umts system

7/29/2019 UMTS System

http://slidepdf.com/reader/full/umts-system 1/359

UMTS Architecture Overview

by Dr Paul Raby



Public Land Mobile Network (PLMN)

▪ A PLMN can be regarded as anindependenttelecommunications entity.

▪ A PLMN is defined as:

▫ One or more switches with:

▪ a common numbering plan

▪a common routing plan▫ Switches act as the interface to

external networks

▪ The PLMN can be separatedinto

▫ Core Network

▫ Access Network

Core Network

Access Network

PLMN

System Architecture Overview



User Equipment

UMTSTerrestrial

Radio AccessNetwork

Core Network

UU IU UE UTRAN CN

UMTS High Level Architecture




Major Network Elements in UMTS

PLMN,PSTN,ISDN

Internet,X25

PacketNetwork

UU

UE

CU

USIM

ME

MobileEquipment

UMTSSIM

CN

MSC/VLR

SGSN GGSN

GMSC

HLR

Serving GSN Gateway

GSN

GatewayMSC

Mobile SwitchingCentre

Home LocationRegister

IU

UTRAN

IUb

IUr

Node B

Node B

Node B

Node B

RNC

RNC

Radio NetworkController


Iu-ps

Iu-cs

IUb




General UTRAN Architecture

UU IU

UE

UTRAN

IUb

IUr

Node B

Node B

Node B

Node B

RNC

RNC



Iu-ps

Iu-cs

IUb

CN (MSC)

CN (SGSN)




Elements of UTRAN

▪ Radio Network Controller

▫ Owns and controls radio resources in its domain (BSC in GSM)

▫ Service Access Point for all services that UTRAN provides for the CN

▫ Note: Service RNC (SRNC) and Drift RNC (DRNC) are subsets

▫ Note: Control RNC (CRNC ) whichever RNC is talking to the UE

▪ Node B

▫ Acts as the radio base station (BTS in GSM)

▫ Converts the data flow between the Iub and Uu interfaces




Major Interfaces in UMTS

▪ There are four major newinterfaces defined in UMTS

▫ Iu

▪The interface between UTRANand the CN

▫ Iur

▪The Interface between differentRNCs

▫ Iub

▪The interface between theNode B and the RNC

▫ Uu

▪The air interface

RNC

Node- B

RNC

UE

CN

Uu

Iu

Iub

Iur




UMTS Interface Implementation

ATM/IP Network

RNC Node B

Node B

Node B

MSC

RNC

SGSN

Node B

Iub Iu_cs

Iu_ps Iur




Handover in UMTS

▪ There are 3 basic types of handover▫ Intra frequency handovers

▪ Handovers between 2 UMTS carriers at thesame frequency

▪ These can be soft handovers

▫ Inter frequency handovers ▪ Handovers between 2 UMTS carriers at different

frequencies

▪ These are hard handovers

▫ Inter system handovers

▪ Handovers between UMTS and GSM carriers

▪ These are hard handovers




Macrodiversity between Node B’s

▪ If Active Set consists of

▫ two connections to cellsparented to different Node Bs

▫ then the combining of the twochannels occurs at the RNC

▪ This is known as a soft handover

▪ This doubles thetransmission „cost‟ of thecall

RNC

Node B

Cell

Cell

Cell

Node B

Cell

Cell

Cell

Iur

Iu

Uu




Maximal Combining between Cells on theSame Node B

RNC

Node B

Cell

Cell

Cell

Node B

Cell

Cell

Cell

Iur

Iu

Uu

▪ If Active Set consists of

▫ two connections to cellsparented to the same Node B

▫ then the combining of the twochannels occurs at the NodeB

▪ This is known as a softer handover

▪ This has no transmissionimplication (but does havecapacity implications) ifcells are collocated.

▪ Uses maximal combining

▫ Adds electrically the

signals making one betterthan each individual




Architecture of a UMTS bearer service

TE TE UE UTRAN CN

edge node

CNgateway

End-to-End

TE/UE LocalBearer UMTS Bearer External Bearer

Radio Access Bearer CN Bearer

Radio Bearer Iu Bearer Backbone Network

UTRA FDD/TDD Physical Bearer

Each bearer service on a specific layer provides services using layers below.


S A hi O i



UMTS Protocol Stratums

▪ Non-access StratumEncompasses layers 4 to

7 of the OSI 7 layermodel, and the upper partof layer 3

N

onA c c e s s S t r a t um

A c c e s s S t r a t um

L1 L1 L1L1

L2L2L2L2

L5L5

L4L4

L6 L6

L7 L7

L3 lower L3 lower L3 lower L3 lower

L3 upper L3 upper

Uu Iu UE UTRAN CN


• Access Stratum

→ Encompasses layers 1and 2 of the OSI 7 layermodel and the lower part oflayer 3

S t A hit t O i



UMTS QoS Classes

▪ Conversational ▫ Speech over CS bearer

▫ Voice over IP, PS bearer

▫ Delay critical, imposed by human perception

▪ Streaming ▫ Multimedia streaming

▫ Using buffers, for non-real time delivery; real-video, real- audio

▪ Interactive ▫ Web browsing, database retrieval

▫ Round trip delay time is a key parameter

▪ Background ▫ E-mail

▫ Delay:- 10s of seconds or even minutes


S t A hit t O i



Protocol Model for UTRAN Interfaces

▪ Protocol structures in UTRAN are designed in layers and planes.

▪ They are seen as logically independent of each other

▫ However they will physically interact.

▪ Being logically independent allows for changes to blocks in the

future

theoretically!





General Protocol Model forUTRAN Terrestrial Interfaces





Horizontal Layers in the General Protocol Model

▪ All UTRAN related issues are only visible in the Radio

Network Layer

▪ The Transport Layer simply represents standard transport

technology for use in UTRAN

▫ e.g. ATM and appropriate ATM Adaptation Layers

▪ AAL2 ( voice ) and AAL5 ( data/control)

▫ UDP/IP or RTP/UDP/IP ( release 6 ? )





Vertical Planes in the General Protocol Model

▪ The Control Plane is provided for all UMTS

specific control signalling including:

▫ Application Protocols

▫ Signalling Bearers

▪ The User Plane is provided for all data sent and

received by the user including:

▫ Data Streams

▫ Data Bearers






▪ The Transport Network Control Plane also includes the Access

Link Control Application Part, ALCAP.

Transport Network User

Plane


Plane

Transport Network Control

Plane

ALCAP




UMTS Technology Overview




CDMA - Direct Sequence Spread Spectrum


f r e q u en c y

time

code

Frame Period (we may still needframes/timeslots for signaling)




CDMA Spreading

•Essentially Spreading involves changing the symbol rate on the air interface

Identical codes

Tx Bit Stream

P

f

Code Chip Stream

Spreading

P

f

Channel

Air InterfaceChip Stream

P

f

Code Chip Stream

Despreading

P

f

Rx Bit Stream

P

f





Spreading and Despreading

Rx Bit Stream


Tx Bit Stream 1

-1

Code Chip Stream

XSpreading

Code Chip Stream

X Despreading





Spreading and Despreading with code Y


Tx Bit Stream 1

-1

Code Chip Stream

XSpreading

X Despreading


Code Chip Stream Y

Rx Bit Stream




Spreading

▪ If the Bit Rate is Rb, the Chip Rate is Rc, the energy per bitEb and the energy per chip Ec then

▪ We say the Processing Gain Gp is equal to:

▪ Commonly the processing gain is referred to as theSpreading Factor

b

ccb

R

R E E

b

c p

R

RG





Spreading in noise

▪ The gain due to Despreading of the signal over widebandnoise is the Processing Gain

Signal

P

f

Spreading Code

Tx SignalP

f

Rx Signal (= Tx Signal + Noise)

f

P

Channel

Wideband Noise/Interference

P

f

Spreading Code Signal

P

f

gy




Visualising the Processing Gain

Signal

Intra-cell Noise

Inter-cell Noise

W/Hz

Before Spreading

f

W/Hz

After Spreading

f

W/Hz

Ec

Io With Noise

f

W/Hz

After Despreading /Correlation

f

W/HzEb

No

Post Filtering Orthog = 0

f

dBW/Hz

Eb

No

Eb /No

f

Eb

No

W/HzPost Filtering Orthog > 0

f

Eb

No

Eb /No dBW/Hz

f

gy




Separating BaseStations

▪ Summarising:

▫ Channelisation Codes

▪Are used to separate channelsfrom a single cell

▫ Scrambling Codes

▪Are used to separate cells fromeach other rather than purelychannels

▪ Different base stations will use thesame spreading codes withseparation being provided by theuse of different scrambling codes.

S1

S2

S3

C1 C2 C3

C1 C2 C3

C1 C2 C3

gy




Separating UE’s

▪ Summarising:

▫ Spreading/Channelisation Codes

▪Are time dependent and so are used in the UL to spread the signalbut not to separate the UE‟s

▫ Scrambling Codes

▪Are used to separate UEs from each other rather.

gy




Spreading Codes = Channelisation Codes

▪ Channelisation codes are orthogonal

▫ Which provides channel separation

▪ Number of codes available is dependant onlength of code

▪ Channelisation codes are used to spread thesignal




Channelisation Code Generation

▪ Channelisation codes can be generated from a Hadamard matrix

▪ A Hadamard matrix is:

▪ Where x is a Hadamard matrix of the previous level

▪ For example 4 chip codes are:

▫ 1,1,1,1

▫ 1,-1,1,-1

▫ 1,1,-1,-1

▫ 1,-1,-1,1

x x

x x

Note: These two codes correlateif they are time shifted




▪ Orthogonal Variable Spreading Factor Codes can

be defined by a code tree:

▫ SF = Spreading Factor of code (maximum 512 forUMTS in the DL, 256 in the UL)

SF = 1 SF = 2 SF = 4

Cch,1,0 = (1)

Cch,2,0 = (1,1)

Cch,2,1 = (1,-1)

Cch,4,0 =(1,1,1,1)

Cch,4,1 = (1,1,-1,-1)

Cch,4,2 = (1,-1,1,-1)

Cch,4,3 = (1,-1,-1,1)

OVSF codes




Vital Parameters

▪ Ec/Io of the Pilot Channel is used to

▫ estimate (“sound”) the channel (multipath characteristics)

▫ decide which server is “best server”

▫ make handover decisions

▫ Typical requirement > -15 dB

▪ Eb/No in both uplink and downlink affects error ratios.

▫ Typical requirement 1 to 10 dB

▫ Required value of Eb/No depends on propagationconditions and sophistication of receiver.

▫ This is your Quality Measure

▪ Noise rise limits path loss and coverage.




Rake Receiver

Correlator

Code Generators

(S & C)

ChannelEstimator

Phase Rotator Delay Equalizer

Matched Filter

I

Q

I

Q

A typical rake receiver withthree fingers




RAKE Receiver

▪ Auto-correlation function of PN sequence is used toproduce multipath estimate of propagation path.

▪ Each finger then acts as separate receiver to provide theoptimum signal

Autocorrelation

-0.5

0

0.5

1

0 1 2 3 4 5 6 7

Delay

A v e r a g e v a l u e

Direct Component Delayed Components




Wideband Implications

▪ Autocorrelation function of PN sequence is 1 for zero delay and zerofor all delays outside a chip period.




Multipath Situation

Autocorrelation can be processedto provide a channel estimation.

Tx

Rx

+




Resolution of Multipath

Chip Period = 0.26 microsecondsCorresponding path length difference = 78 metres

This indicates the sort of resolution possible with theUMTS Rake receiver

Tx

Rx

excel




Noise Rise

▪ The effective noise floor of the receiver increases as the

number of active mobile terminals increases.

▪ This rise in the noise level appears in the link budget andlimits maximum path loss and coverage range.

Background Noise

Three Users

One User

Two Users




Self Assessment Questions

▪ What is the Processing Gain (in dB) if a UMTS

system utilising a chip rate of 3840 kbps is used

with the following user rates?

i) 12.2 kbps

ii) 64 kbps

iii) 128 kbps





Solution:

dB98.2412200

3840000log10

3840000log10Gain

10

10

b R




Self Assessment Question

▪ If a 12200 bps voice channel has a

SNR of -16 dB. Determine Eb/No.

Processing Gain = 25 dB

Eb/No = 25 - 16 = 9 dB





▪ A mobile receives a 12200 bps voice channel from a base

station at a level of –106 dBm. The base station is transmitting

12 additional identical voice channels plus a pilot channel and

common channel that are both received with power levels of –

104 dBm. The background thermal noise level of the mobile is –99 dBm.

Determine the value of Eb/No at the output of the receiver for

the following orthogonality factors.

i) 0 ii) 0.5 iii) 1.0




Example Solution

W

W

W

13-

14-

13-

103.81Total

107.96dBm-1042

103.01dBm-10612

common)andpilot(incchannelsotherfromPowerTotal

Taking Orthogonality to be 0.5:

W 13-101.9PowerEffectiveModified




Solution

W 13-101.9PowerEffectiveModified

W 13-101.26dBm99-PowerNoiseThermal

dBm95

10.163PwrEff.ModifiedNoiseThermal 13-

W

dB11(-95)--106SNR

dB14N

E

dB52GainProcessing

0

b

ebno



The Link Budget

The Link Budget



The Link Budget

▪ GSM and UMTS compared

▫ More thermal noise in UMTS systems. KToB ~ -108.1 dBm

▫ Processing Gain (dBs) in UMTS

= 10 log (3840000/User Rate (bps))

▫ Power Control Margin for imperfect Fast Fading Power

Control must be considered in UMTS systems

▫ Interference Margin for Noise Rise must be considered in

UMTS systems

The Link Budget



Recovering the Wanted Signal

jtotal

j

j

b

jtotal

j

P I

P

R

W

N

E

P I

PSNR

0

The Link Budget




total j

j jb

total

j

j

b

j

j

btotal

j

j

j

b

jtotal

j

I L

R

W

E

N

I

W

R

N

E

W

R

N

E I

P

W R

N E

P I P

0

0

0

0

1

1

The Link Budget




j jb

j

R

W

E

N L

01

1

95

1

1220038400003.01

1

12200 3840000 )dB5.2( 3.00

j

j

b

L

RW E

N

Inserting typical values for a voice service:

The Link Budget




Possibility 1:

•Single user.

•Thermal Noise responsible for

94/95ths of received power.

•SNR=1/94 = -19.8 dB

User j is responsible for 1/95th of received power

Possibility 2:

•Power level so high that thermalnoise is insignificant.

•95 identical users possible.

Practical situations will lie in between these two extremes.

The Link Budget



Considering Thermal Noise

Total Power = Power from Users + Thermal Noise

ReceivedPowerTotaluserswantedfromreceivedPower

1

1

1

1

budget)link (affectsRiseNoise

1

1

UL

UL M j

j j

N

total

N

total

N total

M j

j jtotal

LP

I P

I

P I L I

•This is also the equivalent to loading factor – see later

C

The Link Budget



Considering Thermal Noise

Total Power = Power from Own Cell + Power from Other Cells + Thermal

Noise

Thermal Noise

Own Cell

Other Cell 323

95

New Capacity

Th N i Ri E i

The Link Budget



The Noise Rise Equation

jb

jUL

M j

j j

N

total

R

W

E

N L L

P

I

0

1

1

1

1

1

1

1

If we have M identical users:

jb

M j

j j

R

W

E

N M L

01 1

jb

N

total

R

W

E

N

M P I

01

1

1 RiseNoise

Th N i Ri E i

The Link Budget



The Noise Rise Equation

jb

N

total

R

W

E

N

M P

I

01

1

1

RiseNoise

W E

N R

MR

b j

j

0

1

1RiseNoise

W E

N

b

0

ghputuser throusingle

tthroughputotal1

1

Eff t f N i hb i C ll

The Link Budget



Effect of Neighbouring Cells

Users in other cells cause interference.

Typical ratio of power from other cells to power

from own cell, i, is 0.6

C id i th ll

The Link Budget



Considering other cells

Total Power = Power from Users + Noise from other cells + Thermal Noise

ReceivedPowerTotal

userswantedfromreceivedPower

11

1

11

1

budget)link (affectsRiseNoise

1

11

UL

UL

M j

j

j N

total

N

total

N total

M j

j

jtotal

M j

j

jtotal

i Li

P

I

P

I

P I Li I L I

Th M difi d N i Ri E ti

The Link Budget



The Modified Noise Rise Equation

jb

j

UL

M j

j

j N

total

R

W

E

N Li Li

P

I

0

1

1

1 11

1

11

1

If we have M identical users:

jb

M j

j j

R

W

E

N M L

01 1

jb

N

total

R

W

E

N

i M P

I

01

11

1

RiseNoise

Th M difi d N i Ri E ti

The Link Budget



The Modified Noise Rise Equation

W E

N

i

b

0ghputuser throusingle

1tthroughputotal1

1RiseNoise

i

W E

N

b

1


tthroughpuTotal

as RiseNoise

0

C ll C it

The Link Budget



Cell Capacity

i

W E

N

b

1

ghputuser throusingle 0

is known as “pole capacity”.

i N

b E

W

10

CapacityPole

usersof numberlargeFor

Cell Capacity

The Link Budget



Cell Capacity

i N

E

W

b

1 CapacityPole

usersof numberlargeFor

0

kbps853

5.0133840000 CapacityPole

0.5 (4.77dB) 3Eb/No 3840000W

i

• 50% of this would give a Noise Rise of 3 dB• ( Eb/No goes from 3 to 6 )

•50% of 853 kbps = 426 kbps

Loading Factor

The Link Budget



Loading Factor

R

W

i M N

E

i N

E

W

R M

R M

b

o

b

1

1

FactorLoading

:ratedatawithusersidenticalFor

CapacityPole

ThroughputActual

FactorLoading

0

Noise Rise and Loading Factor

The Link Budget



Noise Rise and Loading Factor

▪ Loading (Capacity) is linked to Eb/No value

▪ Noise Rise is linked to maximum path loss

Noise Rise Loading Factor

1 dB 20%

3 dB 50%

6 dB 75%10 dB 90%

UL 1log10RiseNoise 10

Activity Factor

The Link Budget



Activity Factor

factor.activityusertheis where

1

FactorLoading

R

W

i M N E

o

b

Users are not active 100% of the time.

It is necessary to adjust the loading factor to acknowledge this fact

Downlink Considerations

The Link Budget




The Downlink benefits from orthogonality between channelisation codes.

)1(


CapacityPole

0

i

W E

N

b

is orthogonality factor and has a value between zero and 1.


The Link Budget




The Downlink loading factor.

W

i N

E

o

b

1throughput

DL

maxP

Ptotal DL

Varies between approximately 20% and 75%

Uplink Budget for 144 kbps service

The Link Budget




Thermal Noise: -108 dBm, Noise Figure: 4 dB, Eb/No: 1.5 dB

Processing Gain: 14 dB (10 log[3840/144])

Sensitivity -116.5 dBm

Margins: Noise Rise: 3 dB, Fast Fading: 2 dB

Antenna Gains: 20 dBi

Tx Power: 21 dBm

•Allowable Path Loss: 152.5 dB

Downlink Budget for 144 kbps service

The Link Budget




Allowable Path Loss: 152.5 dB

Sensitivity -113.5 dBm



Required Tx Power: 24 dBm per channel

kbps19115.1

41.1

3840144 CapacityPole

For 3 dB Noise Rise, capacity is halved to 955 kbps or 6

channels.

Total transmitted power required = 6 x “24 dBm” = 31.8 dBm


The Link Budget




Thermal Noise: -108 dBm, Noise Figure: 4 dB, Eb/No: 5 dB

Processing Gain: 27 dB (10 log[3840/8])

Sensitivity -126 dBm



Tx Power: 21 dBm

Allowable Path Loss: 162 dB


The Link Budget




Allowable Path Loss: 162 dB

Sensitivity -123 dBm



Required Tx Power: 24 dBm per channel

For 3 dB Noise Rise, capacity is halved to 407 kbps or 51 channels.

Total required power is 51 x “24 dBm” = 41 dBm

kbps8155.1

3840316.08 CapacityPole

Coverage vs Capacity Comparisons

The Link Budget



Coverage vs. Capacity Comparisons

Coverage vs. Capacity

145.00

150.00

155.00

160.00

165.00

170.00

175.00

180.00

100 200 300 400 500 600 700 800

Throughput (kbps)

M a x i m

u m P

a t h l o s s ( d B )

Uplink

Downlink

144 kbps service 8 kbps service


145.00

150.00

155.00

160.00

165.00

170.00

175.00

180.00

100 200 300 400 500 600 700 800

Throughput (kbps)

M a x i m

u m P

a t h l o s s ( d B )

Uplink

Downlink

Capacity Issues

The Link Budget



Capacity Issues

▪ Cell Throughput affects Noise Rise

▪ Noise Rise affects Link budget.

▪ Adding more cells reduces pathloss andallows for more Noise Rise and hence highercapacity.

Capacity Issues: example

The Link Budget




▪ Network originally provides coverage at NR = 2dB

Radius of each cell is RR

If number of cells isdoubled the radius

reduces to R/ 2

Path loss will reduce by 5.3 dB.

( Assuming simple model for Path loss 137 +

35logR)

NR can increase by 5.3 dB.


The Link Budget




▪ NR of 2 dB corresponds to a loading factor of 37%.

▪ NR of 7.3 dB corresponds to loading factor of 81%

▪ Each cell can now handle more than double the traffic.

▪ Doubling the number of cells has increased the capacity

by a factor of 4.4 ( 0.81/0.37 x 2 )

Note that doubling the number again would give a NR of 12.6 dB

(loading factor 95%) which would not produce such a remarkable

improvement.

( 0.95/0.81 x 2 = 2.34 this is equivalent to 10.3 x capacity of original network )

Target Eb/No Values

The Link Budget



Target Eb/No Values

▪ Capacity is linked to Eb/No value

▫ Typical values: Voice 4 dB; High Speed Data 1.5 dB

▪ Lower overhead in control data for higher speed data.

▪ Also, Eb/No value assumes processing gain of 3840/12.2 for 12200bps voice. Actual transmitted data is 30000 bps.

▪ Target Eb/No controlled by RNC and will depend upon prevailing

conditions.

Single Channel Cells

The Link Budget



Single Channel Cells

▪ UMTS Flexibility allows for one user to take up entire cell‟scapacity.

▪ Noise Rise is now irrelevant as there can be no other users.

▪ Base station has higher capacity making coverage uplink

limited.

▪ Asymmetry in power can be reflected in asymmetry in

throughput. e.g. 2 Mbps possible in downlink with 144 kbpsin uplink.

Asymmetric Traffic

The Link Budget



Asymmetric Traffic

▪ Combination of (symmetric) voice and (highly asymmetric) packet data traffic

will tend to make the service asymmetric.

▪ Predicted asymmetry is 3:1 in favour of downlink.

▪ Achievable through balancing the links so that for the same path loss, downlink

has greater throughput capability.


145.00

150.00

155.00

160.00

165.00

170.00

100 200 300 400 500 600 700 800

Throughput (kbps)

M a x i m u m P a t h l o s s ( d B )

Uplink

Dow nlink

ConclusionsThe Link Budget



Conclusions

▪ Eb/No and Capacity intimately linked.

▪ Link budgets are affected by fast fading and interference

margins.

▪ Uplink and downlink affected differently by increased loading.

▪ Flexibility allows high data rate services to be provided.

▪ Asymmetric traffic requirements can be designed in.



Analysis, Prediction and Optimisation ofDownlink Capacity

The Story so Far

Coverage versus Capacity



The Story so Far

▫ It is possible to make “ball park” estimates of the capacity on

the downlink.

▫ The first step is to estimate a nominal pole capacity

▫ Then estimate the noise rise that can be produced at the“magic spot”.

▫ Hence deduce loading factor.

▫ This is a useful “first pass” planning calculation to perform.

▫ However, it does not consider an unevenly loaded network, nor

does it help us optimise network performance.

i N

E b 1

3840

0

Further Analysis of the Downlink

Downlink Analysis




▫ The concept of the “identical user”.

Identical:

•Bit Rate

•Eb/No

•Path loss

•Orthogonality

•Interference


Downlink Analysis




int

11)1( R

L

Tcom

L

user

L

user N P

L

P

L

P N

L

PP

▫ Power Received by each user:

PTcom

Puser

N-1“other users”

Bit Rate

Eb/No

Path loss

Orthogonality

PRint

i L

NPP

L

user Tcom


Downlink Analysis



u t e a ys s o t e o

RW

P L

P L

P N P

LP N E

R L

Tcom

L

user N

Luser b

int0

111 /

▫ Eb/No delivered to each user:

Total Transmitted Power

=

Tcomuser P NP

Capacity vs Link Loss

Downlink Analysis



Capacity vs. Link Loss

Link Loss(dB) Tx Power for25 users(dBm)

Maximumusers for 43dBm TxPower

110 37.62 64

125 37.69 64

140 39.33 49

145 41.63 33

150 45.26 16

dBm102 ;bit/s12200 dBm;36

dB;6N ;6.0 ;6.0 0

N Tcom

b

P RP

E i

Rapid, Approximate Method

Downlink Analysis



p , pp

▫ As Puser is allowed to approach infinity:

i N E

W RN

b

10

▫ The “Pole Capacity”.


Downlink Analysis



p , pp

▫ Identical Users will experience identical noise rise.

▫ Noise rise can be converted to throughput.

▫ We can predict the noise rise for given circumstances.

0

5

10

15

20

25

0 0.2 0.4 0.6 0.8 1 1.2


Downlink Analysis



p , pp

iP LP

LPiP

Tcom L N

L N T

1

1max

We can predict the noise rise for given circumstances.

The maximum noise rise that can be produced is

Then, capacity is given by

L N T

TcomT

b LPiPPP

N E W PC

1.

RiseNoise11

max

max

0

Effect of Link Loss on Capacity

Downlink Analysis



p y

There is a maximum capacity at low levels of link loss. High

transmit power allows this capacity to be approached at significantlevels of link loss.

0

200

400

600

800

1000

1200

120 130 140 150 160

Link Loss (dB)

C a p a c i t y ( k b i t / s

)

+37 dBm +40 dBm +43 dBm +46 dBm

Maximum Capacity

Downlink Analysis



At negligible levels of link loss, the expression

for noise rise becomes

Tcom

T

P

P max

kbit/s1.

1

3840

max0

T

Tcom

b P

P

i N

E

And capacity can be estimated from

Maximum Capacity

Downlink Analysis



If is taken to be fixed at, for example, 5,

then capacity is given byTcom

T

P

P max

kbit/s

1

3072

0i

N E b

And maximum capacity can be estimated from

kbit/s2.01

1

3840

0

i N E b

Effect of Orthogonality

Downlink Analysis



Graph shows the effect of orthogonality on the downlink

capacity for a link loss of 145 dB and i set at 0.6.

0

200

400

600800

1000

1200

0 0.2 0.4 0.6 0.8 1

Orthogonality

C a p a c i t y ( k

b i t / s )

BTS Power: 37 dBm 40 dBm 43 dBm 46 dBm

Effect of Out-of-Cell Interference

Downlink Analysis



Graph shows the effect of variations in the value of i on the

downlink capacity for a link loss of 145 dB and orthogonality of 0.6.

0

200

400

600

800

1000

1200

1400

0 0.4 0.8 1.2 1.6 2

Out of Cell Interference

C a p a c i t y ( k b i t / s )

BTS Power: 37 dBm 40 dBm 43 dBm 46 dBm

Extending the Validity – The Evenly-loaded Network

Downlink Analysis



So far, “identical users” have been considered.

Consideration is now given to an evenly loaded

network.

Crucially, is there a

representative value oflink loss and out-of-cell

interference that can

be used to estimate

downlink capacity?

Extending the Validity

Downlink Analysis



Experimentation with Monte

Carlo simulation suggests

that:

The effective value of link loss

is 4 dB less than that to the

edge of the cell.

The effective out-of-cell

interference ratio is 0.85.

Max throughput = kbit/s2458

0 N E b

Extending the Validity- An Example

Downlink Analysis



Network of cells with link loss

to edge of 133 dB.

Maximum throughput on

downlink at Eb/No of 7 dB

is 460 kbit/s for 43 dBm

transmit power.

Note:

Pole Capacity = 613 kbit/s

If 20% of power is for

common channels then

Max throughput = 490 kbit/s

for very low linkloss

Uplink-Downlink Balance

Downlink Analysis



Approximate Downlink Max Capacity =

Approximate Uplink Max Capacity =

Initial expectation is that loading factors will be higher

on the downlink.

Uplink Diversity and MHA will favour the uplink.

kbit/s2458

0 N E b

kbit/s2400

0 N E b

The Unevenly-loaded Network

Downlink Analysis



The situation is complicated by the fact that different

users experience different levels of noise rise.

For example, consider the case where there are 24

voice users, split into two, equal groups.

• Link loss = 120 dB

• DL i = 0.3

• NR = 2.1 dB

• Link loss = 140 dB

• DL i = 1.0

• NR = 1.4 dB


Downlink Analysis



For a more general

situation, the

maximum capacity is

often determined

using a Monte Carlo

simulator on a trial

and error basis.


Downlink Analysis



However, if the pole capacity is estimated from

Then a single simulation result can be used to estimatethe maximum downlink capacity.

Reports from the simulation include downlink traffic

channel power and throughput.

i N E b 1

3840

0


Downlink Analysis



As an example, it was found that a cell supported 300 kbit/s at an

Eb/No value of 6 dB using 33.9 dBm of traffic channel power.

Pole Capacity estimated at 772 kbit/s.

Hence representative noise rise estimated as 2.14 dB.

42 dBm of traffic channel power is available.

What noise rise (and hence throughput) would this cause?


Downlink Analysis



If 33.9 dBm causes 2.14 dB of noise rise then 42 dBm would

cause 7.1 dB of noise rise.

Loading factor of 80%.

Resulting throughput of 622 kbit/s at an Eb/No value of 6 dB.

Tested with Monte Carlo simulation and found to be valid for

general situations where the distribution of the new load was

similar to the existing load.


Downlink Analysis



For heavily concentrated “hot spot” situations.

A static analysis can result in an estimate for downlink values of i.

Optimising Throughput – using Pilot SIR

Downlink Analysis



The value of i influences

throughput. Any hotspotsshould be located where i is

low. Examining the Pilot SIR

as part of a static analysis

when the network is heavily

loaded will indicate the

throughput possible.


Downlink Analysis



If the pilot is at +33 dBm, the SIR reported will

be that for any traffic channel with the samepower.

This influences throughput.

E.g. SIR = -6 dB; target Eb/No = 4 dB

Maximum throughput for 33 dBm = 384 kbit/s.

33dBm = 2W so

192 kbit/s/watt (192 kbit/J).


Downlink Analysis



Pilot SIR varies between -5 dB and -12dB.

kbit/J parameter varies by a factor of 5.

Re-directing antennas can cause

variation by a factor of 3 or more.

Conclusions

Downlink Analysis



▫ Downlink capacity can be estimated for dimensioning purposes.

▫ Estimates compared with Monte Carlo simulator predictions.

▫ Estimates less accurate where network is not evenly loaded.

Simulations can lead to a more accurate estimate.

▫ Site location and antenna azimuth have key role in optimising

downlink throughput.



Network Dimensioning

Session ObjectivesNetwork Dimensioning



▪ To answer the questions:

▫ What is dimensioning?

▫ How might we carry out dimensioning?

▫ What are the key issues with dimensioning?

▫ What are the key equations that we need?

▫ What is sensitivity analysis and how is it carried

out?

What is Dimensioning?Network Dimensioning



▪ Dimensioning is the task of estimating the site numbers in a network

▪ Why do we need to know the number of sites?

▫ Project Management

▫ Rollout Strategy

▫ Vendor Comparison

▫ Configuration Comparison

▫ Business Planning

▪ We are NOT talking about dimensioning individual sites/links forcapacity

Dimensioning OutputsNetwork Dimensioning



▪ Site Numbers

▫ By region

▫ By configuration

▫ By environment

▪ Project Milestones and Required Resource

▪ Turnkey Zones

▪ Core Network and Transmission Network DimensioningInputs

▪ Sensitivity Analysis

▫ How sensitive is the dimensioning to changes in inputs...

Types of Dimensioning

etwor

Dimensioning



▪ There are many different ways to dimension anetwork

▫ There is no „right‟ way but there are many „wrong‟ways

▪ These can be generically grouped:

▫ Simple Coverage▫ Simple Capacity

▫ Simple Combined Coverage and Capacity

▫ Interactive Coverage and Capacity

▫ Benchmark Planning

„Spreadsheet‟ based

Planning Tool based

Dimensioning Inputs

etwor

Dimensioning



Environment

SiteConfiguration

Geographic Demographic

Service

Simple Coverage

etwor

Dimensioning



▪ Link Budget based

▫ i.e. simple numerical calculation

▪ Firstly a link budget is created

▪ The maximum path loss is used to calculate the cell

range using a propagation model

▪ The cell range is used to calculate the site area

▪ Site Numbers = (Total Area)/(Site Area)

Create Link Budget

Calculate Range

Calculate Site Area

Calculate Number

of Sites in a givenArea

Max PL

Max Range

Max Area

Key Issues with Simple CoverageDimensioning

etwor

Dimensioning



Dimensioning

▪ Capacity!

▪ Building Penetration

▪ Shadow Fading

▪ Propagation Model

▪ Site Environment Parameters

Shadow Fading and Building Penetration

etwor

Dimensioning



▪ Building Penetration

▫ Mean and standard deviation per environment ▪ Shadow Fading

▫ Typically calculated using „Jakes formula‟

▫ This assumes an isolated omni directional site…

b

aberf

b

abaerf F u

11

21exp1

2

12

2

0

xa

2

log10 10

enb Where: ;

x 0 - = Fade Margin

= location variability ( standard deviation )

n = Propagation Model Exponent

x0 -

x0 -

P(connect)

P(connect)

5.6

0

50%

76% 90%

75%

Point Location Probability

Area Location Probability

RH Clarke, "Statistical Theory of Mobile-Radio Reception," Bell SystemTechnical Journal 47, July 1968, pp. 957-1000

Shadow Fading

etwor

Dimensioning



▪ Shadow Fading

▫ The distribution of power signals is known as log-normal distribution

▪That is the signal measured in decibels has a normal distribution

▫ The process by which this distribution comes about is known ashadowing or slow fading

▫ Any variation in received signal is of the order of 10s to 100s of metres

▫ The standard deviation in decibels is known as the locationvariability

▪On average 5 - 12 dB ( 8 dB at 2GHz in urban environment )

▪Has a tendency to increase with frequency

▪No relationship has been proven between range and

▪With A = 5.2 in urban and 6.6 in suburban and fc in MHz

A f f cc log3.1log65.02

Propagation Model

etwor

Dimensioning



▪ COST 231 Hata typically used

▫ Rural very optimistic▫ No accounting for diffraction

▫ Typically considered inaccurate

▫ Accuracy limited to sites above30m, ranges above 1km,frequencies below 2GHz

▫ R in km

▫ fc in MHz

▫ hb base station height in metres

▫ hm mobile height usually 1.5m

▪ Very simple to implement...

▪ Assuming base station height tobe 30m and carrier of 1910MHzin a medium sized city

areasianmetropolit3dBG

citiessizedmedium0dBG

97.475.11log2.3

log55.69.44

log82.13log9.333.46

log

2

m

b

bc

dB

h E

h B

h f F

G E R BF L

R LdB log35137

Site Environment Parameters

etwor

Dimensioning



▪ Different configurations in different environments

▫ MHA

▫ Xpolar/Space Diversity

▫ Antennas

▪ Requires different link budgets…

▪ Loading on sites may also differ withenvironment

▫ To take advantage of the capacity-coverage

tradeoff

▪ Also different number of sectors

Area Calculation

etwor

Dimensioning



▪ Cells are complex shapes

▪ We assume in dimensioning that cellsconform to a regular shape

▫ Hexagons are commonly used because oftheir close packing properties

▫ K factors used to represent the differencebetween a circle of radius r and the site area

▫ The K factor will depend upon the number ofsectors

K = 0.827

K = 0.62

r

r

2

r K Area

Coverage-based Dimensioning: Example

etwor

Dimensioning



▪ Area to be covered: 80 km2.

▪ Link Budget for NR of 3dB suggestsmaximum path loss of 151 dB can betolerated, assuming sectored antennas areused.

▪ In building margin and shadow fading marginreduce this to 131 dB

▪ Path loss model

K = 0.62

R

dBlog35137 R L

km674.01010 35635137 L R

Coverage-based Dimensioning: Example

etwor

Dimensioning



▪ Area covered by 3-sectored site

▪ Number of sites required =

▪ 90 sites required (270 cells)

K = 0.62

R

km674.0101035635137

L

R

22 km88.062.0 R

9088.080

excel

Environment Distribution

etwor

Dimensioning



▪ Spreadsheets don‟t dealwith topology ormorphology accurately

▫ Hills, parks anddistributed target areas

▫ Interference and trafficcaptured by sites will

vary

▪ Margins for siteacquisition and overlapare required

Urban Area Site Numbers

Suburban

Area

Site Numbers?

Simple Capacity Dimensioning

etwor

Dimensioning



▪ Capacity calculation based

▪ Firstly calculate maximumcapacity per carrier

▪ Calculate maximum offeredtraffic per sector

▪ Calculate site area based ontraffic density

▪ Finally calculate themaximum number of sites inan area

Calculate CarrierCapacity

Calculate SectorOffered Traffic

Calculate MaximumSite Area

Calculate Numberof Sites in a Given

Area

Erlang-B

etwor

Dimensioning



▪ Erlang-B formula provides an

estimate of the peak traffic (notexceeded more than x%(usually 2%) of the time giventhe average traffic (quoted inErlangs).

▪ Erlang-B should only be usedfor:

▫ circuit switched traffic

▫ single services

▪ UMTS is multi-service andpacket switched...

Variation of demand with time

time

demand

average

peak

Erlang B

Accommodating a multi-service system

etwor

Dimensioning



▪ The Erlang B formula relies on the variance of the

demand equalling the mean (a Poisson distribution).

▪ If a particular service requires more than one “trunk”per connection, the demand is effectively linearlyscaled and the variance no longer equals the mean.

▪ Methods to investigate:

▫ Equivalent Erlangs

▫ Post Erlang-B

▫ Campbell‟s Theorem

Equivalent Erlangs Example

etwor

Dimensioning



▪ Let us consider 2 services sharing the same resource:

▫ Service 1: uses 1 trunk per connection. 12 Erlangs of traffic.▫ Service 2, uses 3 trunks per connection. 6 Erlangs of traffic.

▪ We could regard the above as equivalent to 30 Erlangs of service 1:

▫ 30 Erlangs require 39 trunks for a 2% Blocking Probability

▪ Alternatively, we could regard the above as equivalent to 10 Erlangsof service 2.

▫ 10 Erlangs require 17 trunks, (equivalent to 51 “service 1 trunks”) for a 2%blocking probability

▪ Prediction varies depending on what approach you choose.

Post Erlang-B

etwor

Dimensioning



▪ Consider 2 services sharing the same resource:

▫ Service 1: uses 1 trunk per connection. 12 Erlangs of traffic.

▫ Service 2: uses 3 trunks per connection. 6 Erlangs of traffic.

▪ We could calculate the requirement separately

▫ Service 1: 12 Erlangs require 19 trunks for a 2% Blocking Probability

▫ Service 2: 6 Erlangs require 12 trunks (equivalent to 36 “service 1trunks”).

▪ Adding these together gives 55 trunks.

▪ This method is known to over-estimate the number of trunks requiredas can be demonstrated by considering services requiring an equalnumber of trunks.

Post Erlang-B

etwor

Dimensioning



▪ Consider 2 services requiring equal resource:



▪ We could calculate the requirement separately

▫ Service 1: 12 Erlangs require 19 trunks for a 2% Blocking Probability

▫ Service 2: 6 Erlangs require 12 trunks.

▪ Adding these together gives 31 trunks.

▪ The accepted method of treating the above would be to regard it as a

total of 18 Erlangs that would require 26 trunks.

▪ Post Erlang-B overestimates the requirement.

Campbell’s Theorem

▪ Campbell‟s theorem creates a composite distribution where:

etwor

Dimensioning



▪ Campbell s theorem creates a composite distribution where:

▪ c is known as the capacity factor

▪ The amplitude used in the capacity is the amplitude of the target service

▪ Once the offered traffic and Capacity are derived, GoS can be derived withErlang-B -> similarly Required Capacity can be calculated if Offered Trafficand GoS target is known

caC Capacity ii

c

fficOfferedTra

i iii

i

iii

ba

ba

c

2

= mean = variance

i = arrival rateai = amplitude ofservicebi = mean holdingtime

iibγTrafficOfferedService

Campbell’s Theorem Example(1)

etwor

Dimensioning



▪ Consider the same 2 services sharing the same

resource:▫ Service 1: uses 1 trunk per connection. 12 Erlangs of traffic.

▫ Service 2, uses 3 trunks per connection. 6 Erlangs of traffic.

▪ In this case the mean is:

▪ The variance is:

3063121Erlangs iiii aab

6636112Erlangs 2222iiii aab


etwor

Dimensioning



▪ Capacity Factor c is:

▪ Offered Traffic for filtered distribution:

▪ Required Capacity for filtered distribution at2% GoS is 21

2.230

66

c

63.132.2

30 TrafficOffered c

α


etwor

Dimensioning



▪ Required Capacity is different depending upon target

service for GoS (in service 1 Erlangs):

▫ Target is Service 1 C1=(2.2 x 21) + 1 = 47

▫ Target is Service 2, C2=(2.2 x 21) + 3 = 49

▪ Different services will require a different capacity for the

same GoS. In other words: for a given capacity, the

different services will experience a slightly different GoS.

campbell

Traffic Analysis Methods Compared

etwor

Dimensioning



▪ Equivalent Erlangs

▫ Optimistic if you use the smallest amplitude of trunk (39)▫ Pessimistic if you use the largest amplitude of trunk (51)

▪ Post Erlang-B

▫ Pessimistic (55)

▫ Trunking efficiency improvement with magnitude ignored

▪ Campbell‟s theorem

▫ Middle band (47 - 49)

▫ Different capacities required for different services -realistic

▫ Preferred solution for dimensioning, but not ideal...

Capacity Dimensioning with Campbell’sTheorem

etwor

Dimensioning



Theorem

▪ Consider the following service definition and traffic forecast.

Service Amplitude Forecast

Voice 1 250 E64 kbps data 2 63 E

144 kbps data 4 41 E

384 kbps data 8 12 E


etwor

Dimensioning



Theorem

▪ Assuming we have n cells, we can determine the loadingper cell.

nnc

c

nnnnn

nnnnn

210028.3636meantrafficoffered

028.3636

1926

mean

variance

1926821414263250variance

636821414263250mean

222


etwor

Dimensioning



Theorem

▪ Unfortunately, we cannot now look up “210/n” in the ErlangB tables.

▪ We need to introduce a notional capacity per cell in termsof “Service 1 trunks”.

▪ We will assume that each cell has a capacity of 32 suchtrunks.

nnc

210

028.3

636meantrafficoffered


etwor

Dimensioning



Theorem

▪ Considering the equation

▪ C iis predefined as 32. a

idepends on the service we use as our

“benchmark”.

▪ Choosing service 3 as the “benchmark” service make aiequal to 4.

▪ Therefore 9.25 (or, rather, 9) trunks will service 4.34 Erlangs.

caC ii Capacity

25.9

028.3

4323

C


etwor

Dimensioning



Theorem

▪ 9 trunks will service 4.34 Erlangs.

▪ Therefore,

▪ Cell requirement is established at 48cells.

▪ Each of the cells will service:

▫ 5.21 Erlangs of voice

▫ 1.32 Erlangs of 64 kbps data



48

34.4210

n

n

campbell

Key Issues with Simple capacitydimensioning

etwor

Dimensioning



▪ What is the resource?

▫ Bitrate - no…

▫ Loading of individual user - yes…

▫ Calculate traffic analysis using the ratio of single channel loading fordifferent services

▪ Loading is affected by bitrate and E

b /N

0

▪ Note that uplink and downlink will yield different pole capacities

1amplitudefor1amplitudeforratebit

serviceforserviceforratebitamplitudeRelative

0

0

N

E

N E

b

b

Campbell amplitude

Simple Coverage and Capacity Dimensioning

etwor

Dimensioning



▪ Simply carry out both coverage ANDcapacity dimensioning and combine them,taking the maximum value

▪ This should be carried out on a „per

environment basis‟, preferably per region.

Complex Coverage and CapacityDimensioning

etwor

Dimensioning



▪ In this case a link is made between the coverage andcapacity

▪ The coverage is calculated from an initial link budget. Thislink budget will include an assumption of Noise Rise

▪ Then the number of subscribers captured per cell iscalculated

▪ The required loading to support the subscribers to thedesired GoS is calculated.

▪ This can be used to recalculate the Noise Rise…and fedback into the link budget

Complex Coverage and CapacityDimensioning (example)

etwor

Dimensioning



▪ Link budget is created assuming 4 dB Noise Rise

▪ 120 Cells required.

▪ Analysis of traffic forecast suggest each cell will experience

2 dB Noise Rise

▪ Re-create link budget

▪ 90 Cells required

▪ Analyse loading: 3 dB Noise Rise

▪ Re-create link budget

▪ 100 Cells required

▪ Analyse loading: 2.8 dB Noise Rise……….

Benchmark Planning

etwor

Dimensioning



▪ A „first pass‟ nominal plan is created

▪ The problems of non-contiguous clutter and diffraction are

removed

▪ A capacity „check‟ is required to ensure that cells aren‟t

overloaded

▪ Very resource hungry…

▪ Sensitivity analysis is impossible

Other Dimensioning Key Failings...

GSM/UMTS Interaction



▪ GSM/UMTS Interaction

▫ Proportion a percentage of voice traffic to GSM

▫ Don‟t assume that UMTS carries all of the traffic

▪ Microcells

▫ Offer capacity relief to macrocells

▫ This allows macrocells to be larger, potentially with a lower loading

▪ Repeaters

▫ Extend the coverage of macrocells at a lower cost than a new Node-B

▪ Sharing the load▫ Analysis so far has assumed that each cell looks after its own traffic. If

capacity is fully allocated on best server, a connection may be establishedwith a neighbour.



Cell Breathing

Uplink Analysis: Cell Breathing



▪ The base station has to achieve the required Eb/No ratio from

a particular mobile.

▪ Noise and interference is present from:

▫ Thermal Noise

▫ Other mobiles in the same cell

▫ Mobiles in other cells

▪ Remember that all mobiles use the same frequency

Noise Rise vs. Throughput

e om na

Plan



▪ Each new userincreases the

throughput of

the cell but also

increases the

effective noiseexperienced by

all other users.

Noise Rise vs. Throughput

0.00

5.00

10.00

15.00

20.00

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15

Throughput (x100kbps)

N o i s e R i s e

Coverage vs. Capacity Example

e om na

Plan



Two polygons were createdto allow traffic served by aparticular cell to be spreadover two geographicregions.


e om na

Plan



Firstly 50 terminals were

spread over the large outerpolygon to demonstrate thatcoverage existed in that area.


e om na

Plan



Next 340 terminals were spreadover the same area. The quality of

service is significantly reducedsuggesting that there is a capacityproblem. The report suggests that306 users may be the maximum thesite can support.


However, redistributing these

e om na

Plan



, gterminals so as to reduce path lossresults in a very good quality ofservice being restored. An averageof 330 terminals were served.Capacity and coverage are linkedtogether.

Cell Breathing :- “good” or “bad” ?

e om na

Plan



.

▪ Cell Breathing is integral to WCDMA cellular radio systems.

▪ Its disadvantage is that it leads to the creation of gaps in the network

coverage.

▪Its advantage is that it maximises capacity when it is demanded.

▪ The amount of cell breathing can be controlled by limiting the

NoiseRise in the admission algorithm. It cannot, however be

eliminated.

Cell Breathing :- “Good” or “Bad” ?

e om na

Plan



.

▪ Limiting the Noise Rise to 3 dB will restrict throughput to 50% of

theoretical maximum and restrict coverage shrinkage to 33% of its

maximum area.

▪ Allowing Noise Rise to increase to 10 dB will allow throughput to rise

to approximately 90% of its theoretical maximum but coverage

shrinkage will rise to 73% of maximum.

▪ Planning to restrict Noise Rise to 3 dB will necessitate the provision of

extra sites.

Cell Breathing

e om na

Plan



.

▪ Very rough rule of thumb.

Area shrinkage (%) =

Coverage with 3 dBNoise Rise

Coverage with 10 dBNoise Rise

Unloaded Coverage

5.17101100

NR

breathing

http://c/Ali_Stuff/laptop%20dump/china%203g/training/U104/Code/breathing.exe

http://c/Ali_Stuff/laptop%20dump/china%203g/training/U104/Code/breathing.exe



Dimensioning Packet Scheduled Traffic

Packet Scheduled Traffic

mens on ng ac et c e u e

Traffic



▪ Issues

▫ Adding Packet Traffic into gaps in CS demand

▫ Trunking Efficiency when delays are tolerated.

Gaps in CS demand


Traffic



▪ Each cell will have a notional capacity in kbps

▪ Simulation will provide details of mean demand during “busy

hour”.

▪ Difference between the two can be construed as “free

capacity”.

Efficiency using Erlang C

▪ Erlang C formula will predict probability of a given delay


Traffic



Erlang C formula will predict probability of a given delayfor a given number of trunks and Erlangs of offered

traffic

▪ If unlimited delay can be tolerated, efficiency will be100%.

Example:

25 Servers, 20 Erlangs of Traffic offered.P(>0) = 0.21

P(T1/T2>0.4)=0.02

P(T1/T2>0.8)=0.0038

P(T1/T2>1.2)=0.00052

P(T1/T2>1.6)=0.00007

P(T1/T2>2.0)=0.00001

T1 is delay time

T2 is mean holding time

Erlang c

Delay Categories for UMTS


Traffic



Category Example Delay

Conversational Interactive Games <250 ms

Interactive Web Browsing <4 s

Streaming ftp transmissions <10 s

Background e-mail >10 s

Example Result

mens on ng ac et c e u e Traffic



▪ 25 “Trunks”, each with 12200 bps capacity.

▪ Mean delay of 10 ms acceptable. P(T1/T2>1)=0.02

▪ Result: 22 Erlangs of traffic carried.

▪ Note: Erlang B for 0.02 GoS would predict 17.5 Erlangs

carried.

▪ Accepting a delay increases trunking efficiency.

Erlang B

Erlang C for Shoppers




Factors affecting the delay




▪ Delay is proportional to the holding time.

▪ Holding time proportional to packet size.

▪ Need to consider “sensible” packet size.

▪ www model has standard packet size of 3840 bits.

▪ Short text message of 128 alphanumeric characters will

require approximately 1 kbit of capacity.

Probability of the Delay exceeding a givenvalue.




▪ The higher the delay, the lower the probability.▪ The curve is an exponential decay.

Probability vs. Delay

0

5

10

15

20

25

30

0 20 40 60 80 100 120 140

Delay (ms)

P r o b a b i l i t y ( % )

Line 1

Dimensioning “all packet” services




▪ Packet traffic may dominate demand in UMTS.

▪ Efficiency issues for packet data become very

significant.





▪ Tolerance of delays implies tolerance of re-transmissions.

▪ Retransmission possibility allows for very low Eb/No

▪ Target Eb/No as low as 1 dB with 25% FER will provide maximumefficiency.

▪ 33% extra demand due to retransmission more than offset by

increased capacity.


▪ If i = 0 5 Eb/No = 1 dB pole capacity = 2 Mbps




If i = 0.5, Eb/No = 1 dB, pole capacity = 2 Mbps.

▪ At 50% loading factor, a capacity of 1 Mbps could be

expected.

▪ 750 kbps of offered traffic will result in 1 Mbps carried if

retransmission ratio is 25%.

▪ High trunking efficiency will result in very high levels of traffic

serviced.

▪ Delay implications of retransmissions need to be addressed.

▪ Frame period of 10 ms significant.

Packet Scheduling for Web Browsing

▪ Web Browsing involves near-total asymmetry in downlink




Web Browsing involves near total asymmetry in downlinkdirection.

▪ It is possible to plan a cell so that, at the maximum pathloss, the downlink capacity is greater than the uplinkcapacity.

▪ Low Eb/No requirement of packet data can make great useof this extra capacity.

▪ E.g. 320 kbps at Eb/No of 1 dB imposes same loading as100 kbps at an Eb/No of 6 dB.

Packet Scheduling for Web Browsing

E l




▪ Example.

▫ 40 x 12.2 kbps speech channels on uplink and downlink.

▫ Analysis suggests that downlink transmitter is operating 4

dB below maximum.

▫ This extra 4 dB of power can be converted into

approximately 850 kbps of data throughput for packet

scheduled traffic.

Effect of Base Station Power on DownlinkNoise Rise




▪ Increasing Base Station Power will increase the Noise Rise

on the downlink but not in a particularly straightforward way.

▪ The Analysis involves imagining that the base station

transmits an effective thermal noise power along with the user

power.





User Power B dBm

Noise Power A dBm

S

10

10

10

1010

10

10110log

10

101010logRise,Noise

A B

A

B A

X





110

10

10

11010

10101

10

1010

1010

1010

X

B A

X A B

X A B

110

10log10

110

10log10

10

10poweruser

10

10

10

10

NR

X

B

A

The first step in calculating the new noise rise is to assign a value to thepower of the virtual noise source, A.





1010 101log10RiseNoiseNew

AC

Then if the user power is increased from B dBm to C dBm:

Effect of Base Station Power on Downlink NoiseRise




Example: Original Noise Rise = 2 dB. Base Station Transmit Power

= 36 dBm.

Transmit Power is increased to 41 dBm.

Intermediate Parameter A = 38.33 dBm

New Noise Rise is 4.55 dB.

(Noise Rise has increased by 2.55 dB for a Base Station powerincrease of 5 dB. This increase in Noise Rise represents an increase

in loading factor and, hence, an increase in throughput.)

10

10 101log10RiseNoiseNew AC

110

10

log10

110

10log10

10

10poweruser

10

10

10

10

NR

X

B

A

New noise rise

Pause for Breath




▪ General Ideas:

▫ A small amount of extra capacity for CS traffic withhigh Eb/No can be converted to a larger amount oftraffic for a service at a low Eb/No.

▫ Extra Power available in the downlink can beconverted to extra capacity by increasing the Noise

Rise and hence the loading factor.

Calculating Extra Capacity

▪ Example




▪ Example.

▫ CS traffic Eb/No 6 dB, provides 225 kbps of traffic (uplink anddownlink) for a cell.

▫ i = 50%.

▫ Orthogonality on the downlink, = 0.6

▫ 32 dBm user power (max. available 42 dBm) on downlink.

▫ Our challenge is to determine the extra capacity available on the

uplink and downlink for a service requiring an Eb/No of 1.5 dB if

the base station user power can be increased to 42 dBm and the

Noise Rise limit on the uplink is set at 3 dB.

U li k l i





▪ Uplink analysis.

▫ At an Eb/No of 6 dB, the pole capacity isapproximately 640 kbps and throughput at 3 dBNoise Rise would be 320 kbps.

▫ 320 - 225 = 95 kbps available.

▫ 95 kbps at an Eb/No of 6 dB produces anequivalent loading as 268 kbps at an Eb/No of 1.5dB.

▫ 268 kbps of data can be added to the uplink.





▪ Downlink analysis.

▫ At an Eb/No of 6 dB, the pole capacity on the downlink is

approximately 1067 kbps and a throughput of 225 kbps

represents a loading factor of 21% and, hence, a Noise

Rise of 1 dB.

▫ Increasing the base station power from 32 dBm to 42 dBm

would increase the Noise Rise to 5.55 dB.

▫This represents 72% loading factor, an increase of 51%.

Downlink analysis continued





▪ Downlink analysis continued.

▫ At an Eb/No of 1.5 dB, the pole capacity on the downlink

is approximately 3000 kbps 51% loading factor increase

represents 1530 kbps.

▪ Conclusion is that the cell could accommodate additional 268kbps packet data on the uplink and 1530 kbps packet data onthe downlink.

▪ Provision of bearers to service the 1530 kbps additional

packet data would have to be addressed as a separateproblem.

Conclusions




▪ Tolerating Delays increases trunking efficiency -

efficiencies higher than 80% can usually be assumed.

▪ Tolerating retransmissions reduces required Eb/No to

as low as 1 dB.

▪ Packet data that is heavily asymmetric in the downlink

direction can be planned for at the link

budget/dimensioning stage.



Optimisation Procedures

Optimisation Techniques

pt m sat on Procedures



)1(

3840CapacityPole

0

i N

E b

Parameter i is crucial. Limiting mutual interference

between cells will increase capacity

Limiting mutual interference




• Downtilt antennas.• Consider mounting antennas onthe side of buildings.





Controlling the backlobe can produce asmall but significant improvement in

capacity.

0º

0ºElec 6ºMech

0º 0º

6º

6º 6º

6ºElec 0ºMech

0º

6º

6º 0º

6ºElec -6ºMech

0º

-6º

12º

0º





• Key parameter: Frequency Re-use Efficiency(FRE).

(W)ceinterferencell-intertheis

(W)ceinterferencell-intratheis

FRE

Inter

Intra

Inter Intra

Intra

N

N N N

N

Note FRE = 1/(1+i)


• Avoid high sites they become overloaded quickly




• Avoid high sites, they become overloaded quickly.

Balanced siteswill divide load

evenly.

• High sites will

“grab” too much

traffic.•They will gather a

lot of interferenceon the uplink.•They willcontribute a lot ofinterference on thedownlink.

Multi-User Detection

▪ Multi-User detection (MUD) is a method used toi th f f th i b d i th




improve the performance of the receiver by reducing the

noise contributions from other CDMA users.

▪ The concept is based on the fact that noise from CDMAusers, although usually approximated with AWGNcharacteristics, inherently consists of coherent signals.

▪ MUD reception decodes a number of userssimultaneously and subtracts their noise contributions tothe each other

▪ Essentially this results in a more sensitive receiver

Visualising the Processing Gain w/o MUD

W/Hz W/Hz W/Hz




W/Hz W/Hz W/Hz

Ec

No

SignalIntra-cell Noise

Inter-cell Noise

Before Spreading

After Spreading With Noise

f f f

W/Hz


f

W/HzE

b

No

Post Filtering Orthog = 0

f

dBW/Hz

Eb

No

Eb /No

f

Visualising the Processing Gain with MUD

W/Hz

Post Filtering




W/Hz

Signal

Inter-cell Noise


Filtering

f

Other Users

Eb

No

W/Hz

f

Eb

No

W/Hz

f

Eb

No

W/Hz

f

Eb

No

W/Hz

f

Because of MUD the contribution of the other users to the Noise is Reduced.

It is not completely eliminated because of the inaccuracies of the Multiple access interference estimation.

Multi-User Detection




▪ Modelling the MUD receiver can beachieved by adjusting the Eb/No of theservices to account for the improved noisecancellation.

▪ Compensates for cable loss between antenna and base station(t i ll 3dB)

Mast Head Amplifiers




(typically 3dB)

▪ MHA used to increase coverage range

▪ Gain 12dB (adjustable)

▪ MHA‟s are also called TMA (Tower Mounted Amplifiers)

- LNA‟s used on receive path

▪ Increase uplink capacity

▪ Only beneficial in uplink-limited situations

▪ 1.6 dB Noise Figure (NF), Gain 12 dB

▪ Drawback - Insertion loss on

Tx path (~ 1.3 dB)Ant

Bias-TDC

TMA

by pass

Uplink Receive Space Diversity

▪ Common to have two receive antennas per sector at the base station.




▪ Even if highly correlated, coherent combination should yield ~3 dB

improvement.

▪ In practice a gain of 4 dB or more is expected from antennas spaced 2-3 m

apart.

Receiveantenna 1

Receiveantenna 2

Uplink Receive Space Diversity

▪ This is not “conventional” space diversity




▪ This is not conventional space diversity.

▪ Each antenna is connected to a separate finger of the Rake receiver.

▪ This is possible due to the synchronisation and channel estimation

derived from the Pilot channel.

▪ Thus Eb/No is improved, rather than simply an effective power gain.

▪ Final limitation of this technique (why not 100 or 1000 antennas?) is

interesting to consider

▪ Very low individual Eb/No will probably mean a very low pilot level which

will lead to poor coherence and little gain - process becomes “self -

defeating”.

Downlink Transmit Diversity

▪ UMTS explicitly allows the use of transmit diversity from the base station




▪ However it is not possible to simply transmit simultaneously from two closeantennas as this would cause an interference pattern

▪ Mobile terminals must have the capability of implementing downlink transmitdiversity .

Transmitantenna 1

Transmitantenna 2

Downlink Transmit Diversity

▪ The following methods are suggested in the UMTS




Transmit DiversityMethod

Description

TSTD Time Switched Transmit antennaDiversity (open loop)

STTD Space Time block coding Transmitantenna Diversity (open loop)

Closed Loop Mode 1 Different Orthogonal Pilots

Closed Loop Mode 2 Same Pilot

▪ The following methods are suggested in the UMTS

standards to avoid the problem of the interferencepattern

Time Switched Transmit antenna Diversity(TSTD) Synch channel only




▪ Even numbered slots transmitted on Antenna 1, oddnumbered slots on Antenna 2

Antenna 1

Antenna 2

P-SCH

Slot #0 Slot #1 Slot #14Slot #2

P-SCH

P-SCH P-SCH

S-SCH

S-SCH

S-SCH S-SCH

Space Time block coding Transmit antennaDiversity (STTD)

STTD di i ti l i UTRAN STTD t i d t t th UE




b0 b1 b2 b3

b0 b1 b2 b3

-b2 b3 b0 -b1

Antenna 1

Antenna 2

Channel bits

STTD encoded channel bits

for antenna 1 and antenna 2.

▪ STTD encoding is optional in UTRAN. STTD support is mandatory at the UE

▪ Channel coding, rate matching and interleaving is done as in the non-diversity mod

▪ STTD encoding is applied on blocks of 4 consecutive channel bits

▪ The bit bi is real valued {0} for DTX bits and {1, -1} for all other channel bits.

Analysis of STTD




▪ STTD encoding effectively spreads a data bit across more than one bit period.

▪ This leads to a general improvement in noise performance.▪ Further, it allows a stronger signal from one antenna to dominate.

b0 b1 b2 b3-b2 b3 b0 -b1

b0-b2 b1+b3 b0+b2 b3-b1

Processing alternate bits will extract the data

Closed Loop Mode

▪ Channel coding, interleaving and spreading are done as in




g, g p gnon-diversity mode

▪ The spread complex valued signal is fed to both TXantenna branches, and weighted with antenna specificweight factors w1 and w2.

▪ The weight factors are determined by the UE, and signalled

using the FBI field of uplink DPCCH (Dedicated PhysicalControl Channel).

1 radio frame: T = 10 ms

Pilot

Npilot bits

TPC

NTPC bits

Slot #0 Slot #1 Slot #i Slot #14

Tslot = 2560 chips, 10 bits

f

DPCCHFBI

NFBI bitsTFCI

NTFCI bits

Closed Loop Mode

Spread/scramble

w1

CPICH1

Tx

Ant1




Spread/scramble

w2

DPCHDPCCH

DPDCH

Rx

Rx

CPICH2

Ant2

Tx

Weight Generation

w1 w2

Determine FBI message

from Uplink DPCCH

Soft Handover as an optimisation technique

▪ Soft handover is a form of space diversity.




Soft handover is a form of space diversity.

▪ Issues.

▫ All except one signal may be poor

quality.

▫ Handover margin can influence the

gain achieved.

▫ 2 dB Eb/No gain possible.



UTRAN

Architecture and Protocols

Protocol Model for UTRAN Interfaces

▪ Protocol structures in UTRAN are designed in layers and planes

UTRAN



▪ Protocol structures in UTRAN are designed in layers and planes.

▪ They are seen as logically independent of each other

▫ However they will physically interact.

▪ Being logically independent allows for changes to blocks in the

future

theoretically!

General Protocol Model forUTRAN Terrestrial Interfaces

UTRAN



Horizontal Layers in the General Protocol Model

UTRAN



▪ All UTRAN related issues are only visible in the Radio

Network Layer

▪ The Transport Layer simply represents standard transport

technology for use in UTRAN

▫ e.g. ATM and appropriate ATM Adaptation Layers

▪ AAL2 ( voice ) and AAL5 ( data/control)

▫ UDP/IP or RTP/UDP/IP ( released 6 ? )


UTRAN



▪ The Control Plane is provided for all UMTSspecific control signalling including:

▫ Application Protocols

▫ Signalling Bearers

▪ The User Plane is provided for all data sent andreceived by the user including:

▫ Data Streams

▫ Data Bearers


UTRAN



▪ The Transport Network Control Plane also includes the AccessLink Control Application Part, ALCAP.


Plane


Plane

Transport Network Control

Plane

ALCAP

ALCAP - Access Link Control Application Part

UTRAN



▪ ALCAP sets up the transport bearers for the User

Plane

▪ The independence of the User and Control Planeassumes that

▫ ALCAP signalling transaction actually takes place.

Iu, the UTRAN-CN Interface

▪ The Iu has two different instances.

UTRAN



▫ Iu-CS for circuit switched connections

▫ Iu_PS for packet switched connections

▪ Originally proposed as one interface but the differences

between circuit and packet switched services has produced

two.

▪ The Control plane is the same for CS/PS as far as possible!

Protocol structure for the Iu-CS

UTRAN

▪ The Iu-CS control plane



▫ RANAP - Radio Access Network Application Part

▫ Broad-band SS7 ( Network-Network Interface NNI )

▫ ATM - AAL 5

▪ The Iu-CS transport network control plane

▫ Signaling protocol for setting up ATM AAL2

▫ Broad-band SS7

▫ ATM - AAL 5

▪ The Iu-CS user plane.

▫ ATM - AAL2 reserved for each CS connection

Protocol structure for the Iu-CS

UTRAN



▪ The Iu-PS control plane

Protocol structure for the Iu-PS

UTRAN



▫ RANAP - Radio Access Network Application Part

▫ Broad-band SS7 ( NNI ) or IP based signalling bearers

▫ ATM - AAL 5

▪ The Iu-PS transport network control plane

▫ NONE required

▪ the GPRS tunneling protocol only requires the IP addresses

which RANAP has supplied.

▪ The Iu-PS user plane.


UTRAN



p

▫ ATM - AAL5

▫ multiple packet data flows are multiplexed onto several

permanent virtual connections PVCs

▫ Each flow uses

▪ GTP-U,

▪ UDP connectionless transport and IP

addressing.


UTRAN



RANAP

▪ The functionality of the RANAP is performed through Elementary

Procedures EP‟s

RANAP has 12 defined functions

UTRAN



▪ RANAP has 12 defined functions

Relocation Iu Release Radio Access BearerManagement

Paging Location reporting Common IDManagement

Security Mode Control UE-CN signallingtransfer

OverloadManagement

Reset ReportingUnsuccessfultransmitted data

Management ofTracing

Iur, the RNC-RNC Interface

▪ The Iur provides support for four distinct functions:

UTRAN



▪ The Iur provides support for four distinct functions:

▫ Inter RNC Mobility

▫ Dedicated Channel Traffic

▫ Common Channel Traffic

▫ Global Resource Management

▪ The Radio Network System Application Part, RNSAP

is therefore divided into four modules.

RNSAP Iur1: Inter RNC Mobility

UTRAN



▪ Iur1 provides for the mobility of the user between two RNCs but

does not support the exchange of user data traffic.

▪ If the network fails to provide this then the only means of mobility

is to disconnect from RNS1 and reapply for connection in RNS2

Cell

Cell

RNSAP Iur2: Dedicated Channel Traffic

UTRAN



▪ Iur2 sets up and maintains a dedicated channel between two

RNCs

▪ Used in inter-RNC soft handover.

▪ Provides the serving SRNC with the capability to manage the

radio links in drift DRNCs

Cell

Cell SRNC

DRNC

RNSAP Iur3: Common Channel Traffic

UTRAN



▪ If Iur3 is not implemented then every time an inter-RNC cell

update takes place then that RNC becomes the serving RNC.

Cell

Cell

SRNC

RNC

RNC

SRNC

SRNC

RNC

No Iur3

RNSAP Iur4: Global Resource Management

UTRAN



▪ This function is considered optionally as it does not

transmit any user data across the Iur.

▪ However it does provide useful information

▫ transfer of cell measurement between RNCs

▫ transfer of Node B timing information between RNCs

Protocol Structure for Iur

UTRAN



Iub, RNC-Node B Interface

▪ Iub user plane, defines every type of transport channel, using

UTRAN



AAL2

▪ Iub signalling control plane, AAL5, is divided into 2 essential

components;

▫ NBAP-C Common Node B application part

▫ NBAP-D dedicated Node B application part

NBAP-C

Set-up of 1st radio link to the UE

Cell configuration

Handling of RACH/FACH andPCH

NBAP-D

Addition, release of radio links

Handling of softer handover

Handling of dedicated and sharedchannels on a Node B - UE.x basis

Protocol structure for the Iub

UTRAN



Note:

SSCF servicespecific co-ordination functionAAL set-up

UNI .. user -network interface

SSCOP servicespecific connectionoriented protocolAAL set-up



ATM Overview

ATM

▪ ATM Concepts

ATM



▫ ATM is based on a Virtual Circuit Technology

▫ Similar to Circuit Switching, ATM uses signalling protocols toestablish a Circuit before data communication commences.

▫ Unlike Circuit Switching, ATM is based on Statistical Multiplexing(similar to Packet Switching)

ATM

▪ Virtual Circuit Concept

▪ A connection is first established using signallingprotocols

ATM



▫ A route from the source to the destination is chosen

▫ The same route is used for all cells (fixed sizepackets) whilst the connection is maintained.

▪ Consequently no routing decision is needed forevery cell

▪ No dedicated path is required unlike CircuitSwitching

▪ Each Link of the network is shared by a set of

virtual channels▫ Each cell uses only one virtual channel number

ATM

▪ Each packet contains enough information forthe node (switch) to forward it towards the

ATM



destination.

▪ Tables at nodes are filled with connection data

▪ Parameters used for establishing VirtualCircuits

▫ Calling and Called Party Addresses

▫ Traffic Characteristics

▫ QoS Parameters

ATM VCI & VPI Assignment

ATM



IN LINK IN VCI VPI OUT LINK OUT VCI VPIA 20 10 AB 4 3

AB 4 3 B 19 7

ATM

▪ Advantages of Virtual Circuits

ATM



g

▪ In-order delivery of packets or cells

▪ Fast Delivery (no routing decision foreach packet)

▪ Less Header Overhead▪ High efficiency when two stations

exchange data for long time

ATM

▪ Handling Congestion with Virtual Circuits

ATM



▪ Establishing Virtual Circuits alone is not sufficientto avoid congestion

▪ We must declare Traffic Characteristics and QoSrequirements.

▪ Resources must be reserved while establishingVirtual Circuits

▪ ATM network is made up of Switches &

ATM

ATM



Endpoints ▪ Switch

▫ Accepts incoming cell

▫ Reading and updating cell header

▫ Switching cell towards output interface

▪ Endpoints

▫ Contains ATM network interface adapter

▫ Ex. Workstations, routers, DSUs, LAN Switches,Video CODECs

ATM

ATM



ATM SwitchRouter

LAN Switch

Work Station

ATM Switch Router

ATM End points

NNI

UNI UNI

▪ Interfaces

▫ ATM supports two primary interfaces

ATM

ATM



▪ User Network Interface (UNI)

▫ Connects end system to an ATM switch

▫ RNC - Node B

▪ Network Network Interface (NNI)

▫ Connects two ATM switches

▫ RNC - MSC, RNC - RNC

▫ UNI and NNI can further be subdivided to Public and PrivateUNIs and NNIs

ATM - Interfaces

Private ATM Public ATM

ATM



PrivateNNI

PrivateUNI

PublicNNI

PublicUNI

Cell Format

▪ ATM Cell Format

ATM



▪ Each cell consists of 53 bytes

▪ First 5 bytes for cell header

▪ Remaining 48 bytes for payload

8 bits

53 bytes

Payload (48 Bytes)

Header (5 Bytes)

Cell-Header

GFC VPI VCI PT HECC

L

P

at

UNI

ATM



VPI VCI PT HEC

12 16 3 1 8 bits

C

L

P

4 8 16 3 1 8 bits

at

NNI

GFC : Generic Flow Control VPI : Virtual Path IdentifierVCI : Virtual Circuit Identifier PT : Payload TypeCLP : Cell Loss Priority HEC : Header error CheckUNI : User Network Interface NNI : Network-Network Interface

ATM

▪ ATM Cell Header Bits

ATM



▫ Generic flow control GFC▫ Only in the cells transported over UNI

▫ Enables a local switch to regulate flowcontrol

▫ In NNI these 4bits are part of VPI

▫ Virtual Path Identifier VPI

▫ Used for identification/routing purposeswithin network

▫ 8 bits at UNI

▫ 12bits at NNI

ATM

▫ Virtual Circuit Identifier VCI

ATM



▫ Identification/routing purposes within network▫ 16bit field

▫ Payload Type PT

Type of payload carried within a cell

▫ user data

▫ operation and maintenance data (OAM)

▫ congestion indication (CI) bit

▫ CI bit may be modified by any switch to indicate

congestion to end users

ATM

▫ PT Interpretation

ATM



000 User Data; type 0; no congestion

001 User Data, type 1; no congestion

010 User Data; type 0; congestion

011 User Data; type 1; congestion

100 OAM Cell

101 OAM Cell

110 Resource Management Cell (to be defined)

111 Reserved for future use

ATM

▫ Cell loss Probability CLP : (1 bit)

▪ Indicates relative priority of a cell

ATM



▪ Indicates if a cell can be discarded in case ofcongestion

▪ CLP = 0; High priority; cell not to be discarded

▪ CLP = 1; Low priority; cell may be discarded

▪ CLP bit is set by the user or by the service

provider

▫ Header error check HEC

▪ 8bit field

▪ cyclic redundancy check CRC on first four bytes(32 bits) of header

▪ Permanent Virtual Circuits (PVC)

▫ Direct Connectivity

Circuit Provision

ATM



▫ No call setup procedure

▪ Switched Virtual Circuits (SVC)

▫ Created & released dynamically

▫ Flexible Connection

▪ Connectionless services

▫ Multipoint SVC’s used for broadcast IPpackets

•IP over ATM implementation

•RFC-2684 ( PVC‟s, SVC‟s and multipoint SVC‟s )

•RFC-2225 (SVC‟s)

VCs are grouped together to create VPs and VPs are

Recap - Virtual Connections

ATM



bundled together to create a transmission path

Virtual Channel

Virtual Path

Transmission Path

ATM Reference Model

OSI Reference Model ATM Reference Model

ATM



Higher Layers

ATM Adaptation Layer

(AAL)

ATM Layer

Physical Layer

MANAGEMENT

LAYER

P

LA

N

E

CONTROL USER

Physical Layer

Data Link

Network

Transport

Session

Presentation

Application

Planes (Application Functions)

ATM Reference Model

ATM



▪ Control Plane

▪ Generating & managing signallingrequests

▪User Plane

▪ Managing the transfer of user data

▪ Management Plane

▪ Layer management

▪ Plane management

ATM Reference Model

ATM



Layers▫ ATM Layer

▫ ATM Adaptation Layer (AAL)

▫ Higher layers above AAL accept user data,

arrange it into packets and hand it to AAL.

▪ The ATM Layer is responsible for:

ATM Layers

ATM



▫ Addition/removal of correct cell header

▫ Multiplexing cells into a single stream received fromAAL

▫ Demultiplexing of cells and relaying to AAL

▫ Handling of Connection Identifiers

▫ Generic flow control

ATM Ad i L AAL

ATM Layers

ATM



▪ ATM Adaptation Layer, AAL

▫ Convergence function between user layer and ATM layer

▫ Conversion of source information to 48 byte segments

▫ Allocates service classes to support various information

sources

ATM Adaptation Layer

ATM



AAL1 AAL2 AAL3/4 AAL5

Timing

Bit rate

Mode

Yes No

Constant Non-real time

Connection oriented Connectionless

Yes No

Real-time, variable NRT Variable

ATM

▪ AAL Types▫ AAL1

ATM



▪ Constant bit rate CBR Example: Circuit Emulation

▪ Connection oriented

▪ Timing information exists

▫ AAL2

▪ Real Time variable bit rate VBR Eg: Traffic UMTS Node B -RNC


▪ Requires timing information

▪ Ex: Compressed video

ATM

▪ AAL 3/4

ATM



▪ Non-real time VBR Example: FrameRelay

▪ Connection oriented or connectionless

▪ No timing information

▪ AAL5▪ Variable bit rate Eg: Packet data and

signalling UMTS


▪ No timing information

▪ Simpler than AAL 3/4▪ Started in ITU; Completed in ATM Forum



ATM Segmentation & Assembly

ATM AAL2

ATM Segmentation Process



▪ The principle ATM adaptation layer specified forUTRAN connection is AAL2

▪ AAL2 provides for real-time fixed delay traffic

ATM AAL2




▪ The AAL2 receives from the ATM layerinformation in the form of a 48 byte ATM ServiceData Unit (ATM-SDU).

▪ The AAL2 passes to the ATM layer information inthe form of a 48 byte ATM-SDU.

AAL2 Common Part Sublayer (CPS)

▪ The AAL2 CPS provides the capabilities to transfer CPS-SDUsfrom one CPS user to another CPS user through an ATM network.




▪ Two types of CPS users are supported:

▫ Service Specific Convergence Sublayer, SSCS entities

▫ Layer Management.

Iub interface

▪ The service offers a peer-to-peer operation:

▫ data transfer of CPS SDUs of up to 45 (default) or 64 bytes





▫ data transfer of CPS-SDUs of up to 45 (default) or 64 bytes▫ multiplexing and demultiplexing of multiple AAL type 2 channels

▪ CPS-SDU sequence integrity is maintained on each AAL type2 channel. ( useful for voice ! )

▪ However some CPS-SDUs may be lost; and lost CPS-SDUswill not be retransmitted. ( pointless for voice )

▪ The AAL2 CPS possesses the following characteristics:





▫ The AAL2 channel is a bi-directional virtual channel.

▫ The same value of channel identifier value is used for bothdirections.

▫ AAL2 channels can be established over

▪ ATM layer Permanent Virtual Circuit (PVC)

▪ Switched Virtual Circuit (SVC).

▪ The Common Part Sublayer merges several streams of CPS-

Packets onto a single ATM connection

Procedure of AAL2 Common Part Sublayer (CPS)




Packets onto a single ATM connection.

▪ The Common Part Sublayer receives CPS-SDUs from one ormore SSCS transmitter processes.

▪ CPS then multiplexes and packs the CPS-Packets into CPSProtocol Data Units, CPS-PDUs.

▪ At the CPS receiver, the CPS-Packets are unpacked anddemultiplexed and passed to one of the SSCS receivers.

▪ A CPS-Packet consists of a 3 byte CPS-Packet Header (CPS-

PH) followed by a CPS Packet Payload (CPS PP)

Format and coding of the CPS-Packet




PH) followed by a CPS-Packet Payload (CPS-PP).▫ The size and positions of the fields of the CPS-Packet are shown

below

▪ Channel Identifier (CID) „maximum 255 different sources‟

CPS Packet




▪ Channel Identifier (CID) maximum 255 different sources

▫ The CID value identifies the AAL2 CPS user of the channel.

▫ The AAL2 channel is a bi-directional channel.

▫ The same value of channel identification is used for bothdirections.

▫ Only 248 used for user data

▪ Length Indicator (LI)

▫ The LI field is binary encoded with a value that is one less than the

number of bytes in the CPS-Packet Payload

CPS Packet




number of bytes in the CPS-Packet Payload.

▫ The default maximum length of the CPS-Packet Payload is 45bytes; otherwise, the maximum length can be set to 64 bytes.

▫ The maximum length is channel specific, i.e. its value need not becommon to all AAL2 channels.

▫ However, for a given CID value, all CPS-Packet payloads mustconform to a common maximum value.

▪ User-to-User Indication (UUI)

▫ The UUI field serves two purposes:

CPS Packet




▫ The UUI field serves two purposes:

▪ to convey specific information transparently betweenthe CPS users, i.e. between SSCS entities orbetween Layer Management

▪ to distinguish between the SSCS entities and LayerManagement users of the CPS

▪ User-to-User Indication (UUI)

▫ The 5-bit UUI field uses

CPS Packet




The 5-bit UUI field uses

▪ 0 to 27 for SSCS entities,

▪ 30 to 31 for Layer Management.

▪ The values 28, 29 are reserved for futurestandardization.

▫ The contents of the UUI field are used to transport the UUIparameters

▪ CPS-UNITDATA primitive

▪ MAAL-UNITDATA primitive

▪ Header Error Control (HEC)

▫ The transmitter shall calculate the remainder of the division(modulo 2) by the generator polynomial x5 x2 1 of the

CPS Packet




The transmitter shall calculate the remainder of the division(modulo 2), by the generator polynomial x5 x2 1, of theproduct of x5 and the contents of CID,LI and UUI (19 bits) ofthe CPS-PH.

▫ The coefficients of the remainder polynomial shall beinserted in the HEC field with the coefficient of the x4 term

in the most significant bit of the HEC field.▫ The receiver uses the contents of the HEC field to detect

errors in the CPS-PH.

▫ Now that the CPS-Packet is formed we can put it into aCPS Protocol Data Unit. This has a single byte headerconstructed as follows.

CPS-PDU ▪ Format and coding of the Common Packet Sublayer-Protocol Data Unit

CPS-PDU

▫ The CPS-PDU consists of a single byte start field and a 47-byte payload.

▫ The 48-byte CPS-PDU is the ATM-SDU.

Th i d iti f th fi ld f th CPS PDU h b l


http://c/Ali_Stuff/Removable%20Disk%20(E)/software/atm%20aal2.xls



▫ The size and positions of the fields of the CPS-PDU are shown below.

▪ CPS-PDU start field (STF) (8 bits)

▪ a) Offset Field (OSF) (6 bits)

▫ This field carries the binary value of the offset, measured in number ofbytes between the end of the STF and the first start of a CPS Packet or in

CPS-PDU




s e d ca es t e b a y a ue o t e o set, easu ed u be obytes, between the end of the STF and the first start of a CPS-Packet or, inthe absence of a first start, to the start of the PAD field.

▫ The value 47 indicates that there is no start boundary in the CPS-PDUpayload.

▫ Values greater than 47 are not allowed.

▪ b) Sequence Number (SN) (1 bit)

▫ This bit is used to number (modulo 2) the stream of CPS-PDUs.

▫ Alternating 0,1.

▪ c) Parity (P) (1 bit)

▫ This bit is used by the receiver to detect errors in the STF. The transmittersets this bit value such that the parity over the 8-bit STF is odd.

▪ The CPS-PDU payload may carry zero, one or more (completeor partial) CPS-Packets

CPS-PDU Payload




▪ CPS-Packets have a maximum size of 64 bytes.

▪ Unused payload is filled with padding bytes coded with thevalue zero.

▪ A CPS-Packet may overlap one or two ATM cell boundaries.

▪ The overlap point where the CPS-Packet becomes partitionedmay occur anywhere in the CPS-Packet, including the CPS-Packet header.

▪ AUU = ATM-User-to-ATM-User Indication = 0

▪ SLP = Submitted (Cell) Loss Priority = 0

▪ CI = Congestion Indication = 0

Additional ATM-DATA




CI Congestion Indication 0▫ CI "1" indicates that congestion has been encountered

either before transmission or during transfer

▪ CPS Interface Data (CPS-INFO)

▫ This parameter specifies the interface data unitexchanged between the CPS and the SSCS entity.

▫ The interface data is an integral multiple of one byte.

▫ The CPS Interface Data represents a complete CPS-SDU.

▪ CPS User-to-User Indication (CPS-UUI) ▫ This parameter is transparently transported by the CPS

between peer CPS users.

ITU SDL diagrams




▪ The process of forming ATM CPS-Packets will beattempted through the use of SDL diagrams providedby the ITU.

▪ Copyright has been approved for the use of thesediagrams in this course.

SDL

▪ Due to the international nature of telecommunications and theenormous investment needed to develop large telecommunicationsystems, a structured method for recording designs has been

proposed.




p p

▪ The Specification and Description Language enables teams of peopleto work on specific aspects of a system and to communicate theirdesigns in a coordinated way.

▪ The objective of SDL is to break down a system into manageableblocks and processes. These processes can be defined using states,signals and tasks.

SDL Concepts

▪ In complex systems like PABX‟s a large number of independent events occur simultaneously, and any

specification method must be able to cater for theseconcurrent events in a simple manner




pconcurrent events in a simple manner.

▪ The concept of a process for describing concurrent events isimportant.

▪ A process may be a physical item such as a line interface, oran abstract item such as an administrative procedure.

▪ Events are the occurrences that are relevant to the behaviourin question.

▪ The following pictures illustrate how a user's requirements can havevarious interpretations (Meek and Heath, 1980).




Graphical SDL

▪ Start

▪ State




▪ Input

▪ Output

▪ Task

▪ Decision

▪ Stop

▪ Connection

ATM Packet Assembler


I.363.2 SDL Diagrams for the CPS transmitter



ATM Packet Assembler




▪ This ATM-DATA is an ATM-SDU

▪ With the addition of a cell header, 5 bytes this becomes anATM cell



UTRAN Signals & Signaling

Radio Frame Structure

▪ Radio Frame Period Tf = 10ms

▪ Frames are used for channel format control

UTRAN Signalling



▪ 15 slots, #0…#14

▪ Slots are used for power control, and synchronisation.

Tslot = 666.7s = 2560 chips

#0 #1 #2 #i #14

Tf = 10ms = 38400 chips

The Common Pilot Channel CPICH

▪ UTRA has two types of pilot channel

▫ Primary

▪ always uses primary scrambling code – choice of 512fixed channelisation code all 1‟s

UTRAN Signalling



▪ fixed channelisation code, all 1 s

▪ only one per cell/sector

▫ Secondary

▪ can have any channelisation code of length 256

▪ may use a secondary scrambling code – choice of 1 -15for each P-CPICH

▪ The P-CPICH is defined by the power in its signal and takesthe form

ceinterferen overalltheof thattobitspilotin theenergyof ratio o

c

I

E

The Primary Synchronisation Channel

▪ The Downlink Synchronisation Channel SCH, transmits the PrimarySynchronisation Code, P-SCH

▪ This is a 256 chip sequence and is the same in all cells in the network

▪ The channel is transmitted at the start of a timeslot, for the first 66.67s

UTRAN Signalling



The channel is transmitted at the start of a timeslot, for the first 66.67s

P-SCH P-SCH P-SCH

256 chips66.67s

2560 chips666.7s

Timeslot # 0 Timeslot # 1 Timeslot # 2

The Primary Synchronisation Channel

▪ The Primary Synchronisation Code, P-SCH is formed by:

▫ a = <1,1,1,1,1,1,-1,-1,1,-1,1,-1,1,-1,-1,1> 16 bits

UTRAN Signalling



▫ Cp-sch= (1+ j)×< a,a,a,-a,-a,a,-a,-a,a,a,a,-a,a,-a,a,a>

a 16×16 complex array

▪ The P-SCH is chosen for good aperiodic auto correlation

▪ No scrambling is used, & the P-SCH is already spread

▪ This is defined for all networks

The Secondary Synchronisation Channel

▪ The downlink SCH also transmits the Secondary Synchronisation Code.

▪ This is a 256 chip sequence and uses a defined bit pattern

▪ The channel is transmitted at the start of a timeslot, for the first 66.67s, atthe same time as the P-SCH

▪ The S-SCH indicates which one of 64 groups of downlink scrambling

UTRAN Signalling



g p gcodes is in use in the cell

P-SCH P-SCH P-SCH

256 chips66.67s

S-SCH S-SCH S-SCH

2560 chips666.7s



▪The Secondary Synchronisation Code can use time-switched transmit antenna diversity (TSTD) and is the only

UTRAN Signalling



y ( ) ychannel in UTRA FDD that uses TSTD.

▪ The S-SCH have identical real and imaginary components,

does not use scrambling & is already spread.

▪ This is defined for all networks


▪ The S-SCH is used to find the scrambling code group, 1 of 64and the timeslot number.

UTRAN Signalling



▪ By following the sequence of S-SCH codes the UE candetermine where it is in the matrix of group codes and itscolumn position.

Exercise

Locate which group code and time slot position the UE is at, from the following data

15,16,10,7,8,1,10,8,2,16,9,15,1,9,2,15


UTRAN Signalling



▪ The PCCPCH is transmitted continuously at constant power from eachcell and carries the Logical Broadcast Channel BCH

▪ Uses one of the 512 Primary Scrambling Codes

▪ There is only one PCCPCH per cell

The Primary Common Control Physical Channel

UTRAN Signalling



There is only one PCCPCH per cell

27 kbps, SF=256

P-SCH P-SCH P-SCH

256 chips

66.67s

S-SCH S-SCH S-SCH2560 chips

666.7s


2304 chips600s

Data (18 bits)Data (18 bits) Data (18 bits)PCCPCH

30 kbps, SF=256

The Primary Common Control Physical Channel

▪ The PCCPCH does not transmit in the first 256 chips ofeach timeslot.

▫ This provides space for the Primary and SecondarySynchronisation codes to be broadcast.

UTRAN Signalling



y

▪ Used with the Common Pilot Channel to provide forchannel estimation.

▪ To improve performance

▫ The PCCPCH can use open-loop transmission diversity.

▪ There is only one PCCPCH per cell

▪ The SCCPCH carries two different commontransport channels

▫ Forward Access Channel FACH

The Secondary Common Control Physical Channel

UTRAN Signalling



▫ Paging Channel PCH

▪ These two channels can share the same physicalchannel or use separate ones.

▪ There can be multiple SCCPCH channelsdepending on the load

▪ SCCPCH can carry Packet Switched traffic if theFACH load is low

▪ The spreading factor for SCCPCH is fixed at themaximum data rate for FACH and PCH.

The Secondary Common Control Physical Channel

UTRAN Signalling



▪ Sharing the same channel saves power as the signalmust be broadcast at a power level to cover entire cell

▪ A separate paging indicator channel PICH is used to getthe attention of the UE.

▪ The Paging channel and Paging Indicator channel areoperated together.

Paging Indicator Channel PICH

UTRAN Signalling



▪ The paging indicators use a channelisation code of 256.

▪ A PICH radio frame is of length 10ms.

▪ PICH carries 288 bits and 12 bits are left idle

Paging a mobile: PCH

▪ Paging information is carried on the paging channel,PCH, a downlink common channel.

▪ Each terminal is allocated a paging group.

UTRAN Signalling



▪ Each group listens periodically to a Paging IndicationChannel, PICH.

▪ How often a mobile must listen is governed by itspaging group. If the mobile was always on, battery lifewould be very short.

Paging Indication Channel PICH▪ Fixed rate (SF=256, 30 kbps) so that 300 bits occupy a full frame.

▪ N Paging Indicators {PI0, …, PINp-1} are transmitted in each PICH

frame, where Np=18, 36, 72, or 144.

UTRAN Signalling



b1b0

288 bits for paging indication 12 bits (undefined)

One radio frame (10 ms)

b287 b288 b299

▪ These are mapped into 300 bits of which 288 bits are defined.

PICH

▪ A „1‟ in the appropriate bit position indicates to the UE that it

should decode the next PCH

UTRAN Signalling



PICH

S-CCPCH

Paging Indicators

Paging Message

▪ Physical Random Access Channel PRACH

▫ Used to carry the Random Access Channel RACH

Common Uplink Physical Channel

UTRAN Signalling



▫ Based on Slotted ALOHA

▫ UE can start the random-access transmission atdefined time intervals

▫ There are 15 access slots per frame, spaced 5120chips apart.

▪ The RACH procedure has the following steps;

▫ UE decodes the BCH to find

▪ RACH sub-channels

Random Access Channel RACH

UTRAN Signalling



▪ Scrambling codes & signatures

▫ UE randomly selects one of the RACH signatures

▫ UE randomly selects one of the RACH sub-channels

▫ The RACH access slots are taken over 2 × 10ms frames.

▪ The RACH procedure continues with the following steps;

▫ Downlink power level is measured

▫ A preamble of 4096 chips is sent with selected signature

Random Access Channel RACH

UTRAN Signalling



p p g

▫ This is a repeating sequence of a 16 bit preamble

▫ Waits for a Acquisition Indicator channel AICH preamble

▫ When AICH is detected a 10ms or 20ms message istransmitted.

RACH Signatures ( 16 bits )

P0(n) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

P1(n) 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1

P2(n) 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1

P3(n) 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1

P4(n) 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1

P5(n) 1 -1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1

UTRAN Signalling



P5(n) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

P6(n) 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1

P7(n) 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1 1 -1

P8(n) 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1

P9(n) 1 -1 1 -1 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1

P10(n) 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1

P11(n) 1 -1 -1 1 1 -1 -1 1 -1 1 1 -1 -1 1 1 -1

P12(n) 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1

P13(n) 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 1 -1 1 -1

P14(n) 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1

Orange(dark) represents a 1 and Blue(light) represents a -1

▪ The AICH is a fixed rate, SF=256, physical channel whichcarries the Acquisition Indicators, AI.

▪ The AI corresponds to the PRACH signature chosen by the

Acquisition Indicator Channel AICH

UTRAN Signalling



▪ The AIs corresponds to the PRACH signature chosen by theUE.

▪ The phase reference for the AICH is the primary CPICH

▪ Frame duration is 20msec, slot length 5120 chips ( 2 normalTS‟s)

▪ 4096 chips are used with 1024 chips with no transmission

▪ AICH consists of 32 symbols created as follows:

▪ AICH consists of 32 symbols created as follows:


jjbAIa 15

UTRAN Signalling



jss

s jb AI a

,0

▪ a j = <a0,a1,a2…….a31>

▪ bs,j is the sequence bs,0, …. bs,31

▪ AIs can take values 1, or 0

AICH signature patterns

0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1

2 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1

3 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1

4 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1

5 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1

P0(n) 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1

P1(n) 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1 -1 1

P2(n) 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1

P3(n) 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1

P4(n) 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1

P5(n) 1 -1 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1

AIs bs,j

UTRAN Signalling



6 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1

7 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1

8 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1

9 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1

10 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1

11 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1

12 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 1 1 1 1

13 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1 1 1 -1 -1

14 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 1 1 1 1 -1 -1 -1 -1

15 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1

( )

P6(n) 1 1 -1 -1 -1 -1 1 1 1 1 -1 -1 -1 -1 1

P7(n) 1 -1 -1 1 -1 1 1 -1 1 -1 -1 1 -1 1 1

P8(n) 1 1 1 1 1 1 1 1 -1 -1 -1 -1 -1 -1 -1

P9(n) 1 -1 1 -1 1 -1 1 -1 -1 1 -1 1 -1 1 -1

P10(n) 1 1 -1 -1 1 1 -1 -1 -1 -1 1 1 -1 -1 1

P11(n) 1 -1 -1 1 1 -1 -1 1 -1 1 1 -1 -1 1 1

P12(n) 1 1 1 1 -1 -1 -1 -1 -1 -1 -1 -1 1 1 1

P13(n) 1 -1 1 -1 -1 1 -1 1 -1 1 -1 1 1 -1 1

P14(n) 1 1 -1 -1 -1 -1 1 1 -1 -1 1 1 1 1 -1

P15(n) 1 -1 -1 1 -1 1 1 -1 -1 1 1 -1 1 -1 -1

▪ a j is formed by incrementing s from 0 to 15, whichever „s‟ is used bythe UE will have a value 1, the rest will be 0


▪ Example lets choose P8(n) = AI8

115151221111001

0,15150,220,110,000

.....

.....

b AI b AI b AI b AI a


UTRAN Signalling



31,151531,2231,1131,0031

2,15152,222,112,002

1,15151,221,111,001

.....

.

.....

.....



bbbba

0,80,880 1 bb AI a As all other elements have AI = 0

▪ Therefore „a‟ will select the AICH „b‟ signature pattern corresponding to thetransmitted RACH preamble


▪ Continuing

▫ a0 to a31 has been found.

▫ „a‟ is then spread and modulated as bits with a SF of 256

UTRAN Signalling



▫ a is then spread and modulated as bits, with a SF of 256

▫ The modulation scheme halves the bit rate ie 1 symbol for 2

bits

▫ 32 × 256 ÷ 2 = 4096 chips + 1024 chips per slot = 5120

chips

▫ 5120 × 15 = 76800 chips in 20msec = 3,840,000 chips

▫ Once the UE has detected the AICH, the UE becomes part

of the active set for that cell.

PRACH

▪ The PRACH consists of two parts

▫ A preamble

▪ To initiate access▫ A message

UTRAN Signalling



A message

▪ Which can contain a request for a dedicated channel or asmall packet of user data

2 frames = 20 ms

1 PRACH slot = 2 normal timeslots 1 PRACH preamble = 4096 chips

PRACH Message


▪ The AICH indicates whether the PRACH preamble has beenreceived.

▪ If the Node-B receives the preamble it mirrors the preamblesignature back on the AICH

UTRAN Signalling



g

2 frames = 20 ms

1 PRACH slot = 1.25ms

1 PRACH preamble = 4096 chips

PRACH

AICH

Message

1 PRACH preamble = 4096 chips



UTRAN Synchronisation

Node Synchronisation

▪ Node Synchronisation is required to provide a common timingreference among different Node Bs.

▪ In the UTRAN a common timing reference among all the nodes

UTRAN Sychronisation



g gis not provided.

▪ To minimise delay and buffering on the air interface, Uu,

estimates of the timing differences between RNC and Node Bs,are made without the need to compensate for the phasedifferences between RNC's and Node B's clocks.

Synchronisation

▪ Time Alignment

RNC

CN




▪ Transport ChannelSynch

▪ Radio InterfaceSynch

Node B

Node B

UE

TDD only

Time Alignment Synchronisation

▪ The time alignment handling procedure over Iu relates to the

control of DL transmission timing in the CN nodes in order tominimise the buffer delay in the SRNC.




▪ The SRNC invokes this procedure whenever a Iu User PlaneProtocol Data Unit, PDU, is inappropriately timed.

▪ The SRNC indicates to the CN by means of a Time Alignmentcontrol frame.

▪ The delay is by a number of +/- 500µsec steps.

Time Alignment Synchronisation

CN

ACK

Timing Alignment n × 500µs

▪ A supervision timer TTA is started after the SRNC sends a timingalignment control frame

1ms




SRNC

TTA

or

NACK

If TTA expires then the SRNC will sendanother timing alignment control FrameTTA starts

Transport Channel: RNC - Node B Synchronisation

▪ RNC - Node B synchronisation is used for determining gooddownlink DL and uplink UL offset values.

▪ Measurements of node offsets can be made at the start or

restart, as well as during normal operation, to supervise thestability of the nodes.




y

▪ Procedure:-

▫ the RNC sends a DL Node Synchronisation control frame to NodeB containing the parameter t1.

▫ Upon reception of a DL Synchronisation control frame, the NodeB shall respond with UL Synchronisation Control Frame to theRNC, indicating t2 and t3, as well as t1.

Conference Call Example

▪ Assume Conference call in Dallas

▪ You are in England, the time difference is 6 hours

▪ You must control the meeting ( ie CRNC )



▪ You send out email requests for staff to meet at 6pm

▪ The meeting room can be thought of as the Node B

▪ The staff meeting in the room are the UE‟s, they aresynchronised with England time as is the Node B

▪ You ( CRNC ) need to be in your office delivering your speech at12 noon

RNC - Node B Synchronisation

SRNC 4094 4095 0 1 2

3

RFN

t1 t4




Node B

UL Node Synch

[t1=4094.3, t2=201.2, t3=202.25 ]

200 201 202 203 204 205

BFN

DL Node Synch [t1=4094.3]t2 t3

RNC - Node B Synchronisation

SRNC

▪ These two paths ( t2 - t1 ) + ( t4 - t3) give the Round Trip Delay RTD

4094 4095 0 1 23

RFN

t1 t4




Node B

200 201 202 203 204 205

BFN

t2 t3

[t1=4094.3, t2=201.2, t3=202.25, t4=2.3 ]

(t2-t1) + (t4-t3) = (t4-t1) – (t3-t2)

So 2.3 – 4094.3 = -4092 + 4096 = 4, 202.25 – 201.2 = 1.05

Therefore RTD = 4 – 1.05 = 2.95 x 10 msec = 29.5 msec

Transport Channel Synchronisation

▪ The Transport Channel (or L2) synchronisation, provides anL2 common frame numbering between UTRAN and UE.

▪ This frame number is the Connection Frame Number(CFN) and it is associated at L2 to every transport block




(CFN), and it is associated at L2 to every transport blockset, TBS and passed to L1.

▪ The CFN is not transmitted in the air interface for eachTBS, but is mapped by L1 to the cell System FrameNumber, SFN of the first radio frame used for thetransmission of the TBS (the SFN is broadcast at L1 in thebroadcast channel, BCH).

▪ The mapping is performed via the Frame_offset parameter.

▪ In case of soft handover, the Frame_offsets of the differentradio links are selected in order to have a timedtransmission of the diversity branches in the air interface.

Counters and parameters

▪ Counters and parameters as used in the different UTRANsynchronisation procedures.

▪ BFN & RFN, Node B and RNC Frame Number counter. These are theNode B/RNC common frame number counters BFN/RFN can be




Node B/RNC common frame number counters. BFN/RFN can beoptionally frequency-locked to a Network sync reference.

▫ Range: 0 to 4095 frames, 12 bits.

▪ SFN Cell System Frame Number counter. SFN is sent on BCCH onLayer 1. SFN is used for paging groups and system informationscheduling etc.

▫ In FDD SFN = BFN adjusted with T_cell.


▪ CFN, Connection Frame Number, is the frame counter used forthe L2/transport channel synchronisation between UE andUTRAN.

▫ A CFN value is associated to each TBS and it is passed together withit through the MAC-L1 SAP.




▪ CFN provides a common frame reference (at L2) to be used e.g.for synchronised transport channel reconfiguration.

▫ Range: 0 to 255 frames, 8 bits.

▫ When used for PCH the range is 0 to 4095 frames, 12 bits.


▪ Frame_offset is a radio link specific L1 parameter used to map the CFN intothe SFN that defines the specific radio frame for the transmission on the airinterface.

▪ At the L1/L2 interaction, the mapping is performed as:▫ LSB8(SFN) = CFN + Frame_offset (from L2 to L1)




▫ CFN = LSB8(SFN) - Frame_offset (from L1 to L2)

▫ The resolution of all three parameters is 1 frame.

▫ Frame_offset and CFN have the same range (8 bits, 0…255) and only the 8 leastsignificant bits of the SFN are used.

▫ The operations are modulo 256. ie 0 .. 255

▪ In the UTRAN, the Frame_offset parameter is calculated by the SRNC andprovided to the Node B.


▪ OFF The parameter OFF is calculated by the UE and reportedto the UTRAN only when the UTRAN has requested the UE tosend this parameter.




▪ In the neighbouring cell list, the UTRAN indicates for each cellif the Frame_offset is already known by the UTRAN or shall bemeasured and reported by the UE.

▫ OFF has a resolution of 1 frame and a range of 0 to 255.


▪ The DOFF (Downlink-Offset) is used to define Frame_offset and Chip_offsetduring the first radio link RL setup.

▫ The resolution should be good enough to spread load over the Iub and Node B (basedon certain load distributing algorithms).




▫ In addition it is used to spread out the location of the Pilot Symbol in order to reducethe peak downlink power since the Pilot symbol is transmitting at the slot boundary(the largest chips for one symbol is 512 chips).

▪ The SRNC sends a DOFF parameter to the UE when the new RL makes the UEchange its state to the dedicated channel state.

▫ Resolution: 512 chips; Range :0 to 599 (<80ms).


▪ The Chip_offset is used as the offset for the DL DPCH relative to thePCCPCH timing.

▫ The Chip_offset parameter has a resolution of 1 chip and a range of 0 to 38399 (<

10ms).▫ The Chip_offset parameter is calculated by the SRNC and provided to the Node B.




▪ Frame_offset + Chip_offset (sent via NBAP) are rounded together to theclosest 256 chip boundary in Node B.

▪ The 256 chip boundary is used regardless of the used spreading factor, alsowhen the spreading factor is 512. The rounded value (which is calculated inNode B) controls the DL DPCH air-interface timing.


▪ The reported Tm parameter has a resolution of 1 chip and a range of 0to 38399. The Tm shall always be sent by the UE.

▪ T_cell represents the Timing delay used for defining the start of SCH,CPICH and the downlink Scrambling Code(s) in a cell relative to BFN.




▪ The main purpose is to avoid having overlapping SCHs in different cellsbelonging to the same Node B.

▫ A SCH burst is 256 chips long.

▫ SFN in a cell is delayed T_cell relative to BFN.

▫ Resolution: 256 chips. Range: 0 .. 9 x 256 chips.


▪ t1 RNC specific frame number (RFN) that indicates the time whenRNC sends the frame through the SAP to the transport layer.

▪ t2 Node B specific frame number (BFN) that indicates the time




t2 Node B specific frame number (BFN) that indicates the timewhen Node B receives the corresponding DL synchronisationframe through the SAP from the transport layer.

▪ t3 Node B specific frame number (BFN) that indicates the timewhen Node B sends the frame through the SAP to the transportlayer.


▪ TOAWS (Time of Arrival Window Startpoint) is the window startpoint.

▫ DL data frames are expected to be received after this window startpoint.

▫ TOAWS is defined with a positive value relative Time of Arrival WindowEndpoint (TOAWE).

▫ A data frame arriving before TOAWS gives a Timing Adjustment Control frameresponse The resol tion is 1 ms the range is {0 CFN length/2 1 ms}




response. The resolution is 1 ms, the range is: {0 .. CFN length/2 –1 ms}.

▪ TOAWE (Time of Arrival Window Endpoint) is the window endpoint.

▫ DL data frames are expected to be received before this window endpoint.

▫ TOAWE is defined with a positive value relative Latest Time of Arrival (LTOA).


▪ LTOA (Latest Time of Arrival) is the latest time instant a Node Bcan receive a data frame and still be able to process it.

▫ Data frames received after LTOA can not be processed (discarded).

▫ LTOA is defined internally in Node B to be a processing time beforethe data frame is sent in air-interface.

Th i ti (T ) ld b d d i d d t




▫ The processing time (Tproc) could be vendor and service dependent.

▪ TOA (Time of Arrival) is the time difference between theTOAWE and when a data frame is received.

▫ A positive TOA means that data frames are received beforeTOAWE.

▫ A negative TOA means that data frames are received after TOAWE.


▪ Data frames that are received after TOAWE but before LTOA(TOA+TOAWE>=0) are processed by Node B.




▫ TOA has a resolution of 125 µs.

▫ TOA is positive when data frames are received before TOAWE.

▫ The range is: {0 .. +CFN length/2 –125 µs}.

▫ TOA is negative when data frames are received after TOAWE.

▫ The range is: { –125 µs .. –CFN length/2}.

Transport Channel Synchronisation ▪ The DL Data frame number is calculated with the delay in mind. So the

RNC assumes the Node B will process the data at CFN 202.

RNC

194 195 196 197 198 199

CFN

DL Data Frame 202 UL Data Frame 202




Node B

200 202

1235

204

1237 SFN

UE DL

CFN

CFN

200

1233

202 204

CFN to SFN is the frame offset

8Lsb(SFN) - CFN = 8

Receiving Window

TOA

Synchronisation Example




Frame arrives too early





Frame arrives too late

FDD Radio Interface Synchronisation

▪ FDD Radio Interface Synchronisation makes sure that the

UE gets the correct frames when receiving from severalcells.




▪ The UE measures the Timing difference between its DPCHand SFN in the target cell when doing handover and reportsit to SRNC.

▪ SRNC sends this Time difference value in two parametersFrame_offset and Chip_offset over Iub to Node B.


▪ Example: The following set of timing diagrams show an examplewith two cells connected to one UE where handover is done fromsource cell (Cell 1) to target cell (Cell 2).




BFN - Node B Frame Number

SFN - System Frame Number SFN is delayed by Tcell relative to BFN. Tcell is used to

skew cells in the same Node B in order to avoid

colliding with synchronisation channel SCH bursts.





DL - DownLink

DPCH - Dedicated Physical Channel

CFN - Connection Frame Number Chip_offset 2.6ms about a quarter of a frame





Tp1 is the propagation delay between Node B and the UE

TUE Tx The time when UE transmits on the UL DPCH

T0 is a constant of 1024 chips ( 266µs )





OFF is 2 frames

Tm is 3840chips

Tm - measured at UE at handover,

Tm has a range of 0 - 38399 chips





Frame_offset = 2

Chip_offset = 10240Tp - Propagation delay in uplink

FDD Radio Interface Synchronisation ▪ How to determine Tm at UE

▫ Select a time instant where frame N starts at DL SFN2 e.g. frame number

2058, the time from that time instant to the next frame border of DL

DPCHnom equals Tm (if these are in phase with each other, Tm is zero).▫ In the example Tm is 3840 which is 1ms




▪ How to determine OFF at UE

▫ The difference between the frame number selected for time instant (2058)and the frame number starting at instant (8) mod 256 frames equals OFF.

▪ Example:▫ (2058 – 8) mod 256 = 2, another example is (2056 – 6) mod 256 = 2.

UTRAN Signalling



UTRAN Signalling

UTRAN Signalling Procedures

▪ The signalling procedures shown in the following sections do not represent thecomplete set of possibilities.

▪ The standard specifies a set of EPs for each interface, which may be combined indifferent ways in an implementation.

UTRAN Signalling



▪ Therefore these sequences are at present merely examples of a typicalimplementation.

System Information Broadcasting

3. BCCH:System Information

1. System Information Update Request

UE Node B RNC CN

NBAPNBAP

RRCRRC

2. System Information Update Response NBAPNBAP

UTRAN Signalling



1. The RNC forwards a request to node B via a Node B application part NBAP message„System Information Update Request’.Parameters: Master/Segment Information Block(s) (System information to be broadcasted), BCCHmodification time.

4. BCCH: System Information RRCRRC

5. BCCH:System Information RRCRRC

2. The Node B confirms the ability to broadcast the information sendingSystem Information Update Response message to the RNC via NBAP.(If the Node B cannot Broadcast the information as requested, SystemInformation Update Failure is returned to the RNC).

3./4./5. The information is broadcast via BCCH, on the air interface byRRC message System Information.Parameters: Master/Segment Information Block(s) (System information).

Paging for a UE in RRC Idle Mode

▪ This example shows how paging is performed for an UE in radio resource controlRRC Idle Mode.

▪ The UE may be paged for both CS and PS services.

UTRAN Signalling



▪ Since the UE is in RRC Idle Mode, the location is only known at CN level andtherefore paging is distributed over a defined geographical area (Location Area).

Paging for a UE in RRC Idle Mode

UE Node B

1.1

Node B

2.1

RNC

1

RNC

2

CN

RANAPRANAP 1. Paging

RANAP RANAP1. Paging

UTRAN Signalling



2. PCCH : Paging Type 1

3. PCCH : Paging Type 1

1. CN initiates the paging of a UE over a LA spanning two RNCs (i.e.RNC1 and RNC2) via RANAP message Paging.Parameters: CN Domain Indicator, Permanent NAS UE Identity,Temporary UE Identity, Paging Cause.

2. Paging of UE performed by cell1 using Paging Channel PCCH PagingType 1 message.3. Paging of UE performed by cell2 using Paging Type 1 message. The UE detects page message from RNC1 (as example) and the procedure for

non-access stratum NAS signalling connection establishment follows. NAS

message transfer can now be performed.

NAS Signalling Connection Establishment

▪ This example shows establishment of a non-access stratum NASSignalling Connection.

▪ This establishment could be requested by the terminal by itself (forexample to initiate a service) or could be caused by a paging from

UTRAN Signalling



the CN.

NAS Signalling Connection Establishment

UE Serving

RNC

CN

1. RRC Connection Establishment

RRCRRC2. DCCH : Initial Direct Transfer

UTRAN Signalling



RANAP RANAP

3. Initial UE Message

1. Radio resource control RRC Connection isestablished.2. UE sends RRC Initial Direct Transfer to SRNC.Parameters: Initial NAS Message CN node indicator (it indicates the

correct CN node into which the NAS message shall be forwarded).

3. SRNC initiates signalling connection to CN, and sends theRANAP message Initial UE Message.Parameters: NAS PDU

The NAS signalling connection between UE and CN can now be usedfor NAS message transfer.

RRC Connection Establishment

▪ The following example shows establishment of a RRCconnection in a dedicated transport channel (DCH)state

UTRAN Signalling



Dedicated transport Channel DCH Establishment

Parameters: Signalling link termination,T l dd i i f i

UTRAN Signalling



1. The UE initiates set-up of an RRC connection by sending RRC messageConnection Request on common control channel CCCH.

Parameters: Initial UE Identity, Establishment cause, Initial UE Capability.

2. When a DCH is set-up, a NBAP message Radio Link Setup Request is sent to Node B.

3. Node B responses with NBAP message Radio Link Setup Response. 4.SRNC initiates set-up of Iub Data Transport bearer using ALCAP protocol.This request contains the AAL2 Binding Identity to bind the Iub DataTransport Bearer to the DCH. The request for set-up of Iub Data Transportbearer is acknowledged by Node B.

Transport layer addressing information(AAL2 address, AAL2 Binding Identity)for the Iub Data Transport Bearer.

Node B allocates resources, starts physical layer PHY receptionThe SRNC decides to use a DCH for this RRC connection, allocates aradio network temporary identity RNTI and radio resources for the RRCconnection.

DCH Establishment

Parameters: Initial UE Identity,RNTI, Capability update

Requirement, Transport FormatSet, Transport FormatCombination Set, frequency, DL

UTRAN Signalling



5./6.The Node B and SRNC establish synchronism for the Iub and IurData Transport Bearer by means of exchange of the appropriate DCHFrame Protocol frames Downlink Synchronisation and Uplink

Synchronisation.

7. Message RRC Connection Setup is sent on CCCH from SRNC toUE.

8. Message RRC Connection Setup Complete is sent on DCCH fromUE to SRNC.

Parameters: Integrity information, ciphering information.

Then Node B starts downlink DL transmission.

scrambling code (FDD only),Time Slots (TDD only), UserCodes (TDD only), Power

control information.

Soft Handover (FDD)

▪ Radio Link Addition (Branch Addition)

▫ This example shows establishment of a radio link via a Node Bcontrolled by a RNC other than the serving RNC.

▫ This is the first radio link to be established via this RNS, thus macro-

UTRAN Signalling



diversity combining/splitting with already existing radio links withinDRNS is not possible.

Radio Link Addition (Branch Addition)

If this is the first radio link

via the DRNC for this UE,a new Iur signallingconnection is established.

UTRAN Signalling



1. SRNC requests DRNC for radio resources by sending RNSAP messageRadio Link Setup Request.

Parameters: Cell id, Transport Format Set per DCH, Transport Format Combination Set,

frequency, UL scrambling code

SRNC decides to setup a radio link RL via a new cell controlled by another RNC.

This Iur signallingconnection will be usedfor all RNSAP signalling

related to this UE.

2. If requested resources are available, DRNC sends NBAP message Radio Link

Setup Request to Node B.

Parameters: Cell id, Transport Format Set per DCH, Transport Format Combination

Set, frequency, UL scrambling code.

Then Node B start the uplink UL reception. 3. Node B allocates requested resources. Successful outcome is reportedin NBAP message Radio Link Setup Response.Parameters: Signalling link termination, Transport layer addressing information (AAL2

address, AAL2 Binding Identitie(s)) for Data Transport Bearer(s).

4. DRNC sends RNSAP message Radio Link Setup Response to SRNC.Parameters: Transport layer addressing information (AAL2 address, AAL2 BindingIdentity) for Data Transport Bearer(s), Neighbouring cell information.

Radio Link Addition (Branch Addition)

UTRAN Signalling



5. SRNC initiates setup of Iur/Iub Data Transport Bearer using access link controlapplication part ALCAP protocol.

This request contains the AAL2 Binding Identity to bind the Iub DataTransport Bearer to DCH. This may be repeated for each Iur/Iub Data Transport Bearer to be setup.

6./7. Node B and SRNC establish synchronism for the Data TransportBearer(s) by means of exchange of the appropriate DCH Frame Protocol

frames Downlink Synchronisation and Uplink Synchronisation, relativeto already existing radio link(s). Then Node B starts DL transmission.

8. SRNC sends RRC message Active Set Update (Radio Link Addition)to UE on dedicated control channel DCCH.

Parameters: Update type, Cell id, DL scrambling code, Power control information,Ncell information.9. UE acknowledges with RRC message Active Set Update Complete.

Soft Handover „continued‟

▪ Radio link Deletion (Branch Deletion)

▫ This example shows deletion of a radio link belonging to a

Node B controlled by a RNC other than the serving RNC.

UTRAN Signalling



Radio link Deletion (Branch Deletion)

UTRAN Signalling



1. SRNC sends RRC message Active Set Update (Radio Link Deletion) toUE on DCCH.Parameters: Update type, Cell id.

2. UE deactivates DL reception via old branch, and acknowledges with

RRC message Active Set Update Complete. 3. SRNC requests DRNC to deallocate radio resources by sendingRNSAP message Radio Link Deletion Request.Parameters: Cell id, Transport layer addressing information.

4. DRNC sends NBAP message Radio Link Deletion Request toNode B.Parameters: Cell id, Transport layer addressing information.

SRNC decides to remove a radio link via an old cell controlled by another RNC.

Radio link Deletion (Branch Deletion)

UTRAN Signalling



5. Successful outcome is reported in NBAP message Radio Link Deletion

Response. 6. DRNC sends RNSAP message Radio Link Deletion Response to SRNC. 7. SRNC initiates release of Iur/Iub Data Transport Bearer using ALCAP protocol.Node B deallocates radio resources.

Hard Handover

▪ This example shows Inter-RNS Hard Handover

▪ A switch in the core network CN, is in a situation in which the UE is connected to two CNnodes simultaneously.

▪ The CN will end up using one Node B directly under the target RNC after the hardhandover.

▪ The Serving RNC makes the decision to perform the Hard Handover via CN.

UTRAN Signalling



▪ The SRNC also decides into which RNC (Target RNC) the Serving RNC functionality isto be relocated.

Hard Handover with switching in the CN

UTRAN Signalling



1./2. SRNC sends Relocation Required messages to both CN nodes.

Parameters: target RNC identifier, Information field transparent to the CN

node and to be transmitted to the target RNC.

Upon reception of Relocation Required message, the CN elementprepares itself for the switch and may also suspend data traffic between

UE and itself for some bearers

..

3./4. When CN is aware of preparation , CN node conveys aRelocation Request message to the target RNC to allocate resources.

Parameters: bearer ID's requested to be re-routed towards the CN node, fromwhich the Relocation Request originated.

CN indicates in the message whether it prefers point to multipoint typeof connections within CN or hard switch in CN. In this example the

latter is assumed.

Target RNC allocates necessary resources within the UTRAN to supportthe radio links to be used after completion of the Hard Handoverprocedure.

5. Target RNC and CN node establish the new Iu transport bearers for

each Radio Access Bearer related to the CN node.


UTRAN Signalling



6./7./8. The target RNC allocates RNTI and radio resources for the RRCconnection and the Radio Link, then sends the NBAP message Radio LinkSetup Request to the target Node-B.

Parameters: Cell id, Transport Format Set, Transport Format Combination Set,frequency, UL scrambling code (FDD only), Time Slots (TDD only), User Codes (TDDonly), Power control information etc.

Node B allocates resources, starts PHY reception, and responds withNBAP message Radio Link Setup Response. Target RNC initiates set-upof Iub Data Transport bearer using ALCAP protocol. This request containsthe AAL2 Binding Identity to bind the Iub Data Transport Bearer to the DCH.


UTRAN Signalling



9./10. When RNC has completed preparation phase, RelocationRequest Acknowledge is sent to the CN elements.

Parameters: transparent field to the CN that is to be transmitted to the Source RNS.

11./12. When CN is ready for the change of SRNC, CN node sends aRelocation Command to the RNC. Message contains the transparentfield provided by Target RNC.

Parameters: information provided in the Information field from the target RNC.


13 S RNC d RRC Ph sical Channel

UTRAN Signalling



13. Source RNC sends a RRC message Physical Channel

Reconfiguration to the UE.

NOTE 1: The messages used here are only one example of the various

messages which can be used to trigger a handover, to confirm it or to

indicate the handover failure.

The different possibilities are specified in the RRC specification (25.331)


14 /15 When target RNC has detected the UE Relocation Detect

UTRAN Signalling



•14./15. When target RNC has detected the UE, Relocation Detectmessage is sent to the CN nodes. The Target RNC switches theconnection towards the new Iu, when UE is detected.

•After the switch, UL traffic from Node-B's is routed via the newlyestablished Mobile data channel to the new MAC/RLC entities andfinally to the correct Iu transport bearer.

•DL data arriving from the new Iu link is routed to newly establishedradio link control RLC entities, to the medium access control MAC

and to the Nodes B.


UTRAN Signalling



16. When the UE switch from the old RL to the new RL, the source NodeB detect a failure on its RL and send a NBAP message Radio LinkFailure Indication to the source RNC.

17. When the RRC connection is established with the target RNC and

necessary radio resources have been allocated the UE sends RRCmessage Physical Channel Reconfiguration Complete to the targetRNC.

NOTE 1: The messages used here are only one example of the various

messages which can be used to trigger a handover, to confirm it or to

indicate the handover failure. The different possibilities are specified in the RRC specification (25.331)


UTRAN Signalling



18./19 After a successful switch and resource allocation at targetRNC, RNC sends Relocation Complete messages to the involved CNnodes.

Note: At any phase, before the Relocation Complete message issent, the old communication link between the CN and UE is all thetime existing and working and the procedure execution can bestopped and original configuration easily restored.

20./21. The CN node initiates the release of the Iu connections to thesource RNC by sending RANAP message Iu Release Command.


UTRAN Signalling



22. Upon reception of the release requests from the CN nodes the

old SRNC executes all necessary procedures to release all visible

UTRAN resources that were related to the RRC connection inquestion.

23./24. SRNC confirm the IU release to the CN nodes sending

the message Iu Release Complete.

umts system

Documents