new transmission.

� The example above is for 2ms TTI. As mentioned earlier, 2ms TTI means 8 HARQ processes

are needed for continuous transmission. In this example, two retransmissions are needed

for successful transmission of the transport block as a result of two NACKs. Once the

second transmission is successful, UE transmit a new transport block on E-DPDCH. The RSN

is set to 0 on E-DPCCH to indicate the new transmission to the NodeB.



HARQ Operation and RSN (continued)� Maximum number of retransmission allowed by HARQ process is configured by

RRC.

� If number of retransmission exceeds 4, RSN of 3 is used for further

retransmission.

� Connection Frame Number (CFN) is used to distinguish between these retransmission.

� For 2ms TTI, subframe number is used, in addition to CFN.

� The example above is for 10ms TTI. In this example the number of retransmission exceeds

3. Because RSN in only 2 bits, a mechanism is needed to distinguish between retransmitted

PDUs beyond 3 retransmissions. This is achieved by using the connection frame number.

Additional CFN information is used to determine the redundancy version (RV). For 2ms TTI,

CFN is not enough, so the subframe number is used in addition to the CFN to distinguish

retransmissions.



HARQ Feedback

� One E-HICH is set up per radio

link.

� The E-HICH information sent by

the cells belonging to the same

RLS is the same and can be

combined.

� Successful transmission is

declared when ACK is received

on any of the E-HICHs.

NodeBNodeB

UE

Non-serving E-DCH

radio link

Serving E-DCH

RLS

Serving E-DCH

cell

UE considers E-DCH

transport block was

transmitted successfully.

� E-HICH is a dedicated channel. Every cell in the E-DCH active set has an E-CHICH set up.

Every cell sends an HARQ ACK/NACK for a decoded transport block (MAC-e PDU). The

same indication (ACK/NACK) is sent by cells that belong to the same NodeB. If any one

cell in the E-DCH active set sends ACK, the UE considers the transmission is successful and

sends next MAC-e PDU on that HARQ process.



How is Scheduling Performed?

� The UE should send a resource request.

� The UE reports scheduling information.

� The UE sends “Happy Bit” on E-DPCCH.

� UE transmit power should be controlled.

� The NodeB signals granted traffic-to-pilot ratio to the UE, which

determines the data rate at which UE can transmit. Through these grants

NodeB can control uplink interference.

� For delay-sensitive application over HSUPA, non-scheduled

transmission is used.

� Non-scheduled transmissions are sent autonomously and do not need a

grant from the NodeB.

� The scheduling mechanism for the uplink needs to control the allocation of the UE transmit

power and uplink interference.



Uplink Scheduling Information

� There are two mechanism for UE to tell the NodeB about

the amount of resources needed.

� Scheduling Information (SI)

� SI can be transmitted with traffic data or alone on the E-DPDCH.

� “Happy Bit”

� “Happy Bit” is transmitted on the E-DPCCH for every E-DCH

transmission

� Scheduling information (SI) includes the following information:

� UPH: UE transmission power headroom, 5 bits

� Used to indicate the Node B of the power ratio of maximum allowed UE transmit

power to DPCCH pilot bit transmit power.

� Then, Node B scheduler knows how much relative power the UE can use for its

data transmission

� TEBS: Total E-DCH buffer status, 5bits

� Reveals the total amount of data in the UE’s transmission buffer.

� This information can be used by the scheduler for deciding the data rate the UE

could actually use.

� Value range from 0 to 31 to indicate different buffer size ranges

� HLID: Highest priority logical channel ID (4bits)

� Indicates the highest priority logical channel that has data in the UE’s transmission

buffer

� HLBS: Highest priority logical channel buffer status (4bits)

� Indicate the amount of data in the buffer for the logical channel indicated by the

HLID

� The HLID and HLBS can be used by the Node B scheduler for deciding which UEs

should be served first or served with higher data rates

� Value range from 0 to 15 to indicate different buffer size ranges

� The UE can use “Happy Bit” to tell the NodeB the resources allocated to the UE is not sufficient and

the UE is capable of transmitting at higher data rate.



Triggering of Scheduling Information

� If serving grant for E-DCH is non-zero:

� SI is triggered, if serving E-DCH cell has changed and it was not

part of previous serving E-DCH RLS.

� Serving grant becomes zero while there is data in buffer.

� SI can also be triggered periodically.

� If there is not grant available for E-DCH transmission:

� SI is trigger if data arrives in empty buffer.

� SI can also be trigger periodically in this scenario.



Reliability Schemes for Scheduling

Information

� If scheduling information (SI) is transmitted without data:

� ACK from non-serving NodeB is ignored.

� If scheduling information (SI) is transmitted with data:

� Scheduling information is retransmitted with next data packet

until ACK is received from serving NodeB or until the number of

retransmission is reached.

� If only non-serving RL send ACK, SI is sent again with new data packet.



“Happy Bit” Calculation

At every E-DCH transmission

Is UE using allavailable serving

gant?

Does UE has

TX power to transmit

at higher rate?

At current grant & rate,

will UE take more than T to

transmit total data?

Yes

YesSet Happy Bit to0 (unhappy)

Set Happy Bit

to 1 (happy)

No

No

No

Yes

T = happy_bit_dalay_condition

This value is configured by RNC

and sent to UE by RRC signaling.

� From the Happy Bit, the NodeB can know whether the resources allocated to the UE are

adequate or not and whether the UE is capable of transmitting at a higher rate.



Scheduled Transmission

� In the downlink, the NodeB allocates scheduling grants to UE to tell UE

the maximum amount of uplink resource it can use.

� Scheduling grants indicates the maximum allowed E-DPDCH/DPCCH power

ratio (T/P ratio).

� Scheduling grants are used by the UE to compute largest permitted

transport block (E-TFC selection).

� There are two types of scheduling grants:

� The absolute grants provide an absolute limitation of the maximum amount of UL

resources the UE may use.

� The relative grants increase or decrease the resource limitation compared to the

previously used value;

� The serving E-DCH cell can allocate the traffic-to-pilot ratio grant as an absolute value on

E-AGCH.

� All the cells in the E-DCH active set can modify the grant value using UP, HOLD or DOWN

commands on E-RGCH. UP commands can be sent only by cells belonging to the serving

RLS.



Absolute Grant

� Absolute grants are carried by the E-AGCH channel of the

serving E-DCH cell.

� The following information is carried by E-AGCH

� UE ID: E-RNTI specific CRC for addressing

� Grant: An index to one of 31 value for traffic-to-pilot ratio.

� HARQ process control: Deactivate/Active HARQ process

� Absolute grant with value 0 is used to deactivate HARQ process

� Scope: Indicates the grant affects one or all HARQ processes

� For 10ms E-DCH TTI, the scope is always set to all processes.



Relative Grant

� A relative grant can be either of the following:

� Serving relative grant

� Non-serving relative grant

� The serving relative grant is transmitted on E-RGCH from all cells in the

serving E-DCH RLS.

� It can be UP, HOLD or DOWN.

� HOLD indicates that there is no new command.

� Each E-RGCH command is associated with a specific HARQ process.

� The amount by which traffic-to-pilot ratio is increased or decreased is

controlled by RRC layer.



Interpreting the E-RGCH command

� When a DOWN command is

received, the grant is always

reduced in step of 1.

� When a UP command is received,

the step of increasing can be 1, 2

or 3. It is controlled by two

parameters “2 step index threshold”

and “3 step index threshold”. These

two parameters are configured by

RNC and sent to UE by RRC

signaling.

(60/15)220

(67/15)221

(75/15)222

(84/15)223

(95/15)224

(106/15)225

(119/15)226

(134/15)227

(150/15)228

(168/15)229

(95/15)2*430

(150/15)2*231

(119/15)2*432

(134/15)2*433

(150/15)2*434

(168/15)2*435

(150/15)2*636

(168/15)2*637

Scheduled GrantIndex

(60/15)220

(67/15)221

(75/15)222

(84/15)223

(95/15)224

(106/15)225

(119/15)226

(134/15)227

(150/15)228

(168/15)229

(95/15)2*430

(150/15)2*231

(119/15)2*432

(134/15)2*433

(150/15)2*434

(168/15)2*435

(150/15)2*636

(168/15)2*637

Scheduled GrantIndex

� The UE maintains a variable (reference_ETPR) that stores the E-DPDCH to DPCCH power ratio used

as reference for relative grant commands. It is set to E-DPDCH to DPCCH power ratio used for the E-

TFC selected for previous TTI on this HARQ process. The value is in terms of an index. This index is

mapped to the actual value of the grant value (E-DPDCH/DPCCH ratio).

� E-RGCH commands may be UP, DOWN, or HOLD. When an up command is received, the increase

can be in steps of 1 or 2 or 3. RRC configures two thresholds to determine which step size to use:

“2 step index threshold” and “3 step index threshold.”

� If the T/P ratio (E-DPDCH to DPCCH power ratio) used for the E-TFC selected for the previous TTI on

this HARQ process is >= 2 step index threshold, the increase is in steps of 1. If the ratio is < (2 step

index threshold) but >= (3 step index threshold), then the increase is in steps of 2. If the ratio is < (3

step index threshold), then the increase is in steps of 3. Thus the configuration can be such that if

the T/P used for the previous TTI was low, the increase in the T/P grant can be set to a larger step

size and the grant allocated to that UE can be increased at faster rate. If the T/P used for previous

TTI was high, the increase can be in smaller steps to avoid a sudden increase in UL interference.

� When a DOWN command is received, the grant is always reduced in steps of 1.

� The slide above shows two examples. In the first example, the serving grant is set to 29 and the UE

receives a DOWN command. The grant is then reduced by 1. In the second example, the serving

grant is set to 21, the UE receives an UP command, and the 3-step-index-threshold is < current T/P.

As a result, the grant is increased by 3.



Interpreting the E-RGCH command

(continued)

� For the example above it is assumed that T/P ratio for previous E-TFC is > 2-step-index-

threshold. As a result, the increase in the serving grant is by a step of 1.



Non-serving Relative Grant

� Non-serving relative grant can

be HOLD or DOWN.

� The use of a non-serving

relative grant allows

neighboring cells to adjust the

uplink transmit power of a UE

which is not under their

control.

Serv

ing E

-RG

CH

UP

Non

-ser

ving

E-R

GC

H

DO

WN

� A non-serving E-DCH cell can use non-serving grants to control the transmit power of the

UEs which are not under its control. A non-serving cell can not increase the serving grant. I

can only decrease it (DOWN) or keep the grant same (HOLD). So the non-serving E-DCH

cell can control the interference generated by sending DOWN commands.

� Consider a scenario where there is only one UE in the serving NodeB and there are several

UEs in adjacent cells. The serving NodeB would allow the only UE to increase is uplink

transmit power. But this may increase interference caused to the adjacent cell that is

already overload. Under such circumstance the non-serving E-DCH cell of the UE will send

DOWN commands to reduce uplink interference.



Rules for Updating Serving Grant

� Serving E-RGCH commands are

ignored for one HARQ RTT is E-

AGCH is received.

� If a DOWN command is received

from the non-serving cell, the UE

ensures for one HARQ RTT that

serving grant is not higher than

the grant considered after

reception of the non-serving E-

RGCH.

� RTT is round trip time. RTT is 8 HARQ processes for 2ms TTI 4 HARQ processes for 10ms

TTI. The example above is for 2ms TTI.



Rules for Updating Serving Grant

(continued)

� T/P for E-TFC in previous TTI for all the HARQ processes is assumed to be N.

� Non-serving command is HOLD during previous TTI for all the HARQ processes.

� T/P is assumed to be < 3-step-index-threshold

� In the example above the TTI is 10ms.

� In the diagram above, the serving grant for HARQ process 4 in not incremented even

though the serving E-RGCH command is UP and non-serving E-RGCH is HOLD. This is

according to the rule described on previous slide. Once a DOWN command is received

from non-serving E-RGCH, it ensures that for one HARQ RTT the serving grant does not

exceed the resulting from this DOWN command.


Thank youwww.huawei.com

2 wcdma hsupa principle

Documents

maximizing

physical channel

radio link

code block

downlink physical

noset happy

enhanced dedicated

uplink physical