g.fast – shifting the limits of copper - uknof – shifting the limits of copper bell labs,...

COPYRIGHT © 2011 ALCATEL-LUCENT. ALL RIGHTS RESERVED.

G.fast – Shifting the limits of copperBell Labs, Alcatel-Lucent - Jochen Maes 19 January, 2012 @ UKNOF


2

…

…

FTTH - the next big thing for decades …

1988

1988

1989

1988

1988


3

Reality gradual deployment of fiber deeper in network

CO deployment Fiber to the Curb / Distribution Point

Fiber to the Node / Cabinet

CO cross-connect cabinet

CO mini DSLAM

CO micro node

<5km, 24 Mb/s <1km, 100 Mb/s <200m, 1 Gb/s


4

Shifting the limits of copperPON

The Gap

1 Mbps

10 Mbps

100 Mbps

1 Gbps

50 Mbps

500 Mbps

5 Mbps

1995 2000 2005 2010 2020-2030

ADSL

ADSL2ADSL2+

VDSL(2) 8b

VDSL2 17a+ bonding

+ vectoring

+ Phantom

+ G.fast

Copper innovations allow operators to gradually build up their fiber network


5

Vectoring

0

20

40

60

80

100

120

100 200 300 400 500 600 700 800 900 1000 1100 1200

Optimal VDSL2 performance

Reduced field performance due to crosstalk

+150%

Loop length / distance (m)

Ban

dwid

th (

Mbp

s)

Near-optimal field performance with vectoring

+a

-a

+c

-c

+b

-b

+c+a

+c-a

+c

-c–c+b

–c-b

Customer modem

DSLAM

voltage

virtual pair

virtual ‘plus’

virtual ‘minus’

0100200

300400500600700

800900

1000

Sing

le li

ne

Bon

ding

Vec

tore

dPh

anto

m

Phan

tom

on

Phan

tom

Phantom modeD

ata

rate

(M

b/s)

Loop length (m)D

ata

rate

(M

b/s)


6

ITU project G.fast

• Standardizing copper physical layer aspects of FTTdp

• Initiated February 2011

• Earliest opportunity for consent: e.o. 2012• Means, earliest, standard could be approved in 2013; products could be available in 2014

• Contributions by DSL and G.hn vendors

• Main requirements• High peak data rate, e.g. 500 Mb/s at 100 m

• Suited for reverse powering

• Allowing for customer self-install

• Optimized for short loops - not a substitute for FTTCab


7

G.fast is becoming technologically feasible

Timmers et al., proceedings of Access2011

Digital computation complexity relative to ADSL in 2001

Along x-axis, cost of complexity is scaled with Moore’s law

Technologically feasible – analog will be limiting design factor


8

FTTdp architecture

CO

Passive splitter

RE CPE

G.fasttransceiver

Powerextraction

Powerinsertion

G.fasttransceiver

Powersource

RE power supply unit

60 VdcRE CPE

CPE powered network equipment

Up to 1 Gb/s net rate

8-24 users per remote network equipmentFits within TR-156 and TR-167 deployment model


9

0 50 100 150 200 250 300 350 4000

2

4

6

8

10

12

14

16

18

20

Loop length (m)

Load

pow

er (W

)

0.9 mm Cu at 20°C0.5 mm Cu at 20°C0.4 mm Cu at 20°C0.4 mm Cu at 65°C0.4 mm Al at 65°C

Reverse power feed power transfer capability

ETSI TR 102 629 : Reverse Power Feed for Remote Nodes

Source voltage 60 V, within SELV limit, current < 300 m

System design for worst case – when single long line is connected


10

0 50 100 150 200 250 300 350 4000

0.5

1

1.5

2

2.5

3

3.5

4

Tim

e (µ

s)

Loop length (m)

Arrival delayExcess delayTotal delay

Time division duplexing

• Advantages of TDD:• Reduction in analog complexity

• Increases transmission efficiency

• Halves the modulation complexity

• Starting frequency above VDSL2, for legacy compatibility

Req

uir

ed C

E in

FD

D

CE

in T

DD

TDD allows relaxing analog requirements


11

P2P and P2MP system architecture

Switchfunction

Tx filter

Rx filter

LD

amp

DAC

ADC

LD

amp

LD

amp

Control function

G.fastDSP

1 Gb/s aggregation

Tx filter

Rx filter

LD

amp

DAC

ADC

LD

amp

LD

amp

G.fastDSP

Several Gb/s aggregation

Tx filter

Rx filter

DAC

ADC

Tx filter

Rx filter

DAC

ADC

Point-to-multipoint with TDMA MACAnalogy with cable

Point-to-point

Options under consideration in ITU G.fast


12

UK Reference loops P2MP vs. P2P comparison

0 10 20 30 40 50 60 70 80 90 100-50

-45

-40

-35

-30

-25

-20

-15

-10

-5

0

Frequency (MHz)

Hlo

g (d

B)

D1H1D2H2D2H1D3H5D4H5D4H3

D1-H1 D2-H2 D2-H1 D3-H5 D4-H5 D4-H3Loop length 176.1 m 137.1 m 113.1 m 45.6 m 83.3 m 93.3 m2.2 MHz 740 642 1076 1270 1084 92212 MHz 616 541 944 1138 953 80517.7 MHz 552 493 868 1062 876 73430 MHz 428 384 714 895 720 599

UK reference loops

2 4 6 8 10 12 14 16 180

50

100

150

200

250

Node size

Net

sus

tain

able

BH

OL

(Mb/

s)

2.2 MHz12 MHz17.7 MHz30 MHz

P2P rate (vectored)

Average TDMA rate of N-1 users if 1 user is at 500 Mb/s ‘peak rate’

Start frequency

Key question: ‘up to’ or ‘guaranteed’ bandwidth


13

Transmitter Controlled Adaptive Modulation

Operator benefit:

• Robustness Self-install

• Spectral efficiency Throughput

• Green Reverse powering

With minimal implementation impact11 11

01 11

11 01

01 01

00 11

10 11

00 01

10 01

00 00

10 00

00 10

10 10

11 00

01 00

11 10

01 101 1

0 1

1 0

0 0

Traffic drivenpower control

Noise drivengrid control

Automatic adaptation to traffic load and channel conditions


14

CuGAR Copper Gigabit Access Research evaluation platform

SERVER

D6 LAB

FPGA Front end

cable

FPGA

Front end

remote connection

off-line connection

real-time connection

place and

route

Bit

load

ing

Test bed for candidate technologies


15

FTTdp - G.fast

• Continued copper access innovations facilitate operators to spread fiber investments

• G.fast is targeted for short loops up to 200 m from the last distribution point

• Being standardized in ITU-T, promising up to 1 Gb/s

• Transmitter controlled adaptive modulation is proposed, enabling high throughput, self-install and reverse powering

• P2P and P2MP system architectures under study in ITU

• TDMA offers high peak rate, but lower sustainable rate applied to UK reference loops – behaves like cable or PON

g.fast – shifting the limits of copper - uknof – shifting the limits of copper bell labs,...

Documents