3g Ran Nsn Hsdpa Rrm & Parameters

HSDPA RRM & parameters

HSDPA RRM & parameters: Module Objectives At the end of the module you will be able to: •

Explain the physical layer basics of HSDPA technology

•

List the key changes brought by HSDPA and their impact on the network and on the protocol model

•

Explain HSDPA RRM and the related parameters in detail, including packet scheduling, resource allocation, mobility and channel type selection

HSDPA RRM & parameters: Module Objectives At the end of the module you will be able to: •

Explain the physical layer basics of HSDPA technology

•

List the key changes brought by HSDPA and their impact on the network and on the protocol model

•

Explain HSDPA RRM and the related parameters in detail, including packet scheduling, resource allocation, mobility and channel type selection

HSDPA RRM: Contents • • • • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH HSDPA Improvements Other Features Appendix

HSDPA Principles

HSDPAenabled WCEL; 0 = disabled; 1 = enabled

High Speed Downlink Packet Access (HSDPA) based on: • Node B decisions • Multi-code operation • Fast Link Adaptation • Adaptive Modulation & Coding AMC

• Fast Packet Scheduling • Fast H-ARQ • Fast  2 ms TTI* • Downwards Compatibility with R99 • (shared or dedicated carrier) Motivation: - enhanced spectrum efficiency

3GPP Rel. 5; TS 25.308: “HSDPA Overall Description”

- higher peak rates >> 2 Mbps - higher cell throughput - reduced delay for ACK transmission

* TTI = 1 Subframe = 3 Slots = 2 ms H-ARQ: Hybrid Automatic Repeat Request

Principles of DC HSDPA •

Dual-Cell HSDPA of 3GPP Rel8 uses two adjacent WCDMA carriers (same bandwidth) to transmit data for a single UE

•

Can be used with MIMO 2x2 and/or 64QAM

•

DC HSDPA UEs are assigned HS-PDSCHs in the primary serving cell & Secondary Serving High Speed Cell (SSHSC)

•

UL (CQI, ACK/NACK) for DC HSDPA UEs via primary serving cell (no UL in SSHSC)

•

Besides HS-DSCH the primary serving cell is carrying – –

The full set of control & comm on control channels UL transport channels E-DCH HS-DPDCH + optional DC HS-DPCCH (HSUPA UEs)

• SSHSC is left clean from control signaling (max. HS-DSCH capacity) – Among common channels only CPICH is used in SSHSC – E-AGCH, E-RGCH, E-HICH in SSHSC existent but not used by DC HSDPA

Primary serving cell DCellHSDPAEnabled WCEL; 0 = disabled; 1 = enabled

f1

f2 SSHSC

Adaptive Modulation & Coding (1/2) HSDPA uses • QPSK • 16QAM • 64QAM*

16QAM

dynamically based on quality of the radio link

QPSK 2-Bit Keying

4-Bit Keying

0100

1100

0101

1101

Q

1000

0000

Q (0,1)

(1,1)

1001

0001

I

I (0,0)

(1,0)

* defined in 3GPP Rel. 7 / implemented with NSN RU20

0111

1111

1011

0011

0110

1110

1010

0010

Adaptive Modulation & Coding (2/2)

Turbo Coding 1/3 HSDPA Adaptive Coding • based on the R’99 1/3 Turbo Coding • Rate Matching: Puncturing or Repetition  code rate: 1/6 – 4/4

Rate Matching Puncturing / Repetition

• dynamically based on quality of the radio link

Effective Code Rate: 1/4 - 3/4

Multi Code Operation (1/3) SF = 1

2

4

8 C8,0 = [11111111]

C4,0 = [1111] C8,1 = [1111-1-1-1-1]

SF = 16

...

256

512

C16,0 = [.........] C16,1 = [.........] C16,2 = [.........] C16,3 = [.........]

C2,0 = [11] C8,2 = [11-1-111-1-1] C4,1 = [11-1-1] C8,3 = [11-1-1-1-111] C1,0 = [1] C8,4 = [1-11-11-11-1] C4,2 = [1-11-1] C8,5 = [1-11-1-11-11] C2,1 = [1-1] C8,6 = [1-1-111-1-11] C4,3 = [1-1-11] C8,7 = [1-1-11-111-1]

C16,4 = [.........]

SF = 16

C16,5 = [.........]

240 ksymb/s

C16,6 = [.........] C16,7 = [.........] C16,8 = [.........] C16,9 = [.........] C16,10 = [........] C16,11 = [........] C16,12 = [........] C16,13 = [........] C16,14 = [........] C16,15 = [........]

Multi-Code operation:

1..15 codes 0.24 .. 3.6 Msymb/s

Multi Code Operation (2/3) Modulation

QPSK

16QAM

64QAM

Coding rate

5 codes

10 codes

15 codes

1/4

600 kbps

1.2 Mbps

1.8 Mbps

2/4

1.2 Mbps

2.4 Mbps

3.6 Mbps

3/4

1.8 Mbps

3.6 Mbps

5.4 Mbps

2/4

2.4 Mbps

4.8 Mbps

7.2 Mbps

3/4

3.6 Mbps

7.2 Mbps

10.8 Mbps

4/4

4.8 Mbps

9.6 Mbps

14.4 Mbps

3/4

5.4 Mbps

10.8 Mbps

16.2 Mbps

5/6

6.0 Mbps

12.0 Mbps

18.0 Mbps

4/4

7.2 Mbps

14.4 Mbps

21.6 Mbps

RU20 includes 3GPP Rel. 7 features: • 64QAM (RAN1643)

HSDPA64QAMAllowed WCEL; 0 (Disabled), 1 (Enabled)

64QAM 6 bits/symbol

Multi Code Operation (3/3): HSDPA UE capability classes RU20/30 include 3GPP Rel. 7/8 features: • 64QAM (cat 13, 14, 17, 18) • 2x2 MIMO (Dual-Stream MIMO) (cat 15, 1 6, 17, 18) MIMO w/- 64QAM (cat 19, 20) • DC-HSDPA (cat 21, 22) • DC –HSDPA w/- 64QAM (cat 23, 24)

RU40 include 3GPP Rel.9 features: • DC –HSDPA w/-MIMO w/o 64QAM (cat 25, 26) • DC –HSDPA & MIMO & 64QAM (cat 27, 28) HSDPA64QAMAllowed

Modulation

DualStream MIMO supported

Peak Rate

3 (6 ms)

QPSK/16QAM

No

1.2 Mbps

5

3

QPSK/16QAM

No

1.2 Mbps

3

5

2 (4 ms)

QPSK/16QAM

No

1.8 Mbps

4

5

2

QPSK/16QAM

No

1.8 Mbps

5

5

1 (2 ms)

QPSK/16QAM

No

3.6 Mbps

6

5

1

QPSK/16QAM

No

3.6 Mbps

7

10

1

QPSK/16QAM

No

7 Mbps

8

10

1

QPSK/16QAM

No

7 Mbps

9

15

1

QPSK/16QAM

No

10 Mbps

10

15

1

QPSK/16QAM

No

14 Mbps

11

5

2

QPSK only

No

1 Mbps

12

5

1

QPSK only

No

1.8 Mbps

13

15

1

QPSK/16QAM/ 64QAM

No

17.4 Mbps

14

15

1

QPSK/16QAM/ 64QAM

No

21.1 Mbps

15

15

1

QPSK/16QAM

Yes

23.4 Mbps

16

15

1

QPSK/16QAM

Yes

28 Mbps

17

15

1

QPSK/16QAM/ 64QAM or Dual-Stream MIMO

17.4 or 23.4 Mbps

18

15

1


21.1 or 28 Mbps

max. No. of HS-DSCH Codes

min. * Inter-TTI interval

1

5

2

HS- DSCH

category

WCEL; 0 (Disabled), 1 (Enabled)

MIMOEnabled WCEL; 0 (Disabled), 1 (Enabled)

Further details on HS-DSCH categories & other parameters HSPA+ RRM * TTI: Transmission Time Interval

Network Modifications for HSDPA new Node B functionalities: • Acknowledged transmission: Fast H-ARQ UTRAN & UE: • modified PHY layer • modified MAC

faster retransmission / reduced delays ! less Iub retransmission traffic ! higher spectrum efficiency !

• Fast Packet Scheduling fast & efficient resource allocation !

• Fast Link Adaptation Adaptive Modulation & Coding ! compensation of fast fading (without fast PC) higher peak rates & spectrum efficiency !

Uu • modified transport and physical channels

Iub

RNC: functionalities shifted to Node B

r e t r R a n e s d m u i c e s s d i o n

• modified coding • modified modulation

Node B „more intelligence“ new functionalities

UE new UEs

 HSDPA

Capability Classes

HSDPA RRM • • • • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection and Switching Associated UL DCH HSDPA Improvements Other Features Appendix

HSDPA Protocol Model RLC

RLC DCH

MAC-d

R99 PHY

UE

HSDPA (R5) HSDPA (R7)

DPCH

Uu

RLC

PHY UE

DCH FP

DCH FP

TNL

TNL

Node B

Iub

HS-DSCH HS-PDSCH Uu

RNC

RLC

MAC-d flow

MAC-d MAC-hs MAC-ehs

PHY

MAC-d

MAC-d

MAC-hs MAC-ehs

HS-DSCH FP

HS-DSCH FP

PHY

TNL

TNL

Node B

Iub

(e)hs: (enhanced) high speed TNL : Transport Network Layer

RNC

Concepts of MAC Layer, MAC-hs & MAC-ehs MAC: Medium Access Control MAC • •

25.321

Mapping of logical channels onto transport channels Multiplexing of multiple logical channels onto a single transport channel, e.g. of 4 signalling radio bearers (SRB) onto single DCH

• • •

 TS

Complete MAC multiplexing for user plane data currently not supported Multiplexing requires the addition of a MAC header

MAC entities on network side distributed between RNC and Node B

MAC-hs • •

•

•

supports HSDPA with 3GPP Rel. 5 Tasks of MAC-hs within the Node B • Flow control (see section packet scheduling) • Packet scheduling (see section packet scheduling) • H-ARQ (see section layer 1 re-transmission) • Transport format selection (see section link adaptation) Tasks of MAC-hs within the UE • HARQ (see section layer 1 re-transmission) • Disassembly of transport blocks • Re-ordering Header & payload • Payload: Concatenating of one or more MAC-d PDU into single MAC-hs PDU • Header: 21 bits assuming single MAC-d PDU size

MAC-ehs • •

supports enhanced HS-DSCH functions of 3GPP Rel. 7 - 9 must be configured to support features such as: 64QAM (RAN1643), MIMO (RAN 1642), flexible RLC (RAN1638), Dual-Cell HSDPA (RAN1906)

Physical Channel Overview

HS-PDSCH High-Speed Physical DL Shared Channel

HS-SCCH High Speed Shared Control Channel

HS-DPCCH High Speed Dedicated Physical Control Channel

Node B

associated DCH Dedicated Channel (Rel. 99)

MAC-hs

F-DPCH Fractional Dedicated Physical Channel (Rel. 6/7)

HS-PDSCH • HS-PDSCH: High-Speed Physical Downlink Shared Channel • • • • •

•

Transfer of actual HSDPA data 5 - 15 code channels QPSK or 16QAM modulation Divided into 2 ms TTIs Fixed SF16

HS-PDSCH code set parameter •

Specifies whether number of codes channels reserved for HSDPA is fixed* or dynamically adjustable

• •

Minimum 5 code channels / Maximum 15 codes channels Possible numbers of code channels enabled / disabled bit wise

HSPDSCHCodeSet HS-PDSCH code set; WCEL; (-) (-) (5 codes) Examples 00000 00000 100000 = always 5 codes reserved (default) 11010 10100 100000 = number of reserved codes adjustable (5, 8, 10, 12, 14 or 15 codes, recommended) 11-15 codes

6-10 codes

0-4 codes always disabled

HS-SCCH (1/2) • HS-SCCH: High-Speed Shared Control Channel •

L1 Control Data for UE; informs the UE how to decode the next HS -PDSCH frame e.g. UE Identity, Channelization Code Set, Modulation Scheme, TBS, H-ARQ process information Fixed SF128 transmitted 2 slots in advance to HS-PDSCHs NSN implementation with slow power control: shares DL power with the HS-PDSCH more than 1 HS-SCCH required when code multiplexing is used

• • • •

MaxNbrOfHSSCCHCodes

• Code multiplexing

Maximum number of HS-SCCH codes

WCEL; RU10 & earlier: 1..3; 1; 1; RU20: 1..4

SF16 HS-PDSCH 15

User 1

User 2

Subframe 2 ms

User 3

•

HSDPA service for several users simultaneously

•

For each user individual HS-SCCH required

•

available only, if > 5 codes can be reserved for HS-PDSCH

User 4

10

5

TBS: Transport Block Size

Time

HS-SCCH (2/2)

+15 x SF16 HS-PDSCH

SF16

SF32

32

SF64

64

64

64

S-CCPCH1 SF128

SF256

128

256

256

128

128

S-CCPCH2

HS-SCCH

256

256

256

256

256

CPICH AICH P-CCPCH PICH 128

128

Allocated CC

Blocked CC

FACH-s: for Service Area Broadcast (CTCH)

128

Available CC

128

256

128

256

256

256

128

128

HS-SCCH

HS-SCCH

256

256

256

256

HS-DPCCH • UL HS-DPCCH: High-Speed Dedicated Physical Control Channel • • •

MAC-hs Ack/Nack information (send when data received) Channel Quality Information (CQI reports send every 4ms, hardcoded period) Fixed SF 256

1 Slot = 2560 chip HARQ-ACK (10 bit)

2 Slots = 5120 chip CQI (20 bit) Channel Quality Indication

1 HS-DPCCH Subframe = 2ms

Subframe # 0

Subframe # i

TS 25.21: CQI values = 0 (N/A), 1 .. 30; steps: 1; 1 indicating lowest, 30 highest air interface quality

Subframe # N

HS-DPCCH & CQI

CQI

P-CPICH UE observes P-CPICH (Ec/Io)  CQI*

TB Size

# codes Modulation



1

137

1

QPSK

0

2

173

1

QPSK

0

3

233

1

QPSK

0

4

317

1

QPSK

0

5

377

1

QPSK

0

6

461

1

QPSK

0

7

650

2

QPSK

0

8

792

2

QPSK

0

9

931

2

QPSK

0

10

1262

3

QPSK

0

11

1483

3

QPSK

0

12

1742

3

QPSK

0

13

2279

4

QPSK

0

14

2583

4

QPSK

0

15

3319

5

QPSK

0

CQI used for:

16

3565

5

16-QAM

0

• Link Adaptation decision • Packet Scheduling decision

17

4189

5

16-QAM

0

18

4664

5

16-QAM

0

19

5287

5

16-QAM

0

ACK/NACK used for:

20

5887

5

16-QAM

0

21

6554

5

16-QAM

0

22

7168

5

16-QAM

0

• H-ARQ process • Link Adaptation decision • HS-SCCH power adaptation * UE internal (proprietary) process TB Size [bit] CQI value 0: N/A (Out of range)  = Reference Power Adjustment (Power Offset) [dB]

CQI Table (Example) TS 25.214: Annex Table 7b

Cat 8 UE

23

9719

7

16-QAM

0

24

11418

8

16-QAM

0

25

14411

10

16-QAM

0

26

14411

10

16-QAM

-1

27

14411

10

16-QAM

-2

28

14411

10

16-QAM

-3

29

14411

10

16-QAM

-4

CQI Tables TS 25.214: Annex Table 7d Cat 10 UE

TS 25.214: Annex Table 7f Cat 27 UE

TS 25.214 Annex Table 7g Cat 14 UE: CQI29: 14 Codes; 32257 bit CQI30: 15 Codes; 38582 bit

CQI

TB Size

# codes Modulation

CQI



TB Size

# codes Modulation



1

137

1

QPSK

0

1

136

1

QPSK

0

2

173

1

QPSK

0

2

176

1

QPSK

0

3

233

1

QPSK

0

3

232

1

QPSK

0

4

317

1

QPSK

0

4

320

1

QPSK

0

5

377

1

QPSK

0

5

376

1

QPSK

0

6

461

1

QPSK

0

6

464

1

QPSK

0

648

2

QPSK

0

7

650

2

QPSK

0

7

8

792

2

QPSK

0

8

792

2

QPSK

0

9

931

2

QPSK

0

9

928

2

QPSK

0

10

1262

3

QPSK

0

10

1264

3

QPSK

0

11

1483

3

QPSK

0

11

1488

3

QPSK

0

12

1742

3

QPSK

0

12

1744

3

QPSK

0

13

2279

4

QPSK

0

13

2288

4

QPSK

0

14

2583

4

QPSK

0

14

2592

4

QPSK

0

15

3319

5

QPSK

0

15

3328

5

QPSK

0

16

3565

5

16-QAM

0

16

3576

5

16-QAM

0

17

4189

5

16-QAM

0

17

4200

5

16-QAM

0

18

4664

5

16-QAM

0

18

4672

5

16-QAM

0

19

5287

5

16-QAM

0

19

5296

5

16-QAM

0

20

5887

5

16-QAM

0

20

5896

5

16-QAM

0

21

6554

5

16-QAM

0

21

6568

5

16-QAM

0

22

7168

5

16-QAM

0

22

7184

5

16-QAM

0

23

9719

7

16-QAM

0

23

9736

7

16-QAM

0

24

11418

8

16-QAM

0

24

11432

8

16-QAM

0

25

14411

10

16-QAM

0

25

14424

10

16-QAM

0

26

17237

12

16-QAM

0

26

15776

10

64-QAM

0

27

21754

15

16-QAM

0

27

21768

12

64-QAM

0

28

23370

15

16-QAM

0

28

26504

13

64-QAM

0

29

24222

15

16-QAM

0

29

32264

14

64-QAM

0

30

25558

15

16-QAM

0

30

32264

14

64-QAM

-2

Associated DCH (DL & UL) • DL DPCH: Associated Dedicated Physical Channel • L3 signalling messages • Speech - AMR • Power control commands for associated UL DPCH

• UL DPCH: (DPDCH & DPCCH) • L3 signalling messages • Transfer of UL data 16 / 64 / 128 / 384 kbps, e.g. TCP acknowledgements

• Speech - AMR DPDCH / DPCCH (time multiplexed) DPDCH: L3 signalling; AMR DPCCH: TPC for UL DPCH power control

DPDCH: L3 signalling, AMR; TCP ACKs; 16 / 64 / 128 / 348 kbps

DPCCH: TPC, Pilot, TFCI

Fractional DPCH: F-DPCH (DL) The Fractional DPCH (F-DPCH):

• introduced in 3GPP Rel. 6 (enhanced in Rel. 7; NSN RU20 implementation based on Rel. 7) • replaces the DL DPCCH • includes T r a n s m i t P o w e r C o n t r o l ( TP C ) b i t s but excludes TFCI & Pilot bits & SRB – TFCI bits - no longer required as there is no DPDCH – Pilot bits - no longer required as TPC bits are used for SIR measurements – SRB mapped to E-DCH & HS-DSCH

• increases efficiency by allowing up to 10 UE to share the same DL SF256 channelization code -

time multiplexed one after another

• RU20 feature RAN1201 ; – – – –

requires Rel. 7 or newer UE HSDPA & HSUPA must be enabled Feature is licensed using an RNC ON/OFF license License CPC exists and its state is ON

FDPCHEnabled WCEL; 0 (Disabled), 1 (Enabled)

1 time slot 2560 chips 256 chips Tx Off

TPC

Tx Off

Slot #i

HSDPA RRM • • • • • • • • • • • • • •


Summary Characteristic

RU10

RU20

RU30

RU40/RU50

HSDPA users per cell

≤ 64

≤ 72 (RAN1668)

≤ 72

≤ 128 (RAN2124)

Modulation

QPSK/16QAM

QPSK/16QAM & 64QAM (RAN1643)

QPSK/16QAM/64QAM

QPSK/16QAM/64QAM

MIMO

No

Yes (2x2) (RAN1642)

Yes

Yes

Dual-Cell HSDPA

No

Yes (RAN1906)

DC-HSDPA

DC-HSDPA DB DC HSDPA (RAN2179)

Data rate per UE

up to 14 Mbps

up to 42 Mbps

up to 42 Mbps 84 Mbps (RAN 1907)

up to 84 Mbps (RAN1907)

Traffic Classes

Interactive + Background + Streaming

+ CS Voice over HSPA (RAN1689)

all traffic classes

Packet Scheduler

Proportional Fair (PF) + QoS Aware HSPA Scheduling

PF + QoS aware sc heduling

PF + QoS aware scheduling

HSDPA Multi-RAB

multiple RAB HSDPA + AMR

multiple RAB HSDPA + AMR

multiple RAB HSDPA + AMR, +CS64 Conv.

Yes (up to 3)

Yes (up to 4)

Yes (up to 4)

Yes (up to 4)

16, 64, 128, 384 Kbps

16, 64, 128, 384 Kbps

16, 64, 128, 384 Kbps

16, 64, 128, 384 Kbps

Code Multiplexing (Scheduled users per TTI)

UL associated DCH

all traffic classes PF + QoS aware scheduling multiple RAB HSDPA + AMR, +CS64 Conv.

Feature Activation • Most enhanced features must be licensed individually and are activated by setting individual off / on parameter • Some features can be activated on cell level, ot hers on WBTS or even RNC level only HSDPAenabled

HSDPAMobility

WCEL; 0 = disabled; 1 = enabled

Serving HS-DSCH cell change & SHO on/off switch RNFC ; 0 = disabled; 1 = enabled

HSDPA48UsersEnabled RNFC; 0 = disabled; 1 = enabled

HspaMultiNrtRabSupport

HSDPA64UsersEnabled

HSPA multi RAB NRT support WCEL; 0 = disabled; 1 = enabled


HSDPA14MbpsPerUser

HSDPADynamicResourceAllocation

WBTS; 0 = disabled; 1 = enabled

HSDPA Dynamic Resource Allocation RNFC; 0 = disabled; 1 = enabled

HSPAQoSEnabled

HSDPA16KBPSReturnChannel

WCEL; 0..4; 1; 0 = disabled 0 = QoS prioritization is not in use for HS t ransport 1 = QoS prioritization is used for HS NRT channels 2 = HSPA streaming is in use 3 = HSPA CS voice is in use 4 = HSPA streaming & CS voice are in use

RU20/ 30 RU40

HSPA72UsersPerCell WCEL; 0 = disabled; 1 = enabled if enabled, max. 72 HSDPA/HSUPA users can be supported per cell.

HSPA128UsersPerCell WCEL; 0 = disabled; 1 = enabled if enabled, max. 128 HSDPA/HSUPA users can be supported

HSDPA 16 Kbps UL DCH return channel on/off RNFC; 0 = disabled; 1 = enabled

FDPCHEnabled; CPCEnabled WCEL; 0 (Disabled), 1 (Enabled)

HSDPA64QAMAllo wed; MIMOEnabled; DCellHSDPAEnabled; MIMOWith64QAMUsage WCEL; 0 (Disabled), 1 (Enabled)

DCellAndMIMOUsage WCEL; 0=DC-HSDPA and MIMO disabled; 1=DC-HSDPA and MIMO w/o 64QAM enabled; 2=DC-HSDPA and MIMO with 64QAM enabled

Cell Group Definition • SCHs under the same Node B should not overlap with each other • define for each sector offset relative to BTS frame number with parameter Tcell • Cells with offsets within certain range form one cell group – – – –

Group 1 offset = 0-512 chips Group 2 offset = 768-1280 chips Group 3 offset = 1536-2048 chips Group 4 offset = 2304 chips

BTS reference

BTS reference SCH

0 chips

SCH 256 chips

BTS reference SCH

512 chips

Tcell Frame timing offset of a cell WCEL; 0..2304 chips; 256 chips; no default

HSPA 72 / 128 Users per Cell (1/3) • • • • •

HSPA 72 users/cell: RAN1686 (RU20); HSPA 128 users/cell: RAN2124 (RU40); optional RNC License Key required (On-Off) HSPA72UsersPerCell increases the number of simultaneous HSPA users to 72 / 128 per cell WCEL; 0 = not enabled; 1 = enabled both with dedicated & shared scheduler HSDPA, HSUPA, Dynamic Resource Allocation must be enabled, Continuous Packet Connectivity & F -DPCH are recommended for both RAN1686 & RAN2124, HSUPA DL Physical Channel Power Control recommended for RAN2124 HSPA128UsersPerCell • max. 15 codes allocated (HS-PDSCH code set = 11010 10100 10000) WCEL; 0 = not enabled; 1 = enabled • Code multiplexing (max. no. of HS-SCCH codes MaxNbrOfHSSCCHCodes = 4) • HSDPA 16 Kbps UL DCH should be enabled to avoid UL overload

72 /128 users

72 /128 users

Hardware requirements: • Flexi Node B must have Rel2 or Rel3 system module 72 /128 users

Other parameters may restrict max. number of HSPA users, e.g.: -WCEL: MaxNumberEDCHCell - WBTS: MaxNumberEDCHLCG - WCEL: MaxNumberHSDSCHMACdFlows - WCEL: MaxNumberHSDPAUsers - WCEL: MaxNumbHSDPAUsersS - WCEL: MaxNumbHSDSCHMACdFS

HSPA 72 / 128 Users Per Cell (2/3) DL Code allocation DL Code allocation in a cell depends on activated features and t raffic – if HSPA 72 Users Per Cell or HSPA 128 Users Per Cell is enabled, RNC allocates DL codes according to Maximum number of scheduled HSDPA user per TTI (Code Multiplexing) 

MaxNbrOfHSSCCHCodes ; WCEL; 1..4; 1; 1 (4 is recommended in both cases)

– 1 E-RGCH & E-HICH codes is reserved in cell setup; max number of E -RGCH/E-HICH codes is 4 or not limited 

reserved number of E-RGCH/E-HICH codes depend on number of HSUPA users, TTI (2ms or 10ms), whether the cell is serving or non-serving E-DCH cell to the UE, and scheduled or non-scheduled transmission

– if Paging 24 kbps feature is enabled, more DL codes are needed to separate FACH and PCH traff ic 

; WCEL; on/ o ff & NbrOfSCCPCHs ; WCEL; 1..3; 1; 1 PCH24kbpsEnabled SF 16,0 SF 32 S-CCPCH SF 64

E-RGCH E-HICH

HS-SCCH

SF 128

SF 256 0

1

2

3

4

5

6

depending on FACH / PCH configuration

7

8

9

10

11

12

13

14

15

HSPA 72 / 128 Users Per Cell (3/3) Traffic analysis

MaxNbrOf HSSCCHCodes WCEL; RU10 & earlier: 1..3; 1; 1; RU20: 1..4

• No. of HS-SCCH channels increased to 4 to schedule & control increased number of HSPA users in a cell • DL code space limited  dynamic DL control channel allocation mechanism introduced to maximize available codes for HS-PDSCHs  HSUPA RRM (E-RGCH & E-HICH management / dynamic code allocation) • if code tree resources allocated like on previous slide, following traffic is supported: – – – –

15 codes @ SF16 for HSDPA single user per 2 ms TTI (no code multiplexing) MIMO enabled F-DPCH enabled

• m o s t l i k el y RNC will allocate another SF16 branch to increase control channel traffic  r e d u c i n g H S DP A SF16 co des further

Code allocation in case of 4 HS-SCCH:

HSDPA RRM • • • • • • • • • • • • • •


CQI Reporting & Link Adaptation Remember:

P-CPICH

CQI used for:

UE observes P-CPICH (Ec/Io) CQImeasured*

• Link Adaptation decision • Packet Scheduling decision

ACK/NACK used for: • H-ARQ process • Link Adaptation decision • HS-SCCH power adaptation

Link adaptation algorithm

1) Generation of CQImeasured : – –

CQImeasured*

UE monitors EC /I0 UE reads PHS-PDSCH SIG (L3/RRC signalling)

2) UE reports CQImeasured every 4 ms (NSN solution) – can be increased with Mass Event Eandler 3) CQI Correction in Node B Node B corrects reported CQImeasured to CQIcompensated based on: actual HS-PDSCH power PHS-PDSCH TRUE Number of ACK & NACK

– –

4) Link Adaptation decision: Node B decides about TB size for next sub-frame: – – –

Modulation Coding rate Number of codes

* UE internal (proprietary) process

CQI Compensation (1/3)

 signalled

to UE in case of  HS-DSCH setup  Serving cell change

CQImeasured UE generates CQImeasured assuming Tx power PHS-PDSCH SIG = PCPICH + –

calculated by RNC:

+

= f x Min(PtxMaxHSDPA, PtxMax – PtxNonHSDPA) – PCPICH

PHS-PDSCH SIG = (f x Min(PtxMaxHSDPA, Ptxmax – PtxNonHSDPA)) [dBm] +

CQI Compensation in Node B • Node B compensates CQI from differences between assumed HS-PDSCH Tx power & actual HS-PDSCH Tx power P HS-PDSCH TRUE – Part of HSDPA power used for HS-SCCH – HS-PDSCH power can vary because of dynamic power allocation

• Offset X used to convert reported CQImeasured into compensated CQIcompensated CQIcompensated = CQImeasured + X [dB] X = PHS-PDSCH TRUE – (PCPICH +

+

) – A [dB]

correction A estimated by outer loop link adaptation algorithm = Reference Power Adjustment (Power Offset) [dB]  CQI tables PtxMax = max. cell power PtxNonHSDPA = total power allocated to R99 & DL control channels (latest report is taken) PtxMaxHSDPA = max. allowed HSDPA power f = 0.7 for static HS-PDSCH power allocation f = 0.5 for dynamic HS-PDSCH power

CQI Compensation (2/3) Outer loop link adaptation algorithm

correction A

• If ACK received for first transmission of a packet – Correction A decreased by 0.005 dB – But not below -4 dB (maximum CQI improvement towards higher TBS)

• If NACK received for first transmission of a packet – Correction A increased by 0.05 dB – But not above 4 dB (maximum CQI downgrade towards lower TBS) NACK for 1st transmission P0

lower CQI

increase CQI ACK for 1st transmission time

CQI Compensation (3/3) • CQI compensation makes it difficult to map reported CQI from UE log files into expected HSDPA transport block size TBS

CQIMEASURED = 3 233 bits per TB (167 K) e.g. PHS-PDSCH SIG = 37 dBm e.g. PHS-PDSCH TRUE = 40 dBm X = (40 – 37) dB = 3 dB CQICOMPENSATED = 3 + 3 = 6 461 bits per TB (230 K)

X = 3 dB

Spectral Efficient Link Adaptation • Good radio conditions CQICOMPENSATED, but less data to be sent • Node B determines CQI NEEDED required for actual service • Node B reduces HSDPA transmission power by CQICOMPENSATED - CQINEEDED

Example: CQICOMPENSATED = 10 Actual service 384 K Requires 768 bits per TB CQINEEDED = 8 Power reduction = (10 – 8) dB = 2 dB

RAN1244: Spectral Efficient Link Ad aptation

Measurement Examples (1/2) • CQI improves both with increasing: – EC /I0 – HSDPA power

CQI as a function of CPICH Ec/Io 30

) I Q C ( r o 25 t a c i d n I y 20 t i l a u Q l 15 e n n a h C d 10 e t a s n e 5 p m o C 0

-15

PtxMaxHSDPA = 30 dBm PtxMaxHSDPA = 35 dBm

Reported

PtxMaxHSDPA = 40 dBm

Compensated

-14

-13

-12

-11

-10

-9

CPICH Ec/Io (dB)

-8

-7

-6

-5

Measurement Examples (2/2) • CQI estimation differs from one type of UE to the next one Prediction of different values in spite of identical channel conditions

• CQI compensation capable to remove most of these differences Almost same service experienced in spite of proprietary CQI estimation Prior to compensation

After compensation 30

30

I Q C ( r o 25 t a c i d n I y 20 t i l a u Q l 15 e n n a h C d 10 e t a s n e 5 p m o C

Samsung zx20 25

I Q C ( r o 20 t a c i d n I y 15 t i l a u Q l e 10 n n a h C

Novatel U740

Unloaded

Common Channel Loaded

5

Novatel U740

Unloaded

Common Channel Loaded

0

0 -15

Samsung zx20

-14

-13

-12

-11

-10

-9

-8

CPICH Ec/Io (dB)

-7

-6

-5

-4

-3

-15

-14

-13

-12

-11

-10

-9

-8

CPICH Ec/Io (dB)

-7

-6

-5

-4

-3

HSDPA RRM • • • • • • • • • • • • • •


R99 & HSDPA Retransmission R99 DCH

R5 HS-DSCH

RNC

RNC

Re-transmission

Packet

Packet RLC ACK/NACK BTS

w o l f a t a D

DL control moved to BTS BTS L1 ACK/NACK Re-transmission

Terminal

Terminal

H-ARQ: Hybrid Automatic Repeat reQuest

Hybrid Automatic Repeat Request H-ARQ • H-ARQ Objective: – ensures reliable data transfer between UE and Node B – short Round Trip Time between UE and network

• HSDPA connection re-transmission can originate from: – MAC-hs layer between UE and Node B (HARQ) – RLC layer between UE and RNC – TCP layer between UE and application server

• Re-transmission time out – after 3rd L1 re-transmission HSDPA packet discarded (hardcoded threshold)

•

HARQ algorithms:

– Chase combining CC – Incremental Redundancy IR Algorithm selected by operator on BTS level

HARQRVConfiguration WBTS; 0 = Chase Combining, 1 = Incremental Redundancy

HSDPA RRM • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling – – – –

• • • • • • • •

Scheduling Types: Round Robin & Proportional Fair Scheduling & Code Multiplexing Basics of QoS Aware Scheduling and Application Aware RAN In-bearer Application Optimization

Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection and Switching Associated UL DCH HSDPA Improvements Other Features Appendix

Basic Scheduler Types • Supports packet schedulers – Round Robin RR – Proportional Fair PF (requires individual license) – Type of scheduler set by HSDPA.BB.Resource.Allocation commissioning parameter

Round Robin Scheduler



• assigns sub-frames in rotation – User at cell edge served as frequently as user at cell centre • does not account for channel conditions experienced by UE – Low total throughput in cell



• if no data have to be transferred from Node B to certain UE then the sub-frame is assigned to the next one

Proportional Fair PF Scheduler • Takes into account multipath fading conditions



experienced by UE

TTI 1

TTI 2

TTI 3

TTI 4 Scheduled user

– Improved total throughput in cell in comparison to round robin

• Sub-frames assigned according scheduling metric – Ratio instantaneous data rate / average data rate experienced in the past

– User at cell edge served less frequently as user at cell



centre

E s t i m a t e o f i n s t a n t an e o u s l y s u p p o r t e d u s e r throughput B a s e d o n c o m p e n s a t ed C Q I

TP inst TP a ve C a l c u l a te d a v er a g e u s e r t h r o u g h p u t i n t h e p a s t T h r o u g h p u t m e a s u r e d e v e r y 10 m s w i t h 1 00 m s sliding window

USER 2 Es/N0

USER 1 Es/N0

Scheduling / HSDPA Code Multiplexing HSDPA Code Multiplexing: enables simultaneous transmission of up to 4* HSDPA UEs during 1 TTI – each simultan. served HSDPA UEs must have separate HS-SCCH – ≥ 5 codes must be allocated to HS -PDSCH – MAC-hs entity selects (3) best users (based on PF or QoS aware metric)

MaxNbrOfHSSCCHCodes Max. number of HS-SCCH codes

WCEL; 1..4*; 1; 1 (no Code Multiplexing)

for transmission in the next TTI – HS-PDSCH codes & power resources shared, taking into account:  how much data user has in its buffer  Channel conditions of user

Full buffer

Different data amounts

Amount of data in buffer UE1

RU10 & later 15 codes

UE2

UE3

UE1

7

7

3 2

8

8

10

UE2

UE3

5 10

Codes & power are divided optimally between users depending on data amount.

8 * 3 before RU20

Basics of QoS Aware Scheduling • Shortcomings of standard PF – PF metric does not distinguish between traffic classes – No bit rate guarantee, i.e. no streaming services supported – Interactive service not prioritised against background one

• Idea of QoS aware HSPA scheduling (RAN1262) – QoS aware HSPA scheduling enabled with parameter HSPAQoSEnabled – HSDPA dynamic resource allocation must be enabled – Streaming services  Guaranteed bit rate set by RNC

– Interactive IA & Background BG services  Operator can set nominal bit rate (target minimum bit rate)

If not defined, service treated as best effort one  Operator can set service priorities, so that IA services are scheduled more often than BG ones

Services belonging to same traffic class again scheduled according PF

HSPAQoSEnabled WCEL; 0..4;1; 0 = disabled 0 = QoS prioritization is not in use for HS transport 1 = QoS prioritization is used for HS NRT channels 2 = HSPA streaming is in use (RAN1004) 3 = HSPA CS voice is in use (RAN1689) 4 = HSPA streaming and CS voice are in use

Basics of QoS Aware Scheduling • Guaranteed Bit Rate GBR – Set by RNC for streaming services on basis of the RAB profile

• Nominal bit rate NBR = target minimum bit rate The nominal bit rate NBR is set as the target minimum bit rate in the RNC for NRT HS-DSCHs. – Can be specified by operator for NRT services  Individually for each SPI 0..12 and Individually for UL and DL  If Application Aware RAN is enabled SPI is dynamically modified by the RNC PDCP layer but the new NBR value

corresponding to the new SPI is not communicated to the BTS and BTS continues using the old NBR value. RNC ensures that the SPI promotion/demotion for NBR users is performed within the SPI range defined for NBR users

NBRForPri0..12UL UL NBR for Priority value 0..12 (structured parameter) RNHSPA; 0..2000 K; 8 K; 0 K for all priority values

NBRForPri0..12DL DL NBR for Priority value 0..12 RNHSPA; 0..2000 K; 8 K; 0 K for all priority values

NBR: Nominal Bit Rate

Application Aware RAN – Principle Shortcomings of available 3GPP QoS model: • 3GPP bearer-based QoS differentiation model is not widely supported by typical terminals connecting with single PDP context carrying all applications within one bearer. • Subscriber level QoS does not separate different applications within single bearer; although each application has different requirements when utilizing a bearer. Application Aware RAN • Equips operators with QoS tools for typical ter minals carrying multiple applications within one bearer with HSDPA allocated AppAwareRANEnabled • Enables prioritization of the latency sensitive data by increasing WCEL; Disabled (0), Enabled (1); 0 the scheduling priority at the air interface and/or demotion of non-priority P2P traffic (priority drop = P2P traffic share down), and introduces dynamic demotion of UL bulk traffic by the BTS in a single RAB case • Applications requiring the same treatment at RAN are grouped AppGrpId by the operator into Application Groups (up to 6) characterized RNHSPA; 1..6; 1; 255 = not defined with AARConfigTable (consisted of AppGrpId, DSCPCode1..5 DSCPCode1..5 (up to 5 applications per group), Precedence, TargetSPIforSPI0..11) RNHSPA; 0..62; 1; 255 = not defined • Precedence value determines what SPI should be chosen when packets TargetSPIforSPI0..11 belonging to multiple application groups are detected by the RNC RNHSPA; 0..11; 1; 255 = not defined (promote/demote/do nothing) Precedence RNHSPA; 1..6; 1; 255 = not defined

Application Aware RAN – NSN implementation Application Aware RAN solution is implemented in 2 network elements GGSN and RNC 1. In GGSN: Core network based DPI (Deep Packet Inspection) provides application detection and inner (user) IP packet marking with DSCP (Differential Service Code Point - a field in the IPv4 and IPv6 header) DSCP of user packet is marked based on PCC rule action

2. In RNC: Initial Scheduling Priority Indicator of the radio bearer is demoted or promoted in the RNC PDCP layer according to Deep Packet Inspection marking (DSCP marking).

QoS Aware Scheduling and Application Aware RAN (1/2) • Scheduling weights – For each combination of RAB QoS parameters operator can define s e r v i c e p r i o r i t y  Traffic class  Traffic handling priority THP  Allocation & retention priority ARP

– Service priorities & Scheduling Priority Indicators SPI  Defined by multiple parameter QoSPriorityMapping  For services on DCH service priorities just def ine values entering queuing and priority based scheduling (see R99

PS)  For services on HS-DSCH/E-DCH or HS-DSCH/DCH services priorities define directly SPI • It is initial SPI value if AppAwareRANEnabled = 1 (dynamic SPI based on the application type and initial SPI value is set and communicated to BTS using CmCH-PI field in Frame Protocol)  If HSPAQoSEnabled is disabled but AppAwareRANEnabled = 1 then initial SPI f or services with HSDPA can be configured by the operator with is defined by InitialSPINRT; RNHSPA; SPI 5 (5), SPI 6 (6); SPI 5 (5)

– SPI mapped onto scheduling weights:  define how often service of certain QoS parameter set scheduled in comparison to another one with another

QoS parameter set  PF scheduling extended by required activity detection RAD with delay sensitivity DS

Priority for Streaming traffic class with ARP1/2/3: PriForStreamARP1/2/3 (RNPS) (0..15) ( = 1) (13/13/13) Priority for Interactive TC with THP 1 & ARP 1/2/3: PriForIntTHP1ARP1/2/3 (RNPS) (0..11) ( = 1) (11/11/11) Priority for Interactive TC with THP 2 & ARP 1/2/3:

ARP: Allocation & retention priority SPI: Scheduling priority indicators THP: Traffic handling priority

PriForIntTHP2ARP1/2/3 (RNPS) (0..11) ( = 1) (8/8/8) Priority for Interactive TC with THP 3 & ARP 1/2/3: PriForIntTHP3ARP1/2/3 (RNPS) (0..11) ( = 1) (5/5/5)

QoS Aware Scheduling and Application Aware RAN (2/2) • Mapping QoS parameter for DCH QoS parameter

Service

RAB profile

priority

RNC PS Queuing Priority Based Scheduling

Mapping defined by Q o S P r i o r i t y M a p p i n g

• Mapping QoS parameter to scheduling weights for HS-DSCH/E-DCH or HS-DSCH/DCH If A p p A w a r e R A N E n a b l e d = 1 then dynamic SPI setting based on the application type and initial SPI

QoS parameter RAB profile

Node B PS: Service priority

Scheduling weight modifying PF

Mapping defined

Mapping defined by

by Q o S P r i o r i t y M a p p i n g

SchedulingWeightList

SchedulingWeightList • is BTS commissioning parameter • defining Mapping QoSPriorityMapping to SchedulingWeight

SPI: Scheduling Priority Indicators

In-bearer Application Optimization In-bearer Application Optimization introduces service prioritization within one downlink radio access bearer. User traffic marked as latency sensitive is scheduled differently and is prioritized ahead of non-latency sensitive traffic inside RNC. HTTP Mobile Network

The Internet

Priority packets (e.g. HTTP) get more bandwidth within RAB P2P

Promoted traffic Demoted traffic

In Core network: IP Header

H e U a D d P e r

H G e T a P d e U r

GTP-U Payload DSCP

User IP header

User IP payload

DSCP of user packet is marked

In-bearer Application Optimization Without RAN2510

RAB Bandwidth

Web page Background download

Time

Without RAN2510 only generic 3GPP QoS differentiation is possible on RAB level. No content detection with prioritization is possible.

T1 RAB Bandwidth

T2

Web page

With RAN2510

Background download Time

( T2 < T1)

Improved download time

More instantaneous bandwidth granted for prioritized applications’ packets leads to minimized download time in comparison to download time without RAN2510

BETTER QUALITY OF EXPERIENCE

In-bearer Application Optimization

InBearerAppPrioEnabled . WCEL; Disabled (0), Enabled (1);

In-bearer Application Optimization can be enabled with InBearerAppPrioEnabled Parameter. The parameter IBAOHighQueueWeight defines the weight for the high PDCP Priority Queue. IBAOHighQueueWeight

IBAODSCPHighPrioQPart1 RNHSPA; Bit 0: DSCP0, Bit 1: DSCP1, Bit 2: DSCP2, ...... Bit 31: DSCP31

RNHSPA; 50...100 %, step 10 %, 50%

The DSCP code values for PDCP priority high queue are defined by the IBAODSCPHighPrioQPart1 (DSCPs 0 to 31) and IBAODSCPHighPrioQPart2 (DSCPs 32 to 63)

IBAODSCPHighPrioQPart2 RNHSPA; Bit 0: DSCP32, Bit 1: DSCP33, Bit 2: DSCP34, ...... Bit 31: DSCP63

In Radio network: Weighted Fair Queueing

RNC

High priority queue (marked blue), served with bigger weight, resulting in lower delay time and higher bandwith for the higher priority packets than for lower priority packets

HQ LQ

TPU

Core Network

DPI

S

PDCP

RNC

PDCP

RLC MAC

Internet Application marking

Phy New entities in PDCP: HQ, LQ, S.

NodeB

HSDPA RRM • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation – HS-SCCH & HS-DPCCH Power Control – Static & Dynamic HS-PDSCH Power Allocation

• • • • • • •

Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection and Switching Associated UL DCH HSDPA Improvements Other Features Appendix

Overview

HS-PDSCH High-Speed Physical DL Shared Channel Static power allocation Tx power „fixed“ Slowly adjusted in dependence on HS-SCCH Tx power Dynamic power allocation All power not needed for R99 services available for HSDPA Slowly adjusted in dependence on R99 & HSDPA traffic

HS-SCCH Shared Control Channel for HS-DSCH

Fast power control in dependence on: - CQI - Feedback of UE HS-DPCCH Dedicated Physical Control Channel (UL) for HS-DSCH

Fast power control parallel to DPCCH with offset for CQI ACK/NACK

WBTS

UE

associated DCH* Dedicated Channel

Inner loop PC basing DL TPC and CQI F-DPCH* Fractional Dedicated Physical Channel

* F-DPCH can be allocated in DL only if SRB can be mapped to HSPA channels

HS-SCCH Power Control (1/3) HS-SCCH inner loop power control algorithm

• Node B estimates HS-SCCH Tx power according to:

PHS-SCCH = PCPICH + Γ +

– PCPICH: CPICH power – Γ measurement power offset (see section link adaptation)

–

CQI:

power offset taken from CQICOMPENSATED by look up table (next slide)

– P0: correction estimated by HS-SCCH outer loop power control algorithm

• HS-SCCH Tx power – Estimated for each HSDPA connection individually – Updated with each CQI report Example: PCPICH + Γ = 6 W (37.8 dBm) P0 = 0 CQI

TBS

Throughput

4

317

159 K

-7.7 dB

(37.8 - 7.7) dBm = 30.1 dBm (1.0 W)

13

2279

1140 K

-16.6 dB

(37.8 - 16.6) dBm = 21.2 dBm (0.13 W)

CQI

PHS-SCCH

CQI +

P0

HS-SCCH Power Control (2/3) HS-SCCH outer loop power control algorithm • With each feedback ( feedback (ACK ACK or NACK) NACK) from UE – Correction P0 decreased by 0.005 dB – But not below -2 dB (maximum power decrease by factor 1.6)

• If there is no feedback from feedback from UE – Correction P0 increased by 0.5 dB – But not above 4 dB (maximum power increase by factor 2.5)

No feedback P0 0.005 dB 0.5 dB

ACK or NACK time

HS-SCCH Power Control (3/3) • HS-SCCH Tx power increases – in poor channel conditions – with higher HS-PDSCH Tx power

• Link budgets typically assume 0.5 assume 0.5 W HS-SCCH Tx power at cell edge Variance of HS-SCCH Tx power in relatively good channel conditions

Variance of HS-SCCH Tx power in relatively poor channel conditions 18000

30000

16000



25000

14000

PtxMaxHSDPA = 35 dBm 12000

PtxMaxHSDPA = 35 dBm 20000


s e c n 10000 a r u c 8000 c O


s e c n a r 15000 u c c O

6000

10000

4000 5000 2000 0

0 0

100

200

300

400

500

600

700

800

900

1000

0

100

HS-SCCH Tx Power (mW)

200

300

400

500

600

HS-SCCH Tx Power (mW)

Static Power Allocation

700

800

900

1000

HS-DPCCH Power Control • Power offsets – HS-DPCCH Tx power goes parallel to that of DPCCH – for ACK / NACK & CQI fields hardcoded fields hardcoded power offsets in dependence on DPDCH data rate (16 / 64 / 128 / 384 K)

– for UL link budgets ACK ACK / NACK offset more important than CQI one DPDCH

HS-DPCCH

Factor 2.7 dB for 16 K DPDCH 9.5 dB for 384 K DPDCH

CQI

ACK/NACK

DPCCH

HS-PDSCH Power Allocation Static Power Allocation

Dynamic Power Allocation

•

PHSDPA ≤ PtxMaxHSDPA allocated

PHSDPA ≤ min(PtxMaxHSDPA, PtxCellMax)- power to R99 DCH & DL control channels

• •

Fixed load target PtxTargetHSDPA

Dynamically adjusted load target PtxTargetPS

Fixed overload threshold for R99

Overload threshold for R99 goes parallel to load target:

•

PtxTargetHSDPA + PtxOffsetHSDPA

PtxTargetPS + PtxOffset

In case of overload HSDPA might be

In case of overload HSDPA power might be reduced,

released immediately

• •

but usually service not released immediately

Priorities distinguish between R99 & HSDPA users only

PtxMaxHSDPA Maximum allowed HSDPA power WCEL; 0..50 dBm; 0.1 dB; 43 dBm

PtxTargetHSDPA Target for transmitted non-HSDPA power WCEL; -10..50 dBm; 0.1 dB; 38.5 dBm

PtxOffsetHSDPA Offset for transmitted non-HSDPA power WCEL; 0..6 dB;  = 0.1 dB; 0.8 dB

Priorities distinguish between interactive & background users as well

HSDPADynamicResourceAllocation HSDPA Dynamic Resource Allocation RNFC; 0 = disabled; 1 = enabled

PtxCellMax Cell maximum transmission power WCEL; 0 .. 50 dBm; 0.1 dB; 43 dBm

PtxOffset Offset for transmitted power WCEL; 0 .. 6 dB; 0.1 dB; 1 dB

Dynamic HS-PDSCH Power Allocation • BTS may allocate all unused DL power up to maximum cell power • all power available after DCH traffic, HSUPA control & common channels can be used for HSDPA PtxCellMax

PtxMax = min (PtxCellMax, MaxDLPowerCapability)

Cell maximum transmission power 0..50 dBm; 0.1 dB; 43 dBm

PtxHSDPA

PtxNRT

PtxNonHSDPA PtxNC

MaxDLPowerCapability: 0..50 dBm; 0.1 dB; -

Dynamic HS-PDSCH Power Allocation No active HSDPA users • NRT DCH scheduling to – –

PtxTarget + PtxOffset if HS-RACH isn’t set up in the cell PtxTargetPS if HS-RACH is set up in the cell

• RT DCH admission to PtxTarget HSDPA active

No HSDPA users

Active HSDPA users • NRT DCH scheduling to PtxTargetPS • RT DCH admission to – –

PtxTarget no RT HS-SDCH PtxTargetTot at least 1 RT HS-DSCH

No HSDPA users PtxMax

PtxTotal

PtxHSDPA 3 PtxTarget + PtxOffset

1

PtxTargetPS

2

PtxNRT PtxNonHSDPA PtxNC

Dynamic HS-PDSCH Power Allocation • Adjustable load target PtxTargetPS – PtxTargetPSMin (minimum value) – PtxTargetPSMax (maximum value, also initial value, HS-RACH is set up in the cell) – PtxTargetPSMaxtHSRACH (maximum value used if HS-RACH is set up in the cell) PtxMax PtxTargetPSMax -10..50 dBm; 0.1 dB; 40 dBm

PtxHSDPA

PtxTargetPSMax (40 dBm) PtxNRT

PtxTargetPS PtxTargetPSMin (36 dBm) PtxNonHSDPA

PtxNC

PtxTargetPSMin -10..50 dBm; 0.1 dB; 36 dBm

PtxTargetPSMin Min DCH PS target for dynamic HSDPA pwr allocation WCEL; -10..50 dBm; 0.1 dB; 36 dBm

PtxTargetPSMax Max DCH PS target for dynamic HSDPA pwr allocation WCEL; -10..50 dBm; 0.1 dB; 40 dBm

PtxTargetPSMaxtHSRACH Max DCH target power level with HS-RACH for dynamic HSDPA pwr allocation WCEL; 0..40 dBm; 0.1 dB; 32767 dBm (Value set by the PtxTargetPSMax parameter used when the HS-RACH has been setup in the cell)

Dynamic Load Target Ideal load target: Ideal_PtxTargetPS • Dynamic load target adjusted if – High DCH load or total load AND – Current load target deviates from ideal load target

• Ideal load target estimated by RNC in dependence on – Non controllable traffic PtxNC = total non-controllable transmitted DCH power - power used by all HSDPA streaming users of the cell - non-controllable HSDPA power

– NRT DCH traffic (sum over all weights of R99 services Weight DL_DCH) – NRT HS-DSCH traffic (sum over all weights of HSDPA services Weight HS-DSCH)

 Weight DL_DCH   PtxNC   M in Weight HS-DSCH  Weight Ideal _ Ptx Target PS  Max   PtxTarget PSMax     PtxTarget PSMin

  PtxMax  PtxNC  DL_DCH

PtxTargetPSMaxtHSRACH if HS-RACH is set up in the cell

Dynamic Load Target

Structured parameter WeightDCH Weight of NRT DCH UE BG RAB WeightDCHBG (RNHSPA) (0..100) ( = 1) (15)

Weights of individual services Weight of NRT DCH UE THP1/2/3 RAB • Can be set individually for each release WeightDCHTHP1/2/3 (RNHSPA) (0..100) ( = 1) (90/65/40) Structured parameter WeightHSPA – R99 (structured parameter WeightDCH) Weight of HSPA UE BG RAB – HSPA (structured parameter WeightHSPA) WeightHSPABG (RNHSPA) (1..100) ( = 1) (25) • Can be set individually for each traffic class Weight of HSPA UE THP1/2/3 RAB WeightHSPATHP1/2/3 (RNHSPA) (0.100) ( = 1) (100/75/50) – Interactive THP1, THP2, THP3 – Background • In case of multi-RAB the average weight of the individual RABs is taken for that user Ideal Load Target - Example

• 2 HS-DSCH users interactive THP1 + background WeightHS-DSCH = 100 + 25 = 125

• 3 DCH users background WeightDL_DCH = 3 * 15 = 45

• PtxMax = 43 dBm • PrxNC = 37 dBm 

Ideal_PrxTargetPS = 37 dBm + (45 / (125 + 45)) * (43 dBm - 37 dBm) = 38.6 dBm

Traffic Class

HSDPA weight value 0…100

DCH weight value 0…100

Interactive THP1

100

90

Interactive THP2

75

65

Interactive THP3

50

40

Background

25

15

Load Target Adjustment • Required information – Total power PtxTotal measured by Node B – Non HSDPA power PtxNonHSDPA measured by Node B – Both averaged according PSAveragingWindowSize (same parameter as for R99)

• Need for adjustment checked periodically according PtxTargetPSAdjustPeriod • If adjustment needed – Increase by PtxTargetPSStepUp in case of DCH congestion – Decrease by PtxTargetPSStepDown in case of HSDPA congestion PtxTargetPSStepUp

PSAveragingWindowSize

DCH PS target step up for dynamic HSDPA pwr alloc. WCEL; 0..5; 0.1; 1 dB

Load measurement averaging window size for PS WBTS; 1..20; 1; 4 scheduling periods

PtxTargetPSStepDown

PtxTargetPSAdjustPeriod

DCH PS target step down for dynamic HSDPA pwr alloc. WCEL (0..5 dB) ( = 0.1 dB) (1 dB)

DCH PS target adjust period for dyn HSDPA power alloc; WBTS; 1..255; 1; 5 RRI periods

Actions in Case of Congestion DCH congestion only

• Increase PtxTargetPS by PtxTargetPSStepUp, if currently below ideal load target (but not above PtxTargetPSMax)

HSDPA congestion only

• Decrease PtxTargetPS by PtxTargetPSStepDown, if currently above ideal load target (but not below PtxTargetPSMin)

Both DCH & HSDPA congestion

• Increase PtxTargetPS, if currently below ideal load target • Decrease PtxTargetPS, if currently above ideal load target

Example: HSDPA congestion

Decrease by PtxTargetPSStepDown in case of HSDPA congestion

1) HSDPA power congestion, if PtxTargetPSStepDown 0..5; 0.1; 1 dB

Ptxtotal ≥ PtxHighHSDPAPwr 1

PtxMax 43 dBm

PtxHSDPA

PtxHighHSDPAPwr -10..50; 0.1; 41 dBm

PtxTotal PtxTargetPSMax -10..50; 0.1; 40 dBm

PtxTargetPS_ideal PtxTargetPS

PtxTargetPSMin -10..50; 0.1; 36 dBm

2 PtxNonHSDPA

PtxNRT PtxNC

High threshold of PtxTotal for dynamic HSDPA pwr alloc: PtxHighHSDPAPwr (WCEL) (-10..50 dBm) ( = 0.1 dB) (41 dBm)

Example: DCH Congestion Increase by PtxTargetPSStepUp in case of DCH congestion

PtxTargetPSStepUp 0..5; 0.1; 1 dB

2) NRT DCH power congestion, if PtxNonHSDPA ≥ PtxTargetPS - 1dB (hardcoded margin) 1 PtxMax 43 dBm

PtxHSDPA

PtxTotal

PtxHighHSDPAPwr -10..50; 0.1; 41 dBm PtxTargetPSMax -10..50; 0.1; 40 dBm

PtxTargetPS_ideal

PtxTargetPS

PtxTargetPSMin -10..50; 0.1; 36 dBm

2 PtxNonHSDPA

PtxNRT PtxNC

HSDPA RRM • • • • • • • • • • • • • •


Static & Dynamic Allocation (1/3) Number of HSPDSCH codes (full set)

HSDPA

HSDPA

15

10

Codes

Codes

Static code allocation

5

X

X

X

6

-

-

-

7

-

-

-

8

X

X

-

9

-

-

-

10

X

X

-

11

-

-

-

12

X

-

-

13

-

-

-

14

X

-

-

15

X

-

-

HSPDSCHCodeSet 11010 10100 100000


Additionally required HSDPADynamicResourceAllocation = enabled


Static & Dynamic Allocation (2/3) SF=1

Dynamic code allocation applied if: • HSDPA dynamic resource allocation enabled

SF=2

(HSDPADynamicResourceAllocation)

• Maximum number of codes > minimum

SF=4

number (HSPDSCHCodeSet)

• BTS capable of 10/15 codes • HSDPA service starts with minimum number

SF=8

of codes defined by HSPDSCHCodeSet

• Cell-specific scheduler reserves HS-SCCH codes from the spreading code tree according to MaxNbrOfHSSCCHCodes If HSDPA dynamic resource allocation disabled, 5 codes are available only

SF=16 0

1

2

3

4

5 6 7 ……….

8

Rel-99 channels (& HS-SCCH) Rel-99 code area (& HS -SCCH) Shared code area Dedicated HS-PDSCH code area

9

10 11 ……….

12

13

14

HS - PDSCH

15

Static & Dynamic Allocation (3/3)

SF16

+14 x SF16

SF16

HS-PDSCH SF32

32

SF64

64

32

64

64

64

64

S-CCPCH1

SF128

128

128

128

128

128

S-CCPCH2 HS-SCCH SF256 256 256 256 CPICH

256 256 256 256

AICH

128

128

128

HS-SCCH HS-SCCH 256 256 256 256

256 256 256 256

128

128

E-RGCH

E-HICH

256 256 256 256 E-AGCH

P-CCPCH PICH 128

128

128

Allocated CC

Blocked CC

Available CC

Maximum of 14 HS-PDSCH codes possible with 3 HS-SCCH & HSUPA

128

256 256 256 256

Dynamic Allocation Procedure (1/2) HSPDSCHAdjustPeriod

Periodic upgrade • HSDPA service starts with minimum number of codes • RNC attempts periodic upgrade according the timer HSPDSCHAdjustPeriod if •

RNHSPA; 1..60; 1; 10s

Number of currently allocated HS-PDSCH codes < maximum allowed number supported by BTS capability Free SF 16 codes adjacent to currently allocated ones available After upgrade enough SF 128 codes available according HSPDSCHMarginSF128

• • • If all conditions are fulfilled, the next greater value from HS-PDSCH code set is taken

HSPDSCHMargin SF128 WCEL; 0..128; 1; 8 # SF128 codes to be available after Code upgrade

Periodic downgrade • RNC attempts periodic downgrade according the timer HSPDSCHAdjustPeriod if • Number of currently allocated HS-PDSCH codes > minimum allowed number • Not enough SF 128 codes available according HSPDSCHMarginSF128 • If all condition fulfilled, the next lower value from HS-PDSCH code set is taken

Dynamic Allocation Procedure (2/2) • Code congestion events – RT request congested due to lack of code  HS-PDSCH downgrade in any case – NRT request congested due to lack of code  HS-PDSCH downgrade only, if actually for HSDPA too much SF 16 codes in use according DPCHOverHSPDSCHThreshold

• Limitations of congestion triggered downgrade – Not below minimum allowed number of HS-PDSCH codes – Highest still possible number of codes according HSPDSCHCodeSet is taken

HSPDSCH CodeSet WCEL; 5..15; 1; 5

s e d o c 6 1 F S d e t a c o l l a f o r e b m u N

15 14 13 12 11 10 9 8 7 6 5

Maximum code set DPCHOverHSPDSCHThreshold

set relative to max. number of codes

WCEL; 0..10; 1; 5

Minimum code set

Code tree optimization • Code tree optimization procedure tries to re-arrange DPCH codes to make room for HS-PDSCH code upgrade •

DPCH O ver HSPDSCH Threshold

DPCHs having SRB DCH only are not allowed to be re-arranged

CodeTreeOptimisation WCEL; 0 = disabled; 1 = enabled

HSDPA RRM • • • • • • • • •

• • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling HSDPA Power Allocation HSDPA Code Allocation (Basics) HSDPA Mobility – Serving Cell Change – HSPA+ over Iur – Inter-RNC Mobility – Inter-frequency Mobility – Directed RRC Connection Setup HSDPA Channel Type Selection & Switching Associated UL DCH HSDPA Improvements Other Features Appendix

Parameter Templates RNFC 32

RNC

G:InterSystem I:InterFrequency S:IntraFrequency

RNMOBI

100

ADJG / L

FMCG 48

WBTS 100

HOPG

100

ADJI FMCI

WCELL

100

32

HOPI

ADJS

100

FMCS 32

FMCS/I/GId

ADJD

HOPS

HOPS

100

100

identifies parameter set for intra-, inter-frequency & inter-system measurements FMCS/G/I; 1..100; 1; no default

HOPSId

HSDPAFMCS/I/GIdentifier

HOPS identifier: identifies parameter set f or intra-frequency mobility HOPS; 1..100; 1; no default

Identifies FMCS/I/G parameter set to be applied for a HSDPA service within a certain serving cell WCEL; 1..100; 1; no default

HSDPAHOPSIdentifier

RTwithHSDPAFMCS/I/GIdentifier HSDPA FMCS/I/G identifier for AMR multi-service WCEL; 1..100; 1; no default Identifies FMCS/I/G parameter set to be applied for a HSDPA + AMR multi-RAB service within a certain serving cell

ADJS; 1..100; 1; no default identifies parameter set to be applied for a HSDPA service to move to a certain adjacent cell

RTwithHSDPAHOPSIdentifier HSDPA HOPS identifier for AMR multi-service ADJS; 1..100; 1; no default

HSDPA Mobility Methods to handle HSDPA mobility • Serving HS-DSCH cell change • Cell reselection

HSDPAMobility Serving HS-DSCH cell change & SHO on/off switch RNFC; 0 = HSDPA cell reselection 1 = Serving HS-DSCH cell change

with HS-DSCH - FACH channel type switching (  Appendix)

Serving Cell Change SCC (1/5): Candidate Initial cell selection • 1 cell active only: just attempt to establish service • More than 1 cell active

HSDPAServCellWindow CPICH Ec/Io window for serving HS-DSCH cell selection

RNMOBI; 0..6; 0.5; 2 dB

– Initial selection of Serving Cell based on latest reported Ec /I0 – To be candidate, HSDPA capable cell must fulfil following condition:

EC /I0 (active cell*) ≥ EC/I0 (best cell) – HSDPAServCellWindow –

Serving cell is chosen in order of EC /I0

– If allocation of HS-DSCH fails due to any reason, next best candidate cell is attempted Max. allowed difference between the best cell in the Active Set & the Serving HSDSCH cell. If Serving HS-DSCH cell out of this window  Serving HS-DSCH cell change procedure initiated.

* Serving Cell

Serving Cell Change : Ec/Io based /I m e a s u r e m e n t s • Periodic al Intra-frequenc y E started when: C 0

– HS-DSCH MAC-d flow active AND Active set size > 1 ( event 1a) – Measurements stopped if either of the above criteria not true – Higher layer filtering for measurement results before reporting by

–

FMCS; k = 0..6; 1; k = 3

HSDPACPICHReportPeriod

• CPICH EC/I0 measurement reporting by UE: –

EcNoFilterCoefficient

EcNoFilterCoefficient Periodical reporting with reporting interval defined by HSDPACPICHReportPeriod RNC averages reports over HSDPACPICHAveWindow

RNMOBI; 0.25, 0.5, 1, 2, 3, 4, 6, 8, 12; 0.5 s

HSDPACPICHAveWindow RNMOBI; 1..10; 1; 3

• EC/I0 based Serving Cell change triggered if: – Ec /I0 (server) < EC /I0 (best cell) – HSDPAServCellWindow AND – EC /I0 (server) < HSDPACPICHEcNoThreshold Window HSDPAServCell

EC /I0

Serving Cell change window RNMOBI; 0..6; 0.5; 2 dB

CPICH 1 CPICH 2

Addition window

Serving cell change EC/I0 threshold

periodic reports as long process is running New cell detected

Periodic reports

Serving cell change triggered

HSDPACPICH EcNoT h r e s h o l d RNHSPA; -20..0; 0.5;-5 dB

time

Serving Cell Change: SIR error based • for Inter Node B & intra Node B inter-LCG cell change only (not applicable for intra Node B intra-LCG ) • Periodical SIR error measurements started when – HS-DSCH MAC-d flow active – difference between actual SIR & SIRtarget: SIRerror = SIR – SIRtarget • Measurement reporting by Node B HSDPASIRErrorFilterCoefficient – Higher layer filtering for measurement results before reporting by RNMOBI; k = 0..10; 1; 5

SIRerror

HSDPASIRErrorFilterCoefficient – Periodical reporting with reporting interval defined by HSDPASIRErrorReportPeriod (if set to 0 SIR measurement is not used as criteria for SCC) – RNC averages reports over HSDPASIRErrorAveWindow

HSDPASIRErrorReportPeriod RNMOBI; 0..10; 0.5; 0.5 s

HSDPASIRErrorAveWindow RNMOBI; 1..10; 1; 3

• SIR error based Serving Cell change triggered if: SIRerror (Server) < HSDPASIRErrorServCell

HSDPASIRErrorServCell RNMOBI; -10..0; 0.5; -3 dB

Serving cell change SIRerror threshold

Periodic reports as long HSDPA service running HSDPA service established

Serving cell change triggered

time

Serving Cell Change: other trigger

Method

Trigger Event 1B

AS update

Event 1C

on Serving Cell

Event 6F/6G HO to D-RNC

AS update for Serving Cell to D-RNC

Serving Cell Change Target cell selection criteria

• Dynamic Resource Allocation disabled – Cell having HSDPA power allocated already chosen as serving cell – Otherwise serving cell chosen in order of E C/I0

• Dynamic Resource Allocation enabled – Serving Cell is chosen in order of EC /I0

• If serving cell change triggered by Ec /I0 or SIRerror – need SIRerror (target) ≥ HSDPASIRErrorTargetCell

HSDPASI R Error T arget Cell RNMOBI; -10..0; 0.5; -2 dB HSDPA SIRError S ervCell RNMOBI; -10..0; 0.5; -3 dB

• If triggered by other event: – need SIRerror (target) ≥ HSDPASIRErrorServCell

Timing Constraints • min. time interval between consecutive Serving HS-DSCH Cell changes based on Ec/I0: HSDPACellChangeMinInterval

• max. number of repetitive Serving HS-DSCH Cell changes HSDPAMaxCellChangeRepetition during predefined time period HSDPACellChangeRepetitionTime

• if exceeded, HS-DSCH released & switched to DCH0/0 or DCH with initial bit rate

HSDPACellChange MinIn terval RNMOBI; k = 0..10; 1; 3 s

HSDPACellChange RepetitionTime RNHSPA; 0..60; 1; 10 s

HSDPAMaxCell C h a n g eR epetition RNHSPA; 1..16; 1; 4

HSDPA RRM • • • • • • • • •

• • • • •


HSPA+ over Iur Introduction • HSPA over Iur feature improves the end-user performance by maintaining the continuous high data rate HSPA service during the inter-RNC mobili ty.

• The possibility of setting up HSDPA/HSUPA MAC-d flows over Iur interface is introduced by the feature RAN1231 in RU20.

• Prior, HSDPA channel type switch to DCH is performed. (only DCH services are allowed over Iur). one PS NRT RAB

• After the serving-cell change, HSDPA and HSUPA data is transmitted over the Iur.

• HSPA throughput over Iur restricted to 10Mbps in DL

SRNC

and 2Mbps in UL

• HSPAOverIur enables HSPA over Iur. • The DRNC does not read the parameter HSPAOverIur but the license only. HSPAOverIur IUR; 0 (HSPA over Iur disabled), 1 (HSPA over Iur enabled)

UE DRNC

HSPA+ over Iur Extension • Extension of HSPA over Iur feature introduces additionally: - the CS AMR on DCH + 1 PS NRT on HS(D)PA multi-RAB combination over Iur,

• and it can be enabled with HSPAOverIurExt (in RU40).

SGSN HSDPA+ allowed over Iur

HSPAOverIurExt IUR; 0 (Disabled), 1 (Enabled)

• It allows: - HS-DSCH and E-DCH Mac-d flow setup and release over Iur

SRNC DRNC

for single PS NRT RAB.

- The SRNC to set up the HS-DSCH and/or E-DCH RL over Iur during anchoring.

- The SRNC to perform SCC from DRNC cell to DRNC cell during anchoring.

- The SRNC can set up a single HS-DSCH and /or E-DCH Mac-d flow with CS AMR on DCH over Iur.

- The SRNC allows for the PS NRT RAB reconfiguration for HS-DSCH and E-DCH Mac-d flow over Iur.

SCC during anchoring (DRNC cell to DRNC cell) allowed due to RAN2270

HSPA+ over Iur RU50 New improvements (RAN2221) • Introduces the following functionalities to the Iur interface: -

Flexible RLC in DL (FRLCOverIurEnabled )

-

Dual-Cell HSDPA (DCHSDPAOverIurEnabled )

-

HSDPA 64QAM (HSDPA64QAMOverIurEnabled )

FRLCOverIurEnabled IUR; 0 (Disabled), 1 (Enabled)

SGSN

• New HSDPA configurations supported over Iur: - Single cell HSDPA with Flexible RLC DL (14Mbps) - Single cell HSDPA (64QAM) with Flexible RLC DL (21Mbps) - Dual cell HSDPA with Flexible RLC DL (28Mbps) - Dual cell HSDPA (64QAM) with Flexible RLC DL (42Mbps) - For RAN1231 HSPA over Iur throughput in DL was limited to 10Mbps.

HSDPA64QAMOverIurEnabled IUR; 0 (Disabled), 1 (Enabled)

DCHSDPAOverIurEnabled IUR; 0 (Disabled), 1 (Enabled)

SRNC

Up to 42 Mbps DL 2 Mbps UL

DRNC

HSPA+ over Iur RU50 New improvements (RAN2221)

HSUPACCIurEnabled IUR; 0 (Disabled), 1 (Enabled)

• Configurations are supported only with SRB on DCH.

SGSN

• In case of SRB on HSPA reconfiguration to SRB on DCH is done before SCC.

SRB on DCH

• Configurations are supported only with one NRT PS RAB.

DRNC

• HSUPACCIurEnabled enables the HSUPA Congestion Control for Iur E-DCH MAC-d flows in the SRNC, covering also DRNC's Iub part.

• Maximum Bit rate limitations are configured with: - MaxIurNRTHSDSCHBitRate (DL), - MaxTotalUplinkSymbolRate (UL) MaxIurNRTHSDSCHBitRate IUR; 128...41984 kbps, step 128 kbps; 75 kbps

SRNC

MaxTotalUplinkSymbolRate WCEL; 960 kbps, SF4 (0), 1920 kbps, 2*SF4 (1), 3840 kbps, 2*SF2 (2), 5760 kbps, 2*SF2 + 2*SF4 (3)

HSPA+ over Iur RU50 New improvements (RAN2221) • Anchoring: • When HSPAOverIurExt is enabled the SRNC is allowed to: SGSN

- Setup HS-DSCH with Flexible RLC in DL, DC-HSDPA, and/or HSDPA64QAM over Iur

- perform Serving Cell Change from DRNC cell to DRNC cell with: • Flexible RLC in DL, DC-HSDPA, and/or HSDPA64QAM without radio links in serving RNC.

- If there is an attempt to establish AMR call with the existing HSPA+ over Iur RAB:

• DC-HSDPA is reconfigured to SC-HSDPA for enabling AMR+HSPA over Iur .

- When there is an attempt to establish another PS RAB with the existing HSPA+ over Iur RAB,

• DRNC rejects the request by the failure code Requested Configuration not Supported .

SRB on DCH SRNC DRNC

HSPA+ over Iur Nokia and non-Nokia DRNC operation • Neighboring RNC settings (Nokia to non-Nokia RNC) are configured with InterfaceMode and the Neighboring RNC settings for Nokia need to be NRncVersion = Rel 9 or higher. NRncVersion InterfaceMode

IUR; 3GPP, Nokia (0), Mode 1 (1), Mode 2 (2), Mode 3 (3), Mode 4 (4), Mode 5 (5), Mode 6 (6), Mode 7 (7)

DC-HSDPA SCC from SRNC to Nokia DRNC Cell Capability Containers of the neighboring cells of the target cell, and for the target cell are send by DRNC If Flexible RLC and DC-HSDPA (or HSDPA 64QAM) are supported, those will be used on the DRNC cell as well. UE makes SCC to DRNC cell, with Flexible RLC DL and DC-HSDPA or HSDPA 64QAM (if supported).

IUR; R99 (1), Rel4 (2), Rel5 (3), Rel6 (4), Rel7 (5), Rel8 (6), Rel9 (7), Rel10 (8), Rel11 (9), Rel12 (10)

DC-HSDPA SCC from SRNC to non-Nokia DRNC Cell Capability Containers of the neighboring cells of the target cell are send by DRNC If SRNC does not receive the HS-DSCH Support Indicator assumes that HS-DSCH is supported If SRNC does not receive the E-DCH Support Indicator assumes that E-DCH is supported If Cell Capability Container of the target cell is not received from DRNC, intra-frequency SCC over Iur shall be tried with existing RLC. But if new HSDPA is established, then fixed RLC is used. UE makes SCC to DRNC cell, with SC-HSDPA.

HSDPA RRM • • • • • • • • •

• • • • •


HSPA Inter-RNC Cell Change • The HSPA Inter-RNC cell change is applied to Flexi Direct RNC when: -

1. The Iur interface between the Adapters does not exist (is not c onfigured). 2. S-Flexi Direct RNC has one or more radio links (RL) with the RNC. 3. When SHO over Iur is not enabled, that is the RNP parameter EnableInterRNCsho is disabled. 4. Iur interface is enabled and SHO over Iur fails (PS RABS only).

• improves the end user performance by:

SGSN /GGSN

- maintaining a high data rate HSPA service during intra-frequency inter-RNC mobility.

Iu /Gn RNC

RNC

• Capacity gain is achieved at the cells border area: - HSPA instead of DCH can be used.

Iur

Iub

• uses SRNS relocation with UE involvement Serving HSPA RL UL DCH E-DCH non-serving RL / UL DCH

Situation prior to HSPA inter-RNC cell change

HSPA Inter-RNC Cell Change • HSPA intra-frequency inter-RNC cell can be enabled with HSPAInterRNCMobility parameter: - need to be set to Enabled or Enabled without E-DCH trigger .

• With HSPAInterRNCMobility =“Disabled”: - HSPA Inter-RNC cell change is not supported but SRNC applies a switch from HSPA to DCH at the RNC border.

• HSPA Inter-RNC cell change from source RNC to target RNC is performed by means of the “UE involved” SRNS relocation procedure.

SGSN/GGSN

• HSDPAMobility has to be set to “ Enabled”. RNC

• A new serving cell cannot be selected under the DRNC, - if the feature HSPA over Iur is not in use - or the DRNC does not support CS voice over

SRNC

Iu/Gn

Iur

RNC

DRNC Iub

HSPA (virtual cell parameter HSPAQoSEnabled ). Serving HSPA RL

Situation after successful HSPA inter-RNC cell change

HSDPA RRM • • • • • • • • •

• • • • •


Inter-frequency Mobility (Optional feature) • Trigger for IFHO / ISHO process in case of active HSDPA service – – – – – –

Event 1F (too low Ec /I0 or RSCP for all active cells) Event 6A (too high UE Tx power) Too high DL RL power UL quality deterioration IMSI based HO Capability based HO

• General rule for HHO process – Channel type switch HS-DSCH to DCH for ISHO – No channel type switch for IFHO

• Allowed transitions for IFHO process – DCH/DCH to 

DCH/HSDPA



HSUPA/HSDPA

– DCH/HSDPA to 

DCH/DCH



DCH/HSDPA



HSUPA/HSDPA

HSDPA RRM • • • • • • • • •

• • • • •


Directed RRC Connection Setup Basic feature • Target – R5 or newer UEs directed from non-HSDPA supporting carrier to HSDPA supporting one

DirectedRRC ForHSDPALayer Enabled WCEL; 0 = disabled; 1 = enabled

– R99 or R4 UEs directed from HSDPA supporting carrier to non-HSDPA supporting one

– Feature works within same sector defined by SectorID

RNMOBI; 0 = disabled; 1 = enabled

• Required parameter settings – DirectedRRCForHSDPALayerEnabled = enabled – DirectedRRCForHSDPALayerEnhanc = disabled

Basic functionality

Enhanced functionality

• •

• •

Only for 2 layers Service (cause for RRC connection setup) not considered

• •

Load of target layer not considered Cannot be used simultaneously with Layering in Cell_FACH supported (same rules as for RRC con. setup)

More than 2 layers supported Can be restricted to certain types of services

• •

R99 directed RRC connection setup

•

DirectedRRC ForHSDPALayer Enhanc

Load balancing applied R99 directed RRC connection setup simultaneously supported

•

Layering in Cell_FACH supported (same rules as for RRC con. setup)

SectorID WCEL; 0..12; 1; 0 = cell not belonging to any sector HSDPALayering C ommonChEnabled HSDPA layering for UEs in common channels enabled


Directed RRC Connection Setup Enhanced feature • Non-HSDPA UEs – Directed away from HSDPA capable cell if  Load of the target cell not too big (i.e. R99 load balancing back to source

cell not triggered)

• HSDPA UEs – Directed away from non-HSDPA capable cell if  Establishment cause indicated by UE allowed in HSDPA layer  Not too much HS-DSCH users in target cell

– Directed to other HSDPA capable cell if  Load balancing required  Establishment cause indicated by UE allowed in HSDPA layer

• HSUPA UEs – Same rules as for HSDPA UEs, but additionally  Directed to HSUPA capable cell if possible  Not directed away from HSUPA capable cell

Directed RRC Connection Setup: Example Decision algorithm for UEs camping on non HSDPA layer

UE reporting Rel-6 HSDPA & HSUPA capability

UE reporting Rel5 or Rel-6, HSDPA capability

UE HSPA capability = cell HSPA capability

f3, HSDPA&HSUPA

Any other UE

C

A

Yes

B&C

No

current layer (f1) f2 & f3

Establishment cause allowed in HSDPA target layer

f2, HSDPA

B

A

f1, R´99

B&C

No

B&C

Yes


UE HSPA capability = target cell HSPA capabili ty B

f2 or f3 (where more HSDPA throughput)

C

f3

Directed RRC Connection Setup: Example Decision algorithm for UEs camping on HSDPA layer

UE reporting Rel-6 HSDPA & HSUPA capability

UE HSPA capability = cell HSPA capability

UE reporting Rel5 or Rel-6, HSDPA capability

Any other UE

No

B&C

Yes

f1 f2 & f3

Establishment cause allowed in HSDPA target layer

f3, HSDPA&HSUPA

B&C

No

B&C

Yes


UE HSPA capability = target cell HSPA capabili ty

C B

A

f2, HSDPA

A f1, R´99

B

f2 or f3 (where more HSDPA throughput)

C

f3

Directed RRC Connection Setup: Load Balancing Load Balancing

• Load of serving and target cell (both in same sector) is checked only if DirectedRRCForHSDPALayerEnhanc parameter is ON

• Applied if there are 2 or more layers supporting HSDPA • Target layer selection depends on number of active HSDPA UEs, which is checked against HSDPALayerLoadShareThreshold – if number of UEs > HSDPALayerLoadShareThreshold in one cell of sector HSDPA UEs directed to HSDPA layer offering highest HSDPA power per user

– Otherwise HSDPA UEs directed to HSDPA layer with highest value of CellWeightForHSDPALayering

HSDPALayerLoadShareThreshold

CellWeightForHSDPALayering

HSDPA layers load sharing threshold RNMOBI; 0..48; 1; 3

Cell weight for HSDPA layering WCEL; 0.01..1; 0.01; 1

Directed RRC Connection Setup: Load Balancing Number of HS-DSCH users <

Number of HS-DSCH users > HSDPALayerLoadShareThreshold for one layer

HSDPALayerLoadShareThreshold for all layers Max

Max

HSDPALayer LoadShare Threshold RNMOBI; 0..48; 1; 3 1; 3

0 CellWeightFor HSDPALayering WCEL; 0.01..1; 0.01; 1

0 Cell f1

Cell f2

Select cell which has • Highest cell weight (CellWeightForHSDPALayering CellWeightForHSDPALayering)) • Highest number of HS-DSCH users

Cell f3

Cell f1

Cell f2

Cell f3

Select cell which offers highest HSDPA power per user

Directed RRC Connection Setup: Load Balancing HSDPA power per user • If not disabled with DisablePowerInHSDPALayeringDecision DisablePowerInHSDPALayeringDecision,, select cell with highest HSDPA Power per user : HSDP HSDPA APowe ower Pe PerUser rUser 

 Ptx PtxM Max  Ptx PtxN NonHS onHSP PA * CellWeight ForH ForHSD SDP PALa ye yering ring Num Numbe berOf rOfHS HS DPA DPAUsers Users  1

• Otherwise select cell with highest HSDPA Cell Weight of: HSDP HSDPA ACell CellW W eightPerUs er 

PtxMax HSPA power PtxNonHSPA Non HSPA power 0

CellWeight ForH ForHSD SDP PALa yeri yering ng Numb Numbe erOfHS rOfHS DPA DPAUsers Users  1 Number of HS-DSCH users > HSDPALayerLoadShareThreshold for one layer

DisablePowerInHSDPA LayeringDecision Disable power in decision making for HSDPA layering

RNMOBI; 0..1; 0 = not disabled

Directed RRC Connection Setup Interworking with R99 directed RRC connection setup

•

Both parameters DirectedRRCEnabled and DirectedRRCForHSDPALayerEnabled enabled and DirectedRRCForHSDPALayerEnhanc enabled

•

Decision of directed RRC connection setup for HSDPA layer done first – Decision = change layer directed RRC connection setup for HSDPA layer is done – Decision = do not change layer decision of directed RRC connection setup is done

•

If several target candidates exist for R99 directed RRC connection setup – – – – – –

•

UE kept in most suitable layer from capability point of view, if possible Non HSDPA capable UE HSDPA capable UE

 non-HSDPA

 HSDPA

capable cell

or HSDPA and HSUPA capable cell

HSDPA and HSUPA capable UE

 HSDPA

F-DPCH capable UE

capable cell preferred othe rwise HSDPA&HSUPA capable and HSDPA capable cells

 F-DPCH

DC HSDPA capable UE capable cells

 HSPA/DC

& HSUPA capable cell preferred, then HSDPA capable cell

HSDPA capable cell preferred otherwise HSDPA&HSUPA capable and HSDPA

HSDPA/HSUPA HSDPA/HSUPA capable UE in R99 directed RRC connection setup – not transferred away from HSDPA/HSUPA layer if requesting interactive or background service – can be transferred away from HSDPA/HSUPA layer layer if requesting other kind of service

HSDPA RRM • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching CTS – – – –

• • • •

Channel Type Selection Switching from DCH to HS-DSCH Switching from HS-DSCH to DCH Switching from HS-DSCH to FACH

Associated UL DCH HSDPA Improvements Other Features Appendix

Channel Type Selection CTS HS-DSCH selected in case of Capacity Request if all of the following conditions are met:

1)

Traffic class & THP allowed on HS-DSCH: configurable with HSDSCHQoSClasses

2)

UE supports HS-DSCH

2)

Cell supports HSDPA & HS-DSCH is enabled

3)

Multi-RAB combination of UE supported with HS-DSCH HSDPA + AMR to be enabled with AMRWithHSDSCH HSDPA + R99 NRT + AMR / R99 streaming enabled with HspaMultiNrtRabSupport

HsdschGuardTimerHO HS-DSCH guard time after switching to DCH due to HO

RNHSPA; 0..30 s; 1 s; 5 s

HSDSCHGuardTimerLowThroughput HS-DSCH guard timer due to low throughput

RNHSPA; 0..240 s; 1 s; 30 s

5)

No. of simultaneous HS-DSCH allocations in BTS/cell below max. no. supported by base band configuration

6)

H s d s c h G u a r d T i m e r H O & H s d s c h G u a r d T i m e r L o w T h r o u g h p u t guard timers not running for UE

7)

UE not performing inter-frequency or inter-system measurements

8)

Active set size = 1 if H S D P A M o b i l i t y = disabled

9)

If HSDPA dynamic resource allocation disabled and no existing MAC-d flow in the cell PtxNC ≤ PtxtargetHSDPA for HSDPAPriority = 1

PtxnonHSDPA ≤ PtxtargetHSDPA for HSDPA Priority = 2

10) 11) 12) 13)

UE does not have DCHs scheduled with bit rates higher than zero HS-DSCH physical layer category is supported HS-DSCH can be admitted if PS streaming and CS voice RB resource are utilized HSDPA prevention function of the RAN2879: Mass Event Handler feature does not prevent from HS-DSCH allocation HSDPA prevention is started if RNC starts using the prioritized DL power

HSDSCHQoSClasses HS-DSCH QoS classes

RNHSPA; 11111 = background / interactive with THP 1/2/3 / streaming allowed

AMRWithHSDSCH Usage of AMR service with HS-DSCH

RNFC; 0 = disabled; 1 = enabled

HspaMultiNrtRabSupport HSPA multi RAB NRT support

WCEL; 0 = disabled; 1 = enabled THP: Traffic Handling Priority

CTS: DCH to HS-DSCH Trigger 1)

First HSDPA capable cell added to the Active Set (UE enters HSDPA coverage) Example: SHO of HSDPA capable UE

2)

RAB combination of UE is changed so that it supports HS-DSCH Example: Release of video call (multi RAB NRT support disabled)

3)

Initial HS-DSCH reservation not successful for temporary reason (DCH allocated although HS-DSCH supported) Example: No dynamic power allocation, initially too high non controllable load

4)

HS-DSCH to DCH switch done for IFHO/ISHO measurement, but IFHO/ISHO not performed due to unsatisfied measurement results Example: No suitable adjacent IF/IS cell found

f1

SWITCH

f2 CTS: Channel Type Switching

non-HSDPA

HSDPA

CTS: DCH to HS-DSCH General Conditions 1)

UE has RAB combination supporting HSDPA

• • 2)

Not more than three NRT RABs (if multi RAB NRT support enabled) No R99 streaming or NRT RAB (if multi RAB NRT support disabled)

UE and at least 1 active cell HSDPA capable

•

If HSDPAMobility = disabled, active set size must be 1

3)

No inactivity or low utilization detected on DCH (DL/UL)

4)

No guard timers running to prevent HS-DSCH selection

• • • 5)

HsdschGuardTimerHO HSDSCHGuardTimerLowThroughput HSDSCHCTSwitchGuardTimer

RAB attribute “Maximum bit rate” does not prevent use of HSDPA

HSDSCHCTSwitchGuardTimer HS-DSCH channel type switch guard timer RNHSPA; 0..30 s; 1 s; 5 s

CTS: DCH to HS-DSCH Ec/Io condition for HS-DSCH candidate:

Ec/Io (candidate) > Ec/Io (best cell) – HSDPAChaTypeSwitchWindow

Periodic Ec/Io measurements • • • •

EC /I0

Filtering based on EcNoFilterCoefficient as for any mobility functionality Reporting period defined by specific parameter HSDPACPICHCTSRepPer RNC averaging over HSDPACPICHAveWindow reports* RNC needs as least 1 report to initiated channel type switch

HSDPACPICHAveWindow RNMOBI; 1 .. 10; 1; 3

HSDPAChaTypeSwitchWindow RNHSPA; 0..4; 0.5; 0 dB CPICH 1 R99

Addition window

CPICH 2 HSDPA

Addition Time

HSDPA cell detected

Periodic reports

Channel type switch

HSDPACPICHCTSRepPer RNHSPA; 0.5; 1; 2; 3; 4; 6 s; 2 s

time

* as for any HSDPA mobility functionality

CTS: HS-DSCH to DCH • Trigger – – – – –

Last HSDPA capable cell dropped Event 1F (too low Ec/I0 or RSCP for all active cells) Event 6A (too high UE Tx power) Too high DL RL power UL quality deterioration

• DCH allocation – attempted in next scheduling period with initial bit rates defined by InitialBitRateUL & InitialBitRateDL

– If initial bit rates can not be allocated, DCH 0/0 is offered only

Only if ISHO process triggered In case of IFHO process switch not required

CTS: HS-DSCH to FACH • HS-DSCH released & channel type switching to Cell_FACH in Cell_FACH in following cases: – Low utilization – Low throughput – In case of Multi-RAB with AMR no channel type switching to Cell_FACH, but to Cell_DCH with AMR + NRT DCH 0/0

• Throughput calculated by counting all transmitted bits during configurable sliding measurement window MACdflowthroughputAveWin – Parameter = 0  throughput measurements switched off – Otherwise  throughput measurements averaged over sliding window – Sliding measurement window moved every HS-DSCH MAC-d scheduling interval MACdflowThroughputAveWin WAC; 0..10 s; 0.5 s; 3 s

Switching from HS-DSCH to FACH: Low Utilisation • HS-DSCH released & CTS to Cell_FACH in following cases: – Low utilization – Low throughput

MACdflowThroughputAveWin WAC; 0..10 s; 0.5 s; 3 s

– In case of Multi-RAB with AMR no CTS to Cell_FACH, Cell_FACH, but to Cell_DCH with AMR with AMR + NRT DCH 0/0

• Low Utilisation indicated when

MAC-d PDU in buffer

– MAC-d flow throughput below MACdflowutilRelThr – AND RLC does not have any data to send – AND there are no are no more data in the BTS buffer (normal release) release)

MACdflowutilRe l Thr Thr Low utilisation threshold of the MAC-d flow WAC; 0..64000 bps; 256 bps; 256 bps

Throughput Timer started Timer started

MACdflowutilTimetoTrigger

Timer reset

Low utilization time to trigger of the MAC-d flow WAC; 0..300 s; 0.2 s; 0 s

MACdflowutilRelThr

Time

Switching from HS-DSCH to FACH: Low Throughput • Low Throughput indicated when – MAC-d flow throughput below throughput below MACdflowthroughputRelThr – AND there is still data in the BTS buffer (abnormal (abnormal release) release) – After MAC-d flow release release HS-DSCH not allowed until guard timer HsdschGuardTimerLowThroughput expires MACdflowthroughputRe l Thr Thr Low throughput threshold of the MAC-d flow bps; 0 bps WAC ; 0..64000 bps; 256 bps; 0

MAC-d PDU in buffer

MACdflowthroughputTimetoTrigger Low throughput time to trigger of the MAC-d flow WAC ; 0..300 s; 0.2 s; 5 s; 5 s

HsdschGuardTimer LowThroughput LowThroughput RNHSPA; 0..240 s; 1 s; 30 s; 30 s

Throughput

Timer started

Timer reset

Timer started

Timer started

Timer started

HsdschGuardTimerLowThroughput

MACdflowutilRelThr MACdflowthroughputRelThr Time

HSDPA RRM • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH – Bit Rates – Packet Scheduling

• HSDPA Improvements • Other Features • Appendix

UL Return channel - Bit Rates RB mapped onto HS-DSCH in DL

DCH (or E-DCH) allocated as UL return channel

• data rates for UL DCH return channel: – – – –

HSDPA16KBPS ReturnChannel RNFC; 0 = disabled; 1 = enabled

16, 64, 128 &384 kbit/s independent on R99 settings 16, 64, 128 kbit/s if PS streaming is mapped on HS-DSCH 16 kbps UL DCH return channel*: HSDPA16KBPSReturnChannel

HSDPAminAllowedBitrateUL: min. allowed bit rate -> this parameter is also used to limit UL DCH date rate if RAN2879 Mass Event Handler is used

PS: HS-DSCH ((DL) DL)

PS: DCH ( UL)

HSDPAm i n A l l o w e d B i t r a t e U L Min. bit rate for HSDPA a-DCH

WAC; 16 K, 64 K, 128 K, 384 K

Packet Scheduling: HSDPA with UL associated DCH • HS-DSCH allocation triggered by UL: – high traffic volume indicated  RNC tries to allocate return channel with highest possible bit rate – low traffic volume indicated  RNC tries to allocate return channel with initial bit rate itrateUL HSDPAi n i t i a l B Initial bit rate for HSDPA a-DCH WAC; 16 K, 64 K, 128 K, 384 K

• HS-DSCH allocation DL triggered:  RNC

tries to allocate HSDPAinitialBitrateUL

• Direct DCH to HS-DSCH switch  UL

a-DCH bit rate can be same as existing DCH UL bit rate

• initial bit rate cannot allocated  UL/DL DCH • UL a-DCH functionalities: – PBS & overload control – Decrease of retried NRT DCH bit rate – RT over NRT – Throughput based optimisation – Upgrade of NRT DCH data rate HS-DSCH not possible

kbps

Example Initial bit rate = 64 K Minimum bit rate = 16 K

Decrease of the retried NRT DCH bitrate

384



(normal or flexible) enabled by DynUsageHSDPAReturnChannel Dynamic usage of UL NRT a-DCH HSDPA return channel RNFC; 0 or 1; 0 = disabled

PBS RT-over-NRT 128 Initial bitrate 64 kbps 64

Min. bitrate 16 kbps

16 0

t t1

Capacity Request (TVM Low)

t2

t3

Capacity Request (TVMHigh)

t4

t5 Capacity Request (TVMHigh)

HSDPA RRM • • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH HSDPA Improvements – – – – – – –

64QAM (RAN1643) MIMO (RAN1642) MIMO 42Mbps (RAN1912) Dual-Cell HSDPA (RAN1906) DC-HSDPA with MIMO 84Mbps (RAN1907) Flexible RLC in DL (RAN1638) Dual Band HSDPA (RAN2179) (RU50)

• Other Features • Appendix

Multicarrier HSPA Evolution in Release 9/10 & beyond •

3GPP Rel. 7 UE can receive and transmit only on 1 frequency even if the operator has total 3-4 frequencies Uplink 1 x 5 MHz

• •

Downlink 1 x 5 MHz

Rel. 8 brought DC-HSDPA, Rel. 9 defined DC-HSUPA Further Releases bring multicarrier HSDPA which allows UE to take full benefit of the operator’s spectrum Uplink

Downlink

2 x 5 MHz

8 x 5 MHz

HSPA Data Rate Evolution RU20 / RU30 / RU40 / RU50

3GPP R5 14 Mbps

0.4 Mbps

3GPP R6 14 Mbps

3GPP R7 21-28 Mbps

64QAM or 16QAM + MIMO (2x2)

5.8 Mbps

11 Mbps

16QAM

3GPP R8 42 Mbps

3GPP R9 84 Mbps

3GPP R10 168 Mbps

4-carrier DC-HSDPA HSDPA DC-HSDPA, + 64QAM + 64QAM + 64QAM + MIMO + MIMO MIMO (2x2) (2x2) (2x2)

11 Mbps

16QAM

23 Mbps

23 Mbps

3GPP R11 336 Mbps

8-carrier HSDPA + 64QAM + MIMO (2x2) or 4-carrier HSDPA + 64QAM + MIMO (4x4) 70 Mbps

DC-HSUPA DC-HSUPA + 64QAM DC-HSUPA + 16QAM + MIMO (2x2) + 16QAM

64QAM: RAN1643

HSDPA64QAMAllowed WCEL; 0 (Disabled), 1 (Enabled)

64QAM 6 bits/symbol

Modulation

QPSK

16QAM

64QAM

Coding rate

15 codes

1/4

1.8 Mbps

2/4

3.6 Mbps

3/4

5.4 Mbps

2/4

7.2 Mbps

3/4

10.8 Mbps

4/4

14.4 Mbps

3/4

16.2 Mbps

5/6

18.0 Mbps

4/4

21.6 Mbps

• optional Feature; RNC License Key required (ON-OFF) • HSDPA peak rate up to 21.1 Mbps • UE categories 13,14,17 & 18 supported • optional feature for UE Prerequisites: • Flexible RLC, HSDPA 14.4 Mbps, Dynamic Resource Allocation, HSUPA

HSDSCH category

max. HSDSCH Codes


13

15

1

14

15

1

17

15

1


17.4 or 23.4 Mbps

18

15

1


21.1 or 28 Mbps

MIMO support

Peak Rate

QPSK/16QAM/ 64QAM

No

17.4 Mbps

QPSK/16QAM/ 64QAM

No

21.1 Mbps

Modulation

64QAM: Channel Quality Requirements •

good channel conditions required to apply / take benefit of 64QAM – –

1/6

CQI 26 !

64QAM requires 6 dB higher SNR than 16QAM average CQI typically 20 in the commercial networks

1/4

2/4

3/4

QPSK no gain from 64QAM

0 Mbps



2/4

3/4

2/4

3/4

5/6 4/4

CQI > 15

CQI > 25

16QAM

64QAM

some gain from 64QAM 10 Mbps

only available with 64QAM

14 Mbps

21 Mbps

64QAM: CQI Tables TS 25.214: Annex Table 7d Cat 10 UE

TS 25.214: Annex Table 7f Cat 13 UE

TS 25.214 Annex Table 7g Cat 14 UE: CQI29: 14 Codes; 32257 bit CQI30: 15 Codes; 38582 bit

CQI

TB Size

# codes Modulation

CQI



TB Size

# codes Modulation



1

137

1

QPSK

0

1

136

1

QPSK

0

2

173

1

QPSK

0

2

176

1

QPSK

0

3

233

1

QPSK

0

3

232

1

QPSK

0

4

317

1

QPSK

0

4

320

1

QPSK

0

5

377

1

QPSK

0

5

376

1

QPSK

0

6

461

1

QPSK

0

6

464

1

QPSK

0

7

650

2

QPSK

0

7

648

2

QPSK

0

8

792

2

QPSK

0

8

792

2

QPSK

0

9

931

2

QPSK

0

9

928

2

QPSK

0

10

1262

3

QPSK

0

10

1264

3

QPSK

0

11

1483

3

QPSK

0

11

1488

3

QPSK

0

12

1742

3

QPSK

0

12

1744

3

QPSK

0

13

2279

4

QPSK

0

13

2288

4

QPSK

0

14

2583

4

QPSK

0

14

2592

4

QPSK

0

15

3319

5

QPSK

0

15

3328

5

QPSK

0

16

3565

5

16-QAM

0

16

3576

5

16-QAM

0

17

4189

5

16-QAM

0

17

4200

5

16-QAM

0

4672

5

16-QAM

0

18

4664

5

16-QAM

0

18

19

5287

5

16-QAM

0

19

5296

5

16-QAM

0

20

5887

5

16-QAM

0

20

5896

5

16-QAM

0

21

6554

5

16-QAM

0

21

6568

5

16-QAM

0

22

7168

5

16-QAM

0

22

7184

5

16-QAM

0

23

9719

7

16-QAM

0

23

9736

7

16-QAM

0

24

11418

8

16-QAM

0

24

11432

8

16-QAM

0

25

14411

10

16-QAM

0

25

14424

10

16-QAM

0

26

17237

12

16-QAM

0

26

15776

10

64-QAM

0

27

21754

15

16-QAM

0

27

21768

12

64-QAM

0

28

23370

15

16-QAM

0

28

26504

13

64-QAM

0

29

24222

15

16-QAM

0

29

32264

14

64-QAM

0

30

25558

15

16-QAM

0

30

32264

14

64-QAM

-2

64QAM: Link Simulations • UE peak data rate increased to 21.1 Mbps (L1 - theoretical) • Max application level throughput ~17.9 Mbps (ideal channel) • 64QAM is applicable for better radio conditions

64QAM Parameter – Bitrate control MaxBitRateNRTMACDFlow can be used to restrict the maximum bit rate of NRT MAC-d flow. The bit rate used in the reservation of the resources for the MAC-d flow is the minimum value of 1) max. bit rate based on UE capability, 2) max. bit rate of the RAB, 3) activated HSDPA bit rate features and 4) the value of this parameter. This parameter does not limit the maximum instantaneous bit rate on air interface. The value of the parameter is compared to the user bitrate of the NRT MAC-d flow excluding MAC-hs header, RLC header and padding. RNHSPA; 128..83968; 128; 65535*

Features enabled No license for HSDPA 15 codes 10 / 15 codes

Suggested Parameter Setting 3456 kbps 6784 kbps

10 / 15 codes & 10Mbps per user

9600 kbps

10 / 15 codes & 14Mbps per user 10 / 15 codes & 64 QAM MIMO DC HSDPA DC HSDPA & MIMO

13440 kbps 21120 kbps 27904 kbps 42112 kbps 84224 kbps

MIMO Principle Signal from jth Tx antenna

S j

Input

T1

R1

T2

R2

• • •

• • •

Tm

MxN MIMO system

MIMO Processor

Output

Rn

• MIMO: Multiple-Input Multiple Output • M transmit antennas, N receive antennas form MxN MIMO system • huge data stream (input) distributed toward m spatial distributed antennas; m parallel bit st reams (Input 1..m) • Spatial Multiplexing generate parallel “virtual data pipes” • using Multipath effects instead of mitigating of mitigating them

MIMO Principle Signal from jth Tx antenna

Signal at ith Rx antenna

h1,1

S j

T1

hn,1

h2,1

R1

h1,m

h1,2

Yi

MIMO

h2,2

Input

hn,2

T2

• • • Tm

R2

h2,m

hn,m

• • •

MxN

Rn

P r o c e s s o r

Output

MIMO

H=

h1,1

h1,2



h1,m

• Receiver learns Channel learns Channel Matrix H

h2,1

h2,2



h2,m

• inverted Matrix H-1 used for recalculation of original input data streams 1..m

  

  

hn,1

hn,2

   

hn,m

yi 

m

h

i , j

j 1

 s j  ni

ni: Noise at receiver

MIMO: RAN1642

MIMOEnabled

• RU20 (3GPP RU20 (3GPP Rel. 7) introduces 2x2 MIMO with MIMO with 2-Tx/2-Rx

WCEL; 0 (Disabled), (Disabled), 1 (Enabled)

– Double Transmit Transmit on BTS side (D-TxAA ( D-TxAA), ), 2 receive antennas on UE side – System can operate in dual stream (2x2 MIMO) or MIMO) or single stream (Tx diversity) mode

• MIMO 2x2 enables 28 Mbps peak Mbps peak data rate in rate in HSDPA – 28 Mbps peak rate in combination with 16QAM – 64QAM: no simultaneous support of 64QAM & MIMO (not yet) – Dual-Cell HSDPA: HSDPA: not possible to enable MIMO & DC-HSDPA in a cell in parallel

WBTS: 2 Txantennas

• Benefits: MIMO increases single user peak data rate, overall cell capacity, average cell throughput & coverage

• UE categories for MIMO support: Cat. 15, 16, 17 & 18 UE: 2 Rxantennas

• optional Feature (ASW) • RNC License Key required (ON-OFF) (ON-OFF)

HSDSCH category

max. HSDSCH Codes


Modulation

MIMO support

Peak Rate

15

15

1

QPSK/16QAM

Yes

23.4 Mbps 23.4 Mbps

16

15

1

QPSK/16QAM

Yes

28 Mbps 28 Mbps

17

15

1


17.4 or 23.4 or 23.4 Mbps

18

15

1


21.1 or 28 or 28 Mbps

Prerequisites:

• double Power Amplifier units & antenna lines per cell; • must be enabled: HSDPAEnabled, HSUPAEnabled, HSDPA14MbpsPerUser, HSDPADynamicResourceAllocation, FDPCHEnabled, HSDPAMobility, FDPCHEnabled, FRLCEnabled; must not be enabled: DCellHSDPAEnabled

MIMO S-CPICH Power & Code allocation

• MIMO enabled cell: S-CPICH is broadcast for DL channel estimation in UE

tx power = S-CPICH

– S-CPICH transmission power is controlled with existing parameter

PtxPrimaryCPICH -10..50; 0.1; 33 dBm

• UE must be able to estimate each of the 2 signals separately – P-CPICH is broadcast along with data stream 1 – S-CPICH (new with RU20) is broadcast along with data stream 2 – SF 256 spreading code must be allocated in DL to support S-CPICH transmission

SF 16,0

SF 32

S-CCPCH

SF 64

E-RGCH HS-SCCH

E-HICH SF 128

SF 256

0

1

2

3

4

5

6

depending on FACH / PCH configuration

7

8

9

10

11

12

13

14

15

RNC checks following conditions, before MIMO allocation to a UE:

MIMO 2x2 / 28 Mbps Allocation of MIMO for a UE

Start

(if at least one of the conditions is false during active MIMO allocation, MIMO will be deactivated) no

MIMO Parameter Enabled optional feature for UE

yes

BTS is MIMO capable

MIMOEnabled WCEL; 0 (Disabled), 1 (Enabled)

no

yes

UE is MIMO capable

no

yes

RAB configuration for UE allows MIMO

no

yes

Streaming RAB state changes to inactive

no

yes F D PCH E nabled WCEL; 0 (Disabled), 1 (Enabled) F RL C E nabled WCEL; 0 (Disabled), 1 (Enabled)

SRB* can be mapped to HSPA (F-DPCH)

no

yes

MAC-ehs can be allocated (Flexible RLC)

no

yes

yes – allocate MIMO

no – do not allocate MIMO * i.e. SRB must be mapped to HSPA

MIMO: Layering & Mobility Layering:

• RU20 MIMO supports following site configurations: – 1 / 1 / 1 – 2 / 2 / 2 – 3 / 3 / 3 • more than one MIMO layer not possible in RU20.

Mobility • Once allocated to a UE, MIMO will be kept also during mobility procedures – Service Cell Change can be used to allocate / de-allocated MIMO for a UE

– If target cell is not supporting MIMO or MIMO can not be enabled, RNC deactivates MIMO for the UE

• Compressed Mode is started for a UE having MIMO allocated

• MIMO Mobility over Iur interface NOT supported in RU20

MIMO Layer

MIMO 2x2 / 28 Mbps Performance mean cell throughput vs.

UE throughput at the Cell Edge, middle of the cell & cell center

various scheduling schemes

Single-stream

Dual-stream

Single-stream

CLM: Closed Loop Mode; Single-Stream with Rx- & Tx-Diversity

Dual-stream

MIMO 42Mbps (RAN1912) 64QAM 6 bits/symbol

2x2 MIMO & 64QAM up to 42 Mbps

WBTS: 2 Txantennas

2x2 MIMO MIMOWith64QAMUsage

Basics: WCEL; 0 (Disabled), 1 (Enabled) • optional Feature; RNC License Key required (ON-OFF) • RU20 enables either 2x2 MIMO (RAN1642) or 64QAM (RAN1643) • RU30 enables simultaneous 2x2 MIMO and 64QAM operation (RAN1912) • Peak Rates: up to 2 x 21 Mbps = 42 Mbps HSmax. HSMIMO Peak • 3GPP Rel. 8 DSCH DSCH Modulation support Rate category Codes • new UE Categories: 19, 20 19

15

QPSK/16QAM/ 64QAM

Yes

35.3 Mbps

20

15

QPSK/16QAM/ 64QAM

Yes

42.2 Mbps

Requirements

• Flexible RLC, F-DPCH, MIMO 28 Mbps, HSDPA 64QAM

Allocating MIMO 42Mbps • 64QAM is allocated with MIMO whenever possible • Switching can occur when conditions change, i.e. when it becomes possible to support MIMO with 64QAM, or when it is no longer possible to support MIMO with 64QAM • The conditions required to support MIMO 42Mbps are: – – – –

it must be possible to support MIMO it must be possible to support HSDPA 64QAM The WCEL MIMOWith64QAMUsage parameter must be set to enabled The BTS and UE must support simultaneous use of MIMO and 64QAM

• If MIMO with 64QAM is not possible but MIMO without 64QAM, or 64QAM without MIMO is possible, MIMO shall be preferred

DC-HSDPA Principles • prior to 3GPP Release 8, HSDPA channel bandwidths limited to 5 MHz • Dual-Cell HSDPA : 3GPP Rel. 8 allows 2 adjacent channels t o b e c o m b i n e d  effective

HSDPA channel bandwidth of 10 MHz (RU20 feature)

• 3GPP Rel. 8: Dual Cell HSDPA can be combined with 64QAM but not with MIMO (Release 9 allows combination with both, 64QAM & MIMO) 

DCell H SDPAEnabled

42 Mbps HSDPA peak rate

WCEL; 0 (Disabled), 1 (Enabled)

Basic Approach

Dual Cell A p p r o a c h

2 UE, each using 5 MHz RF Channel Peak Connection Throughput = 28 Mbps 5 MHz

5 MHz

F1

F2

MIMO (28 Mbps), or 64QAM (21 Mbps)

1 UE, using 2 × 5 MHz RF Channels Peak Connection Throughput = 42 Mbps 10 MHz F1

F2

DC-HSDPA & 64QAM (42 Mbps)

DC-HSDPA Principles • DC-HSDPA provides greater flexibility to the HSDPA Scheduler, i.e. the scheduler can allocated resources in the frequency domain as well as in the code and time domains Gains o f DC-HSDPA:

1) Improved Load Balancing 2) Frequency Selectivity 3) Reduction of Latency 4) Higher Peak Data Rates 5) Improved Cell Edge “User Experience” Channel conditions good on both RF carriers

Channel conditions good on RF carrier 1

F2

F1

UE1

F1

UE1

Channel conditions good on RF carrier 2

F2

UEx

F2

F1

UEx

UE1

DC-HSDPA: UE Cat & Requirements • • • •

RU20 (3GPP Rel. 8) introduces DC-HSDPA (RAN1906) DC-HSDPA & 64QAM enable DL 42 Mbps peak rates UE categories for DC-HSDPA support: Cat. 21, 22, 23 & 24 optional feature; requires long term RNC license for specific number of cells

• following features must be enabled: • • • • • • • • •

HSDSCH category

max. HSDSCH Codes

Modulation

MIMO support

HSDPA (HSDPAEnabled) 21 15 QPSK/16QAM HSUPA (HSUPAEnabled)* 22 15 QPSK/16QAM HSDPA 15 codes (HS-PDSCHcodeset) QPSK/16QAM/6 HSDPA 14 Mbps per User (HSDPA14MbpsPerUser ) 23 15 4QAM HSDPA Serving Cell Change (HSDPAMobility) QPSK/16QAM/6 24 15 Fractional DPCH (FDPCHEnabled) 4QAM DL Flexible RLC (FRLCEnabled) Shared Scheduler for Baseband Efficiency HSPAQoSEnabled must be configured with the same value in both DC-HSDPA cells • MaxBitRateNRTMACDFlow (def. 65535 = not restricted) should be configured to allow the peak throughput • RU20: MIMO + DC-HSPDA must not be enabled for all cells belonging to the Node B ( MIMOEnabled); ; • RU40: MIMO + DC-HSDPA possible DC-HSDPA + MIMO possible in RU40

Peak Rate

No

23.4 Mbps

No

28 Mbps

No

35.3 Mbps

No

42.2 Mbps

DC-HSDPA: Requirements SectorID = 1

• DC HSDPA cells require: • adjacent RF carriers  UARFCN • same sector  SectorID • same Tcell value

SectorID = 3 SectorID = 2

RF Carrier 2 SectorID = 1

SectorID = 3 SectorID = 2

RF Carrier 1

DC-HSDPA: Tcell Configuration (I) Tcell = 0

• 2+2+2 Node B with DC-HSDPA requires: • each cell belonging to the same sector must have the same Tcell value • Tcell values belonging to different sectors must belong to different Tcell groups

• •

Tcell = 3 RF Carrier 2 Tcell = 0

Configuration requires 3 HSDPA Efficient Baseband Schedulers RF carriers 1 & 2 must be adjacent

Tcell = 6

Tcell = 6 Tcell = 3

RF Carrier 1

Tcell: defines start of SCH, CPICH, Primary CCPCH & DL Scrambling Code(s) in a cell relative to BFN

DC-HSDPA: Tcell Configuration (II) • 3+3+3 Node B with DC-HSDPA requires: • each DC-HSDPA cell belonging to same sector to have same Tcell value • DC-HSDPA Tcell values belonging to different sectors must belong to different Tcell groups

• Configuration requires 4 HSDPA Efficient Baseband Schedulers

Tcell = 0

• Cells belonging to RF carrier 3 must be

Tcell = 3

Tcell = 6 Tcell = 9

RF Carrier 2

within a further LCG

Tcell = 3 RF Carrier 1

LCG: Local Cell Group

Tcell = 1 Tcell = 2

RF Carrier 3

• RF carriers 1 & 2 must be adjacent • Cells belonging to RF carriers 1 & 2 must be within the same LCG

Tcell Groups • Group 1: Tcell values 0, 1, 2 • Group 2: Tcell values 3, 4, 5 • Group 3: Tcell values 6, 7, 8 • Group 4: Tcell value 9

Tcell = 6 Tcell = 9

DC-HSDPA: HSDPA Scheduler

• A single HSDPA shared scheduler for baseband efficiency is required per DC-HSDPA cell pair • 3 HSDPA shared schedulers are required for a 2+2+2 Node B configuration with DC-HSDPA • Each scheduler is able to serve both HSDPA & DC-HSDPA UE on both RF carriers • Link Adaptation is completed in parallel for each RF carrier

HSDPA UE on f 2 Shared Scheduler per DC-HSDPA cell pair

DC-HSDPA UE with serving cell on f 2 HSDPA UE on f 1 DC-HSDPA UE with serving cell on f 1

DC-HSDPA with MIMO 84Mbps 64QAM 6 bits/symbol

DC-HSDPA, 2x2 MIMO & 64QAM up to 84 Mbps

WBTS: 2 Txantennas

2x2 MIMO

Basics: • enables simultaneously: DC HSDPA, MIMO & 64QAM • MIMO uses Single Stream or Double Stream transmission • DC-HSDPA uses 2 cells (in 1 sector) at same BTS; same frequency band & adjacent carriers to a UE • 64QAM  6 bits/symbol Benefits: • higher Peak Rate: up to 2 x 2 x 21 Mbps = 84 Mbps • better Coverage due to DC-HSDPA & MIMO • More robust transmission due to MIMO & DC HSDPA usage

Dual-Cell (DC-) HSDPA

DC-HSDPA with MIMO 84Mbps 42 Mbps

42 Mbps

56 Mbps

MIMO + 64QAM

DB-DC-HSDPA + 64QAM

DC-HSDPA + MIMO

RAN1912 / 3GPP Rel. 7

RAN2179 / 3GPP Rel. 9

3GPP Rel. 9

max. Peak Rate in RU40

84 Mbps DC-HSDPA + MIMO + 64QAM

w/o 64QAM

both supported by RAN1907

3GPP Rel. 9

Feature Enabling: • DC-HSDPA with MIMO 84 Mbps: optional feature; but: w/o own license; required licenses:   

RAN1642 MIMO (28 Mbps) RAN1643 HSDPA 64QAM RAN1906 DC-HSDPA 42 Mbps

• DC-HSDPA + MIMO can be enabled w/o 64QAM  Peak

Rate up to 56 Mbps

• to enable Peak Rate = 84 Mbps



DCellAndMIMOUsage must be enabled &

DCellAndMIMOUsage WCEL; 0 (DC-HSDPA & MIMO disabled), 1 (DC-HSDPA & MIMO w/o 64QAM enabled), 2 (DC-HSDPA & MIMO with 64QAM enabled) MIMOWith64QAMUsage WCEL; 0 (Disabled), 1 (Enabled)

DC-HSDPA: UE Categories & Requirements

MaxBitRateNRTMACDFlow* can be used to restrict max. bit rate of NRT MAC-d flow

Requirements • RAN1642 MIMO 28 Mbps

RNHSPA; 128... 83968 ; 128; 0

UE Categories

• RAN1638 Flexible RLC • RAN1906 DC HSDPA • RAN1643 64QAM • RAN1912 MIMO 42Mbps

(3GPP Rel. 9; TS 25.306)

64QAM with MIMO (w/o DC-HSDPA) DC-HSDPA (w/o MIMO, 64QAM) DC-HSDPA with 64QAM (w/o MIMO) DC-HSDPA with MIMO (w/o 64QAM) DC-HSDPA with MIMO 84Mbps

value 0 / 65535 (before): HSDPA peak rate not limited by the RNC

HSDSCH category

max. HSDSCH Codes

Modulation

MIMO support

DCHSDPA support

19

15

QPSK/16QAM/ 64QAM

Yes

No

20

15

QPSK/16QAM/ 64QAM

Yes

No

21

15

QPSK/16QAM

No

Yes

23.4 Mbps

22

15

QPSK/16QAM

No

Yes

28 Mbps

23

15

QPSK/16QAM/ 64QAM

No

Yes

35.3 Mbps

24

15

QPSK/16QAM/ 64QAM

No

Yes

42.2 Mbps

25

15

QPSK/16QAM

Yes

Yes

46.7 Mbps

26

15

QPSK/16QAM

Yes

Yes

56 Mbps

27

15

QPSK/16QAM/ 64QAM

Yes

Yes

70.6 Mbps

28

15

QPSK/16QAM/ 64QAM

Yes

Yes

84.4 Mbps

Peak Rate 35.3 Mbps 42.2 Mbps

DC-HSDPA: Mobility Hard Handover HHO

• DC-HSDPA with MIMO can be maintained, activated or de-activated during mobility • Availability of DC-HSDPA with MIMO checked in target cell when SCC or HHO initiated • If DC-HSDPA with MIMO cannot be used in the target cell mobility proceeds without it: – DC-HSDPA or MIMO is used if possible, according to the parameter DCellVsMIMOPreference

• If HSUPA IFHO can be used DC-HSDPA & MIMO is not be deactivated but is maintained during Inter-Frequency measurements • If HSUPA IFHO cannot be used, E-DCH to DCH switch is completed before inter -frequency measurements; DC-HSDPA with MIMO is deactivated at the same time

• DC-HSDPA with MIMO is not supported across the Iur • S-RNC does not configure DC-HSDPA with MIMO if there are radio links over the Iur in the active set

DCellVsMIMOPreference RNHSPA; DC-HSDPA preferred (0), MIMO preferred (1)

defines whether RNC primarily activates DC-HSDPA or MIMO for a UE, which supports both DC-HSDPA & MIMO in case simultaneous usage of DC-HSDPA & MIMO is not possible. SCC: Serving Cell Change

DC-HSDPA: Gain in Throughput & Coverage Gain of DC-HSDPA & MIMO compared to SCHSDPA: • Throughput: + 220% • Coverage: + 57% Furthermore: Some 29% more subscriber can be served

T h r M o u o g r h e p u t

more Coverage

SC-HSDPA: Single Carrier HSDPA DC-HSDPA: Dual-Carrier HSDPA TP: Throughput

Flexible RLC (DL): RAN1638 • included in RU20 basic software package – no license needed • HW Prerequisites: Flexi Rel2, UltraSite with EUBB • Flexible RLC used, if: – Cell Flexible RLC capable & enabled – UE supports Flexible RLC – AM RLC is used – HS-DSCH & E-DCH selected as transport channels – Dynamic Resource Allocation enabled

prior Rel. 7

PDCP

IP packet (max. 1500 byte) segmentation

RLC

••• RLC PDU: 336 bit or 656 bit 16 bit RLC Header  4.8% or 2.4% Overhead

concatenation

MAC-hs

FRLCEnabled RNFC; 0 (Disabled), 1 (Enabled)

Rel. 7 Flexible RLC IP packet (max. 1500 byte) no segmentation IP packet (max. 1500 byte) adapts RLC-PDU size to actual size of higher layer data unit

segmentation

TBS (depending on scheduling) AM: Acknowledged Mode

DL Flexible RLC Background

• Prior to Rel. 7: RLC layer segments high layer data units (IP packets) in RLC PDU sizes of 336 and 656 – 336 is 320 net bit plus 16 bit RLC OH – 656 is 640 net bit plus 16 bit RLC OH

• On MAC-d layer did not increase Overhead – Data was passed directly to MAC-hs layer (MAC-d)

• Several MAC-d PDUs were concatenated to form a MAC-hs data block • BTS selects proper MAC-hs data block size based on – available user date in BTS buffer and – radio conditions for that UE

• With DL Flexible RLC the RNC adapts the RLC-PDU size to the actual size of the higher layer data unit (IP) – maximum size of 1500 Byte is supported (IP packet length in Ethernet)

DL Flexible RLC Advantages

• Major improvements with DL Flexible RLC – – – – – –

less processing in RNC & UE higher end user application throughput lower latency for packet access Significantly lower Overhead Much less padding bits Lower risk for RLC stalling because of too small transmission windows 50% 45% Rel. 6 with RLC PDU Size of 336 bits

40%

d a e h r e v O

Rel. 6 with RLC PDU Size of 656 bits 35% Rel. 7 Flexible RLC 30% 25% 20% 15% 10% 5% 0% 0

100

200

300

400

500

600

700

800

900

1000 1100 1200

1300 1400 1500

IP packet size [byte]

Dual Band HSDPA: With and Without the Feature (RU50) • This feature introduces for a single UE the possibility of using simultaneously two carriers in DL that are situated on two different WCDMA frequency bands

• Feature enables achieving 42 Mbps peak rate for user in DL (assuming 64QAM and 15 codes usage on both frequencies) Without DB-HSDPA feature there is no possibility to establish data connection with to different band at the same time DC-HSDPA DL transmission options 2 x 5 MHz f 1

2 x 5 MHz f 1

f 2

U2100

f 2

U900

SC-HSDPA DL transmission options 5 MHz

5 MHz

f 1

f 1

U2100

U900

With DB-HSDPA feature there is possibility to establish data connection with to different band at the same time DB-HSDPA DL transmission options 2 x 5 MHz f 1

f 2

U2100

U900

DBandHSDPAEnabled WCEL; (0) Disabled, (1) Enabled

n o r e t a l d e t n e s e r p s n o i t p o s n o i t a r u g i f n o c d e l i a t e d y r a l p m e x e y l n o e r a s d n a b y c n e u q e r f d e t n e s e

HSDPA RRM • • • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH HSDPA Improvements Other Features – – – – –

Continuous Packet Connectivity CPC (RAN1644) CS Voice over HSPA (RAN1689) Fast Dormancy (RAN2136) Fast Dormancy Profiling (RAN2451) High Speed Cell_FACH (DL) (RAN1637)

• Appendix

CPC: Continuous Packet Connectivity Introduction

• Discontinuous UL DPCCH Transmission & Reception during UE UL traffic inactivity (UL DPCCH gating + DRX at BTS) – CQI reporting reduction (switched from periodical to synchronized with DPCCH burst) – Stopping E-DPCCH detection at NodeB during DPCCH inactivity

CPCEnabled WCEL; 0 (Disabled), 1 (Enabled)

• Discontinuous DL Reception (DRX at UE) – Stop receiving HS-SCCH, E-AGCH & E-RGCH when not needed

• Faster response times – Increased number of low activity packet users in CELL_DCH state

Motivation / Benefits: • Increased capacity for low data rate applications • Longer battery life • Network: – optional feature; ON-OFF RNC License • Prerequisites: – UE must support CPC – F-DPCH enabled

CPC “Sub-features”: • UL DPCCH Gating (UL DTX) • CQI Reporting reduction • Discontinuous UL Reception (MAC DTX) • Discontinuous DL Reception (DL DRX)

CPC: UL Gating (UL DTX) UL Gating (UL DTX): reduces UL control channel (DPCCH) overhead • no data to sent on E-DPDCH or HS-DPCCH  UE switchs off UL DPCCH • D P CC H G a t i n g i s p r e c o n d i t i o n f o r o t h e r 3 s u b - f e at u r e s

Rel99 Service

DPDCH

Voice (20ms)

DPCCH

Rel6 Voice 2ms (Rel6 VoIP)

Rel7 Voice 2ms (Rel7 VoIP) UL DPCCH Gating

E-DPDCH DPCCH

E-DPDCH DPCCH

CPC: UL Gating UL DPCCH Gating (UL DTX)

• UE specific Packet Scheduler provides CPC parameters • These are service & UL TTI specific & part of parameter groups – Voice 2ms, 10ms; RNHSPA: CPCVoice10msTTI , CPCVoice2msTTI – Streaming 2ms, 10ms; RNHSPA: CPCStreaming10msTTI , CPCStreaming2msTTI – Interactive, Background 2ms, 10ms; RNHSPA: CPCNRT10msTTI , CPCNRT2msTTI Following parameters are parameters from CPCNRT2msTTI group (per sub-feature): DPCCH Gating (UL DTX): • N2msInacThrUEDTXCycl2 : number of consecutive E-DCH TTIs without an E-DCH transmission, after which the UE should immediately move from UE DTX Cycle 1 to UE DTX Cycle 2. RNHSPA; Range:1 (0), 4 (1), 8 (2), 16 (3), 32 (4), 64 (5), 128(6), 256 (7); default: 64 (5) TTIs

• N2msUEDPCCHburst1: UL DPCCH burst length in subframes when UE DTX Cycle 1 is applied. RNHSPA; Range:1 (0), 2 (1), 5 (2); default: 1 (0) subframes

• N2msUEDPCCHburst2: UL DPCCH burst length in subframes when UE DTX Cycle 2 is applied. RNHSPA; Range:1 (0), 2 (1), 5 (2); default: 1 (0) subframes

• N2msUEDTXCycle1: UL DPCCH burst pattern length in subframes for UE DTX Cycle 1. RNHSPA; Range: 1 (0), 4 (1), 5 (2), 8 (3), 10 (4), 16 (5), 20 (6); default: 8 (3) subframes

• N2msUEDTXCycle2: UL DPCCH burst pattern length in subframes for UE DTX Cycle 2. RNHSPA; Range: 4 (0), 5 (1), 8 (2), 10 (3), 16 (4), 20 (5), 32 (6), 40 (7), 64 (8), 80 (9), 128 (10), 160 (11); default: 16 (4) subframes

CPC: UL Gating / DPCCH Gating 2m s TT I UL Gating, E-DCH 2ms TTI example: CPCNRT N2msInacThrUEDTXCycl 2 CFN

10ms Radio Frame

10ms Radio Frame

RNHSPA; 1, 4, 8, 16, 32, 64, 128, 256; 64 TTIs

10ms Radio Frame

Inactivity Threshold for UE cycle 2 E-DPDCH Tx, 2ms TTI

no data on E-DPDCH

N2msUEDPCCHburst 1

DPCCH

RNHSPA; 1, 2, 5; 1 subframe(s)

pattern

N2msUEDPCCHburst 2 synch reference

UE_DTX_Cycle_ 1 UE_DTX_Cycle_ 2

RNHSPA; 1, 2, 5; 1 subframe(s)

N2msUEDTXCycle 1 RNHSPA; 1, 4, 5, 8, 10, 16, 20; 8 subframes

UE_DTX_Cycle_2

N2msUEDTXCycle 2 DPCCH with

RNHSPA; 4, 5, 8, 10, 16, 20, 32, 40, 64, 80, 128, 160; 16 subframes

E-DCH, 2ms TTI 2ms subframe UE_DTX_DRX_offset is UE specific offset granted from BTS

cycle 1

cycle 2

switch to UE cycle 2 CFN: Connection Frame Number; used for any synchronized procedure in UTRAN Pre/Postambles not shown here

CPC: Reduced CQI Reporting CQI Reporting reduction: • CQI Reporting Reduction r e d u c e t h e T x p o w e r o f t h e U E b y r e d u c i n g t h e C Q I r ep o r t i n g ; this means to reduce the interference from HS-DPCCH in UL when no data is transmitted on HS-PDSCH in DL

• Reduced CQI reporting takes place only if the CQI reporting pattern defined by the last HS-DSCH transmission and CQI cycle overlaps the UL DPCCH burst of the UE DTX pattern

• N2msCQIDTXTimer : defines the number of subframes after an HS-DSCH reception, during which the CQI reports have higher priority than the DTX pattern . RNHSPA; 0 (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10), Infinity (11); 64 (7) subframes

• N2msCQIFeedbackCPC : defines the CQI feedback cycle for HSDPA when the CQI reporting is not reduced because of DTX. RNHSPA; 0 (0), 2 (1), 4 (2), 8 (3), 10 (4), 20 (5), 40 (6), 80 (7), 160 (8); default: 8 (3) ms; Note: CQI transmission time defined by Bigger CQI reporting cycles 10ms are not recommended. HS-DSCH reception

CQI_DTX_Priority

CQI_DTX_TIMER

ACK/NACK transmission

7.5 slots

CQI_DTX_Priority

set to 0

set to 1

DPCCH pattern UE_DTX_cycle_1

UE_DTX_cycle_1

UE_DTX_cycle_2

CQI transmission CQI period 2ms CQI period 4ms

CQI period 8ms

UE_DTX_cycle_2

UE_DTX_cycle_2

CQI period, but not overlapping with DPCCH transmission no CQI transmission CQI Transmission

CPC: Discontinuous UL & DL Reception (MAC DTX & DL DRX) During E-DCH inactivity, E-DPCCH detection happens at the BTS only every MAC_DTX_Cycle subframes. It is stopped at Node B after MAC_inactivity_threshold subframes of E-DCH inactivity. As a consequence, the UE experiences a delay regarding the transmission start time. The UE-specific offset parameter UE_DTX_DRX_Offset allows to stagger the processing of several UEs in time to save the BTS resources.

Discontinuous UL Reception (MAC DTX):

• N 2 m s M A C D T X C y c l e: length of MAC DTX Cycle in subframes. This is a pattern of time instances where the start of the UL E-DCH transmission after inactivity is allowed. RNSHPA; Range: 1 (0), 4 (1), 5 (2), 8 (3), 10 (4), 16 (5), 20 (6); default: 8 (3) subframes

• N 2 m s M A C In a c T h r : E-DCH inactivity time in TTIs after which the UE can start E-DCH transmission only at given times. RNHSPA; Infinity (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10) TTIs; default: Infinity (0)

Discontinuous DL Reception (DL DRX):

• N 2 m s I n a c Th r U E D R X Cy c l e : number of subframes after an HS-SCCH reception or after the first slot of an HS-PDSCH reception, during which the UE is required to monitor the HS-SCCHs in the UE's HS-SCCH set continuously. RNHSPA; Range: 0 (0), 1 (1), 2 (2), 4 (3), 8 (4), 16 (5), 32 (6), 64 (7), 128 (8), 256 (9), 512 (10); default: 64 (7) subframes

• N 2 m s U E D R XC y c l e : HS-SCCH reception pattern (UE DRX Cycle) length in subframes. This parameter is a multiple or a divisor of the parameter UE DTX Cycle 1. If the value is not allowed, the parameter value minus 1 is used to calculate a new value, and so on. RNHSPA; Range: 0.5 (0), 1 (1), 2 (2), 3 (3), 4 (4); default: 2 (2) subframes

CPC: Discontinuous UL Reception Discontinuous UL Reception (MAC-DTX) – NSN implemented parameters

N2msMACInacThr RNHSPA; Infinity, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512; Infinity subframes

DTX

N2msMACDTXCycle

UE can transmit E-DPDCH data only at predefined time instances.

length of MAC DTX Cycle RNHSPA; Infinity, 1, 4, 5, 8, 10, 16, 20; 8 subframes

CPC: Discontinuous DL Reception Discontinuous DL Reception (DL DRX)

• When the UE DRX is enabled, the U E m a y t u r n o f f t h e r e c ei v e r w h e n t h e r e i s n o n e e d t o r e c e i v e anything in DL

• The DL DRX can be enabled only in conjunction with UL DTX N2msInacThrUEDRXCycle UE DRX Inactivity threshold RNHSPA; 0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512; 64 subframes

DRX N2msUEDRXCycle length of UE DRX Cycle RNHSPA; 0.5, 1, 2, 3, 4; 2 subframes

DL DRX only with UL DTX !

• N2msInacThrUEDRXCycle : number of subframes after an HS-SCCH reception or after the 1st slot of an HS-PDSCH reception, during which the UE is required to monitor the HS-SCCHs in the UE's HS-SCCH set continuously; UE DRX Inactivity threshold; RNHSPA; 0, 1, 2, 4, 8, 16, 32, 64, 128, 256, 512; 64 subframes • N2msInacThrUEDRXCycle : HS-SCCH reception pattern (UE DRX Cycle) length in subframes; RNHSPA; 0.5, 1, 2, 3, 4; 2 subframes

CPC & Power Control Power Control

• New parameter introduced to control step size for DL Inner Loop PC

DLInLoopPCStepSizeCPC RNSPA: 0.5..2; 0.5; 1.5 dB

DLInLoopPCStepSizeCPC: used by the W CDMA BTS to calculate the power increase/decrease step size when receiving TPC commands. It is applied when CPC (UE DTX, etc.) is activated for the UE. Note: If CPC is not used for a UE, BTS applies DownlinkInnerLoopPCStepSize DownlinkInnerLoop PCStepSize

RNAC: 0.5..2; 0.5; 1 dB

CPC: Extra-inactivity timer for Transition from CELL_DCH to CELL_FACH CELL_ DCH

UE

Node B

InactivityTimerDownlinkDCH InactivityTimerUplinkDCH

RNC

Range: 0 .. 20 s; Step: 1 s; default: • for 8, 16 & 32 kbps: 5 s • for 64 kbps: 3 s • for 128, 256, 320 & 384 kbps: 2 s

PDU Transport on the DCH/DPCH All data sent & RLC-U buffer empty Inactivity detected

Start

InactivityTimerDownlinkDCH InactivityTimerUplinkDCH Radio Bearer Reconfiguration CELL_ FACH

Expiry

Radio Bearer Reconfiguration Complete

as soon as L2 in RNC indicated RB inactivity, RNC allocates “ extra inactivity timer ” to keep the UE in Cell_DCH This depends on: – CPC is allocated for a UE or not (CPC or NonCPC) – UE Device Type – RNC knows from UE capabilities  UE benefits / does not benefit from Power Consumption Optimization (BatOpt / NoBatOpt)

InactCPCNoBatOptT: 180 s InactCPCBatOptT: 0 s InactNonCPCNoBatOptT: 0 s InactNonCPCBatOptT: 0 s all parameters: RNHSPA; 0s..48h & infinity; several steps;

CS Voice Over HSPA (RAN1689) Requirements Network: • optional RU20 feature; ON-OFF RNC License

HSPAQoSEnabled WCEL; 0..4*; 1; 0 = disabled 0 = QoS prioritization is not in use for HS t ransport 1 = QoS prioritization is used for HS NRT channels 2 = HSPA streaming is in use 3 = HSPA CS voice is in use 4 = HSPA streaming & CS voice are in use

UE: • must support CSvoiceOverHSPA • optional feature in Rel. 7/8

required Network Features: • • • • •

QoSPriorityMapping RNPS; 0..15; 1; 14 for CS Voice over HSPA • Priority must be lower than SRB (15) • Priority must be higher than Streaming 13)

HSDPA Dynamic Resource Allocation QoS Aware HSPA Scheduling CPC F-DPCH HSPA with simultan. AMR Voice

• SRB must be mapped to HSPA • supported RAB combinations: • • • • •

Speech CS RAB Speech CS RAB + PS streaming PS RAB Speech CS RAB + 1...3 IA/BG PS RABs Speech CS RAB + PS Streaming PS RAB + 1...3 IA/BG PS RABs

Load based AMR selection algorithm not used while CS Voice is mapped on HSPA BG: Background IA: Interactive

for Voice, SRB & other services Codecs supported for CS Voice Over HSPA: • AMR (12.2, 7.95, 5.9, 4.75), (5.9, 4.75) & (12.2) • AMR-WB (12.65, 8.85, 6.6) * if HSPA streaming or CS voice is activated, then QoS prioritization for NRT HSPA connections is in use, too

Enabling the feature: CS Voice Over HSPA Pre-conditions

• CS voice over HSPA license exists & state is 'On‘ • HSDPA with Simultaneous AMR Voice Call license exists & state is 'On' • HSUPA with Simultaneous AMR Voice Call license exists & state is 'On' • AMRWithHSDSCH & AMRWithEDCH : HSPA with Simultaneous AMR Voice Call enabled • HSDPAenabled & HSUPAenabled : HSPA enabled in all Active Set cells • HSDPA Dynamic Resource Allocation license exists & state is 'On‘ • H S D P A D y n a m i c R e s o u r c e A l l o c a t i o n is enabled • QoS Aware HSPA Scheduling license exists & state is 'On‘ • HSPAQoSEnabled is set to “HSPA CS voice” in all Active Set cells • CPC & Fractional DPCH licenses exists & state is 'On‘ • CPCEnabled in all Active Set cells • FDPCHEnabled : Fractional DPCH enabled in all Active Set cells

CS Voice Over HSPA Efficiency

• Two different voice transmission scenarios are being considered with IP: – VoIP – UE connects with network as in standard Packed Data transmission and by using “web communicators” a connection can be established (hard to establish appropriate charging schemes)

– CS voice over IP – voice is being carried by HSPA transport channels transparent for the user Assumed IP Header Compression

[REF. WCDMA for UMTS – HSPA Evolution and LTE, HH AT]

CS Voice Over HSPA Concept / Protocol Stack • In UL there is a so called Dejitter buffer implemented in RNC PDCP • used to align the UL data stream before routing to MSC or MSS system • In DL MAC-ehs is used to support flexible RLC PDU sizes • supporting different AMR rates

CS Core

CS Voice over DCH

RAN CS Voice over HSPA

CS Core

Dejitter buffer

RAN

PDCP

TM RLC

UM RLC

DCH

HSPA

• Inter system mobility between 2G & 3G is as today, the CS Voice Over HSPA is just RAN internal mapping and it is not visible outside of the RAN. Handover signaling is not affected and RAN provides the measurement periods for UE using compressed mode as today

• AMR rate adaptation can be used to provide even higher capacity gains by lowering the AMR coding rate • Voice related RRM algorithms like pre-emption are expanded to cover also the Voice Over HSPA • Air interface capacity gain of the feature depends on parameterisation of HSUPA including CPC parameters, allowed noise rise and voice activity

CS Voice Over HSPA Admission Control: CS Voice over HSPA connection admitted if:

PtxNCTDCH + PtxNCTHSDPA + Pnew < PtxTargetTot && PtxNCTHSDPA + Pnew < PtxMaxHSDPA PtxTargetTotMax

PtxMaxHSDPA max. allowed HSDPA power WCEL; 0..50 dBm; 0.1 dB; 43 dBm

PtxCellMax HSDPA PS streaming PtxTargetTotMax HSDPA NRT

PtxTargetTot PtxTargetTotMin

PtxNCTHSDPA

HSDPA voice + SRBs DCH PS streaming DCH NRT DCH RT + SRBs

PtxNCTDCH

(excluding PS streaming)

Common Channels

PtxTarget NCT Tx power target for DCH

max. target pwr for NCT* load WCEL; -10..50; 0.1; 32767 dBm Special value: Use of dynamic DL target power is disabled

NCT Tx power target for DCH + HSPA PtxTargetTotMin min. target pwr for NCT* load WCEL; -10..50; 0.1; 32767 dBm Special value: Use of dynamic DL target power is disabled

PtxTargetTot is calculated always when NCT* load services are admitted

PtxNCTHSDPA: power used by HSDPA conversational services PtxNCTDCH: power used by DCH services associated as NCT load

* Non-Controllable Traffic NCT: CS services & PS conversational services

Dynamic target power for NCT load The min. & max. value for dynamic target power for NCT load (CS services & PS conversational services) can be set :

Rules:

PtxTargetTotMin WCEL; -10..50 dBm; 0.1 dBm; 32767 dBm

PtxTargetTotMin

PtxTargetTotMax WCEL; -10..50 dBm; 0.1 dBm; 32767 dBm

PtxTargetTotMax PtxCellMax

PtxTargetTotMin PtxTargetTot PtxTargetTotMax PtxTargetTotMax

PtxTarget PtxTargetTotMin

Dynamic target power is used when in cell there are SRBs or conversational services (NCT load) mapped to HS-DSCH transport channel. Dynamic target power varies between PtxTargetTotMin & PtxTargetTotMax depending on the mix of services mapped to DCH & HS-DSCH transport channels. However, NCT load caused by services mapped to DCH transport channels must still stay below PtxTarget. Power margin between PtxCellMax & PtxTargetTotMax is needed to protect the already admitted services mapped to HStransport channels by giving time for the overload control to adjust PS DCH load before high priority HS-DSCH load is affected. PtxTargetTot is calculated whenever a NCT connection is admitted

PtxTargetTot = PtxTargetTotMax - PtxNCTDCH

(

PtxTargetTotMax PtxTarget

-1

)

PtxNCTDCH: power used by DCH services associated as NCT load NCT: Non-Controllable Traffic

PtxTargetPS Target Calculation • The introduction of CS Voice over HSPA impacts the calculation of the target for PtxTargetPS • The original calculation in RAS06 was: PtxTargetPSTarget = P tx_nc + [(P max - P tx_nc ) x Weight Ratio]

• This calculation shares the power left over from non-controllable load between HSDPA & NRT DCH connections

• The calculation was updated in RU10 to account for HSDPA streaming: PtxTargetPSTarget = P tx_nc + [(P max - P tx_nc - P tx_hsdpa_stream) x Weight Ratio]

• The updated calculation reduces the quantity of power to be shared by effectively including HSDPA streaming power as non-controllable power

• The calculation is further updated when CS Voice over HSPA is enabled PtxTargetPSTarget = P tx_nc + [(P max - P tx_nc - P tx_hsdpa_stream- P nc_hsdpa) x Weight Ratio] CS Voice over HSPA transmit power

UL Power Allocation: dynamic threshold PrxTargetAMR •

PrxTargetMax

PrxTargetAMR is used for the admission of UL DCH

max. UL target power for CS speech service allocation WCEL; 0..30; 0.1; 465535 dB

& E-DCH, SRB & CS AMR connections

•

PrxTargetAMR shall be applied always, w/o

considering the activation of the feature CS voice over HSPA.

•

PrxTargetAMR varies between PrxTarget &

PrxTargetMax HSUPA NRT PrxTargetAMR

PrxTargetMax depending upon the UL load of data services

•

PrxTargetAMR is calculated by cell specific AC

inside RNC

•

NCT can always use power up to PrxTarget

•

Standalone SRB & CS AMR can be admitted even if the NC interference power exceeds PrxTarget as long as the RSSI is below PrxTargetAMR

•

SCT load of the HSUPA & UL DCH streaming services can take all power left from the NCT load up to PrxTarget

•

DCH PS NRT services can use power up to dynamic UL DCH target PrxTargetPS

•

HSUPA PS NRT services can take all power left from all other services

P r xTarget

HS/DCH CS AMR DCH CS data

PrxTargetPS

DCH PS NRT PrxDataDCHNST HSUPA PS streaming DCH PS streaming other interference, Noise power Semi-Controllable Load

Controllable Load

Non-Controllable Load NST: Non-Scheduled Transmission

HSUPA Non-Scheduled Transmission NST • NST is used for the UL of CS Voice over HSPA • HSUPA TTI = 2 ms  1 HARQ process is allocated for the E-DCH MAC-d flow • EDCHMuxVoiceTTI2 & EDCHMuxVoiceTTI10 define whether or not other E-DCH MAC-d flow data can be multiplexed within the same MAC-e PDU as CS Voice

• The max. Number of Bits per MAC-e PDU for NST indicates the number of bits allowed to be included in a MAC-e PDU per E-DCH MAC-d flow configured for non-scheduled transmissions

• Generally the MAC-d flow of the SRB has higher SPI value, being prioritized over the CS voice in the ETFC selection

•

 The

max. SRB bit rate will be limited so that the at least 1 CS voice frame can always transmitted together with the signaling when the max. puncturing is applied, for minimizing the CS voice delay

• 2 ms TTI is selected whenever possible, otherwise 10 ms TTI is used The maximum target value for the RTWP in UL for CS speech service allocation:

PrxTargetMax defines the max. target value for the RTWP in the UL resource allocation for the CS speech services. A dynamic target of RTWP is applied in the resource allocation for the CS speech services and for the establishment of the link. Dynamic target is the closer to the value of this parameter, the less there is PS NRT R99 data traffic and RT data R99 and HSPA traffic in the cell . Establishment of the stand alone signaling link or a single service CS speech can be admitted in UL even the received noncontrollable interference exceeds the value of the parameter "Target for received power" so long as the RTWP keeps below the dynamic target value defined with this parameter. WCEL: 0..30 dB; 0.1 dB; 465535 dB NST: Non-Scheduled Transmission

Fast Dormancy: Background Smart phones with many applications, requiring frequent transmission of small amount of data # (always-on) To save battery power, 3GPP defines transition from states with high power consumption (Cell_DCH, Cell_FACH) to those with low consumption (Cell_PCH, URA_PCH) approx. battery consumption in different RRC states:

UTRA RRC Connected Mode URA_PCH

CELL_PCH

CELL_DCH

CELL_FACH

•Idle = 1 (relative units) •Cell_PCH < 2*1 •URA_PCH ≤ Cell_PCH*2 •Cell_FACH = 40 x Idle •Cell_DCH = 100 x Idle

Idle Mode Typical terminal power consumption

300 ] 250 A m [ 200 n o i t p150 m u s100 n o c r 50 e w o P 0

*1

URA_PCH / Cell_PCH / Idle

Cell_FACH

Cell_DCH

depends on DRX ratio with Idle & mobility *2 < in mobility scenarios, = in static scenarios # e.g. sending frequent ‘polls’ or ‘keep-alives’

Fast Dormancy: Background Problem for UE: many networks with rel. long inactivity timers for Cell_DCH & Cell_FACH and/or PCH states not activated


CELL_PCH

CELL_DCH

CELL_FACH



UE vendors introduced proprietary Fast Dormancy: •UE completes data transfer •UE sends Signaling Connection Release Indication SCRI (simulating a failure in the signaling connection)

•RNC releases RRC connection

 UE

to RRC Idle mode

Idle Mode Disadvantages : •increasing signaling load due to frequent packet connection setup (PS RAB), •large number of “signaling connection failures” •increased latencies

Fast Dormancy: Principle 3GPP Rel. 8: Fast Dormancy •modifying SCRI message; new cause value indicating packet data session end


CELL_PCH

CELL_DCH

CELL_FACH

•RNC can keep UE in RRC connected mode, moving it into CELL_PCH/URA_PCH 

UE

battery life remains prolonged because power consumption in CELL_PCH/ URA_PCH is low Network

again in charge of RRC state; clarification of “signaling connection failures” Reduction

Idle Mode

of signaling load & latency times

3GPP TS 25.331 10.3.3.37a Signalling Connection Release Indication Cause „This IE is used to indicate to the UTRAN that there is no more PS data for a prolonged period.“

Cause value of ‘UE Requested PS Data Session End’ defined

SRCI: Signalling Connection Release Indication

Fast Dormancy FastDormancyEnabled RNFC; 0 (Disabled), 1 (Enabled)

SIB1: T323

MSActivitySupervision

BTS

SCRI: „UE Requested PS Data Session End”

MSActivitySupervision

„Physical Channel Reconfig.”  move to CELL_PCH

RNC; 0..1440; 1; 29 min

UE

RNC

RAN2136: Fast Dormancy (FD) • • •

Basic SW; no activation required; enabled by default MSActivitySupervision to be configured with value > 0 to enable PCH states Enabling FD results in T323 being broadcast within SIB1

T323: • •

Inclusion of T323 within SIB1 allows UE to detect that network supports FD Setting a min. delay between 2 SRCI messages for FD; prevents, that UE is sending a flow of SCRI messages, if network is temporarily unable to move UE to a battery-saving state T323 RNC; 0..7; 1; 0 s (hardcoded)

Fast Dormancy - RNC Actions: After receiving SCRI message with cause value ‘UE Requested PS Data Session End’: •FD functionality overrides inactivity timers •RNC instructs UE to make state change to CELL_PCH/URA_PCH If RNC receives an SCRI message without a cause value then the existing legacy functionality is applied & the UE is moved to RRC Idle mode

Fast Dormancy Profiling: RAN2451 • Included in RU40 application software package – license is required

Brief description: • Identifies legacy Fast Dormancy phones which cause unnecessary signaling load • Provides with better network resources utilization due to shorter inactivity timers • Less signaling load because LFD (Legacy Fast Dormancy) Phones are being forced to stay in Cell_PCH

Benefits: • Signaling load reduction on Iub, UU and Iu interfaces • Signaling load reduction in the RNC • Longer UE battery life Overview: • RAN supports Fast dormancy • Application has no more data to transfer • UE wants go to more battery efficient RRC state SCRI

RNC: Data session ended RNC: UE move to more battery efficient state

SIB1 contains info about T323

Go to URA/Cell_PCH

Fast Dormancy Profiling: Background Legacy Fast Dormancy phone detection: • The UE is detected as Legacy Fast Dormancy phone (LFDphone) when network receives RRC:Signaling Connection Release Indication without any cause

• If the Fast Dormancy Profiling feature is ac tivated then RRC state transition is performed according to Fast Dormancy functionality SCRI - without any cause

RNC checks if the license is ON

If the license is available - Go to Cell_PCH

Handling the PS Connection Establishment: • The LFD Phone after sending SCRI without any cause may st ill silently goes to Idle • After receiving RRC: Initial Direct Transfer, RNC checks if Iu-PS connection already exists • If yes, then all existing PS RAB resources locally and the old Iu connection are released • New Iu connection is established for pushing RRC: Initial Direct Transfer to SGSN RNC checks if Iu-PS connection for this UE already exists

Iu RRC: Initial Direct Transfer

Fast Dormancy Profiling: Principle Shorter Inactivity Timers for LFD Phone and Smartphones: • Shorter inactivity timers should be used for moving smartphones and LFD Phones to Cell_PCH state - saving UE battery

• It gives possibility to avoid unnecessary movement to IDLE_mode – less signaling load

Higher Traffic Volume Thresholds for LFD Phone and Smartphones: • Higher traffic volume thresholds should be used for moving smartphones and LFD Phones to Cell_DCH state

• It gives possibility to avoid unnecessary movement to Cell_DCH – only for sending keep-alive message

• Stored IMSI gives possibility to faster usage of higher traffic volume thresholds

High Speed Cell_FACH (DL): RAN1637 • Included in RU30 application software package – license required • HW prerequisites: Flexi rel.2 • Can be used if: Flexible RLC Downlink is active

Brief Description:

HSFACHVolThrDL WCEL; Infinity, (8, 16, 32, 64, 128, 256, 512, 1024, 2048, 3072, 4096, 8192, 16384, 24576, 49152) bytes

• This feature enables Fast Cell_PCH to Cell_FACH switching (transition <200ms) • Feature reduces signaling load on Iub and Iur interf aces • Reduces code tree occupation • Saves BTS baseband resources • Increases number of supported smartphones • Increases possible throughputs on common channels to 1. 80Mbps in DL

DL channel mapping: Logical channels

Transport channels

BCCH

BCH

CCCH DCCH

FACH FACH

FACH

DTCH

HS-DSCH

PCCH

FACH

PCH

S-CCPCH

S-CCPCH

3GPP Rel7

Physical channels

P-CCPCH

S-CCPCH

HS-PDSCH

High Speed Cell_FACH (DL): With and Without the Feature RAN1637

RAN1637

Data transmission

Not activated

Activated

HSDPA only on dedicated channels

HSDPA also on common channels

Common channels Common channels Dedicated channels

Dedicated channels

Significant setup time reduction •

to Cell_DCH state change Cell_PCH

•

Cell_PCH to Cell_FACH state change

•

Cell Update required • 600 ms

•

Cell Update not needed

•

Data appears in buffer

Cell update

Channel type switch

<200 ms Data appears in buffer

Transmission/recepti t [ms] on in Cell_DCH

Transmission/recepti on in Cell_FACH Channel type switch

t [ms]

HSDPA RRM • • • • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH HSDPA Improvements Other Features Appendix: – Static HS-PDSCH Power Allocation – Cell Reselection – Iub Flow – Congestion Control

Static HS-PDSCH Power Allocation (1/2) • Required parameter settings – –

PtxMaxHSDPA Maximum allowed HSDPA power WCEL; 0..50 dBm; 0.1 dB; 43 dBm

Dynamic HS-PDSCH power allocation disabled “Fixed” HS-PDSCH power defined with PtxMaxHSDPA

• Rules for HSDPAPriority = 1 (higher priority for HSDPA) – A: 1st HSDPA users enters cell  

Non-controllable traffic PtxNC ≤ PtxTargetHSDPA  HSDPA allowed Otherwise R99 only

– HSDPA already active 

R99 scheduled up to PtxTargetHSDPA

– B: Overload for total R99 traffic PtxnonHSDPA > modified overload threshold 

Standard R99 overload actions

– C: Overload for PtxNC > modified overload threshold  HSDPA released

Node B Tx power

Max power

HSDPAPriority 1,2; 1 = HSDPA priority Ptxtotal

PtxTargetHSDPA

PtxTarget

Target for transmitted non-HSDPA power -10..50 dBm; 0.1 dB; 38.5 dBm

PtxOffsetHSDPA

PtxOffsetHSDPA

PtxnonHSDPA

Offset for transmitted non-HSDPA power 0..6 dB; 0.1 dB; 0.8 dB

PtxTargetHSDPA

PtxNC

A

B

C

Static HS-PDSCH Power Allocation (2/2) • Rules for HSDPAPriority = 2 (higher priority for R99) – A: 1st HSDPA users enters cell  Total R99 traffic PtxnonHSDPA ≤ PtxTargetHSDPA  Can have HSDPA  Otherwise can have R99 only

– HSDPA already active  R99 scheduled up to PtxTargetHSDPA

– B: Overload for total R99 traffic PtxnonHSDPA > modified overload threshold  HSDPA released

– C: Standard overload for total R99 traffic PtxnonHSDPA > standard overload threshold  Standard R99 overload actions

Node-B Tx power HSDPAPriority

Max power

1,2; 1 = HSDPA priority Ptxtotal

PtxOffset

PtxTargetHSDPA Target for transmitted non-HSDPA power -10..50 dBm; 0.1 dB; 38.5 dBm

PtxTarget

PtxOffsetHSDPA

PtxTargetHSDPA

Offset for transmitted non-HSDPA power 0..6 dB; 0.1 dB; 0.8 dB

PtxOffsetHSDPA PtxnonHSDPA PtxNC

A

B

C

HSDPA RRM • • • • • • • • • • •

HSDPA Principles HSDPA Protocols & Physical Channels RU50 Capabilities & Baseband Configuration HSDPA Link Adaptation HSDPA H-ARQ HSDPA Packet Scheduling Basics of HSDPA Power Allocation Basics of HSDPA Code Allocation Basics of HSDPA Mobility HSDPA Channel Type Selection & Switching Associated UL DCH

• HSDPA Improvements • Other Features • Appendix: – – – –

Static HS-PDSCH Power Allocation Cell Reselection Iub Flow Congestion Control

Cell Re-selection (1/3) HSDPAMobility Serving HS-DSCH cell change & SHO on/off switch RNFC; 0 = HSDPA cell reselection; 1 = Serving HS-DSCH cell change

• HSDPAM o b i l i t y set to disabled • IF- mobility handled by HSDPA Cell Reselection, not by serving cell change • IF- / IS- mobility handled by same events as for serving cell change

• HSDPA cell reselection • Transition to CELL_FACH based on event 1a • Handling depends on setting of EnableRRCRelease

EnableRRCRelease Enable RRC connection release HOPS; 0 = disabled; 1 = enabled

 if disabled  1a triggers transition to Cell_FACH immediately  if enabled  1a triggers IF- measurements only; transition to cell_FACH triggered by release margins

• HSDPARRCdiversity • can disable SHO for stand alone SRB of HSDPA capable UE (e.g. according addition window) • reduces capacity consumption due to stand alone RRC connections (more capacity available for HSDPA) • if conditions for HSDPA mobility fulfilled, SHO for stand alone SRB is allowed in any case (e.g. triggered by release margins)

HSDPARRCdiversity IF: Interfrequency IS: Intersystem

SHO of the HSDPA capable UE RNHSPA; 0 = disabled; 1 = enabled

EnableRRCRelease = disabled

Cell Re-selection (2/3)

Risk of ping-pong But UE connected mostly to optimum cell

Ec /Io Addition Window

CPICH 2

Addition Time

CPICH 1

time

AdditionWindow FMCS; 0..14.5 dB; 0.5 dB; 4 dB Recommended 0 dB

HSDPA

CELL_FACH Measurement Reports

HSDPA

Cell Re-selection (3/3) EnableRRCRelease = enabled No ping-pong But UE often connected to non optimum cell

Ec /Io Addition Window

ReleaseMarginAverageEcNo ReleaseMarginPeakEcNo

One margin need to be exceeded only CPICH 2 Addition Time

CPICH 1

Measurement Reports

HSDPA ReleaseMarginAverageEcNo Release margin for average Ec/Io HOPS; -6..6; 0.1; 2.5 dB

CELL_FACH

time

HSDPA

ReleaseMarginPeakEcNo Release margin for peak Ec/Io HOPS; -6..6; 0.5 dB; 3.5 dB

HSDPA RRM • • • • • • • • • • •




Iub Flow Control (1/4) •

•

Objective

–

Node B has to offer sufficient data for HSDPA

–

to avoid overflow of its buffer

–

to be performed per HSDPA connection on Iub

Node B informs RNC about

–

•

Max. number of MAC-d PDUs (credits) allowed to be sent by RNC for unlimited 10ms periods. That means that the RNC can send data according to latest capacity allocation as long as new capacity allocation is received

Number of assigned credits are recalculated by BTS each 10ms and signaled to the RNC (if differs enough from the previously signaled). Calculated capacity allocation depends on

– Air interface throughput estimation (the higher, the more credits) – Buffer occupancy (the higher, the less credits)

•

BTS prevents packet loss due to buffer overflow by reducing the capacity allocation in case of air interface congestion and ensures that the HSDPA capacity can be reached by having enough data to fill the reserved power allocation

Iub Flow Control (2/4)

RNC

Node B

CAPACITY REQUEST

Priority User buffer size in RNC Priority

CAPACITY ALLOCATION

User buffer size in Node B Credits (number of MAC-d PDUs) Repetition period (number of time intervals) Credit validity interval (duration of time interval) Priority

DATA

User buffer size in RNC Length of MAC-d PDU MAC-d PDUs

Example: Credits = 4 Repetition period = 3 Credit validity interval = 10 ms


Number of credits allocated per user decreases and the HSDPA connection throughput decreases as the number of connections increases

•

Number of PDU transferred drops frequently when 1 HSDPA connection is act ive only

Raw data

Averaged data

60

60 1 active UE

MAC-d PDU sent to Node B

s 50 t i b 6 3 3 ( 40

MAC-d PDU sent to Node B 50

Credits allocated by Node B

U D P d C 30 A M f o r 20 e b m u N

D P 40 d C A M30 f o r e b m20 u N

2 active UE

3 active UE

Credits allocated by Node B

4 active UE

10

10

0

0 0

20

40

60

80

100

120

140

160

Time (seconds)

180

200

220

240

260

280

0

20

40

60

80

100

120

140

160

Time (seconds)

180

200

220

240

260

280


Node B buffer occupancy can be evaluated as follows number of acknowledged MAC-d PDU - number of MAC-d PDU transfer red from the RNC

•

Comparison with previous slide shows, that number of credits decreases also because of high buffer occupancy

Raw data

600

Averaged data 600

2 active UE D P 500 d C A M ( 400 y c n a p u 300 c c O r e f f 200 u B B e d o 100 N

Connection 1 Connection 1

U D 500 P d C A M400 ( y c n a p u 300 c c O r e f f 200 u B B e d 100 o N

Connection 2 Connection 3 Connection 4 3 active UE 4 active UE

1 active UE

0

Connection 2 Connection 3 Connection 4

0 0

25

50

75

100

125

150 Time (s)

175

200

225

250

275

300

0

25

50

75

100

125

150 Time (s)

175

200

225

250

275

300

HSDPA RRM • • • • • • • • • • •




Iub Congestion Control CC (1/2) •

Objective: – –

•

RNC informs Node B by DL Frame Protocol about: – –

•

Too strong delay of frames Loss of frames 3 thresholds (BTS commissioning parameter) Minimum threshold Thmin: 0..5000 ms; 50 ms Intermediate threshold Thmid: 0..5000 ms; 250 150 ms Maximum threshold Thmax: 0..5000 ms; 1000 250 ms

Actions: – – – –

•

WBTS; 0 (disabled); 1 (enabled)

Delay thresholds: – – – –

•

Build up delay Sequence number

HSDPACCEnabled

Node B thus can detect: – –

•

RNC can not see Iub congestion towards Node B after hub node Iub congestion must be detected by Node B

Delay < Thmin no action Thmin ≤ delay ≤ Thmid Node B reduces credits for RNC with low probability (depending linearly on delay with low slope) Thmid ≤ delay ≤ Thmax Node B reduces credits for RNC with high probability (depending linearly on delay with high slope) Delay > Thmax or frame loss Node B reduces credits for RNC in any case

If QoS aware scheduling applied: –

for high priority service Node B reduces credits for RNC with lower probability than for low priority service

3g Ran Nsn Hsdpa Rrm & Parameters

Recommend Documents