TUTORIAL

We.3.A.1

ECOC 2008, 21-25 September 2008, Brussels, Belgium

TUTORIAL The future of optical networks Ken-ichi Sato Nagoya University; Furo-Cho, Chikusa-Ku, Nagoya, 464-8603 JAPAN [email protected] Abstract The transport network paradigm is moving toward NGN (Next Generations Networks) which aims at IP convergence, while architectures and technologies are diversifying. Video technologies including ultra-highdefinition TV (more than 33M pixels) continue to advance and future communication networks will become videocentric. The inefficiencies of current IP technologies, in particular the energy consumption and throughput limitations of IP routers, will become pressing problems. Harnessing the full power of light will resolve these problems and spur the creation of future video-centric networks. Extension of optical layer technologies and coordination with new transport protocols will be critical, and are discussed in detail.

Ken-ichi Sato Ken-ichi Sato is currently a professor at the graduate school of Engineering, Nagoya University. Before joining the university in April 2004, he was an executive manager of the Photonic Transport Network Laboratory at NTT. His R&D activities cover future transport network architectures, network design, OA&M (operation administration and maintenance) systems, photonic network systems including optical cross-connect/ADM and photonic IP routers, and optical transmission technologies. He has authored/co-authored more than 200 research publications in international journals and conferences and fourteen books. He holds 35 granted patents and more than 100 pending patents. He is an NTT R&D Fellow, a Fellow of the IEICE of JAPAN, and a Fellow of the IEEE.

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

Vol. 6 - 93


We.3.A.1

Extended Abstract Until ten years ago, we looked at N-ISDN and SONET/SDH as the next step technologies on which to base the development of access and core networks, respectively; agreement was universal. The next step, IP convergence, is supported by the advent and penetration of IP, new technical developments including WDM and photonic network technologies, rapid advances in access technologies, and the emergence of IP-based control protocols such as MPLS and GMPLS; all provide powerful tools for creating the next generation networks. While IP convergence is well recognized among carriers and vendors, the key issues that they must consider include the divergence of architectures and technologies. The technology alternatives are extremely varied and this allows us to develop optimized networks that match each country's or region's and/or carrier's situation. Broadband access including ADSL and FTTH is now being rapidly adopted throughout the world and, as a result, traffic is continually increasing; around 50 % every year in North America and Japan. The number of FTTH subscribers exceeded ten million in Japan and two million in USA in 2007. In order to cope with the traffic increase, optical transmission and node technologies are being extensively developed. The maximum number of WDM wavelengths per fiber exceeds one hundred, and WDM transmission systems with a channel speed of 40 Gb/s are now being introduced in some countries. The key enabling technology that enhances node throughput while simultaneously reducing node cost, is the optical path technology that exploits wavelength routing. Wavelength routing using ROADMs has recently been introduced, and a large scale deployment is being conducted in North America and Japan. GMPLS controlled OXCs (Optical Cross-connects) have also been used to create nation-wide testbed networks. Video technologies including IP TV and high-definition and ultra-high-definition TV (more than 33M pixels) are advancing and further traffic expansion is expected in the near future. Future communication networks will become video-centric. Recent advances in video technologies, which include three-dimensional TV, are described. Cutting-edge applications including e-science, all of which need enormous bandwidth, have also been conceived. The inefficiencies of the TCP/IP protocol will become more and more tangible. The power consumption and throughput limitations of IP routers are expected to limit the scale of Internet expansion in terms of bandwidth and the number of users, and the approach of using only IP convergence will not be the best to creating future bandwidth abundant networks. The details, including the limits of IP routers and protocol bottleneck, are discussed from various viewpoints. One important direction that can resolve these problems is the enhancement of photonic networking technologies. In future networks the number of wavelength paths and hence optical node throughput must be greatly increased. The enhancement of optical path capabilities and the introduction of new protocols including fast optical circuit switching will play key roles. In realizing the networks needed, wavebands (bundles of optical paths) and hierarchical optical path cross-connects (HOXCs) will become basic technologies. The presentation will elucidate the merits and issues of introducing higher order optical paths. The introduction of wavebands will substantially reduce optical switch size at cross-connects, which mitigates one of the major barriers to the implementation of large throughput optical cross-connect systems. One of the obstacles, network design complexity, will be shown to be effectively resolved by a new optical path network design that introduces a traffic demand expression in a Cartesian product space. Some of the key component technologies of the HOXCs, a new waveband MUX/DEMUX, and a waveband selective switch (WBSS) have been developed. The hierarchical optical path network will be implemented in the not so distant future when traffic volumes warrant it.

Vol. 6 - 94

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

6.3

32

1.5

100

6.3

45

45N

Europe

2

8

34

139

IP/MPLS Ethernet NG-SONET/SDH Photonic WDM

FTTH Cable Wireless DSL Satellite

and Divergence of Architectures & Technologies

Extensive choices made available through recent rapid technical advancements.

Death of the monopolies and enhanced competition that strongly drives the optimization of architectures and technologies according to the different physical/geographical and regulatory situations.

N-ISDN

SONET/SDH FR, ATM

The technology alternatives are extremely varied and this allows us to develop optimized networks that match each country's or region's or carrier's situation.

N-ISDN and SONET/SDH were recognized universally as the next step technologies on which to base the development of access and core networks.

IP Convergence,

2005

1995

Universal Standard

Analog (huge number of variations)

Japan NA

Access

Metro

Core

400

PDH

1985

c Copyright 2008, Nagoya University

Nagoya University

[email protected]

Ken-ichi Sato

September 24, 2008 ECOC 2008, Tutorial

The Future of Optical Networks

IP/MPLS Ethernet NG-SONET/SDH Photonic WDM Wireless Cable FTTH Access DSL Satellite

Metro

Core

Hub Office

Local Office Ether Sw Router VoIP Router Router ADM

ADM ADM ATM Switch Remote DSLAM Router

WDM PON EPON GPON/BPON

FTTH Single Star (Media Converter) Double Star Active Star (Ether Switch) Passive Star (PON)

Access

ADM/DXC Router

ATM Centric Architecture

ADM/DXC ADM/DXC MPLS Router Router

SDH Centric Architecture

MPLS Router

IP Centric Architecture

Core Office

Metro

DXC-based mesh or concatenated ring architectures

Integrated or separated NW approach for provisioning of legacy(voice/data) and internet services.

Core

Divergence of Architectures & Technologies

2005

c Copyright 2008, Nagoya university

6. Conclusions

Recent technical advances allow us to utilize a different set of network element. However, in order to strengthen the scalability, manageability, and costeffectiveness, new technologies are required.

Transfer Mode/Protocol Transparency Wavelength routing with OXC/ROADM(Linear/Ring) Enhanced Transparency in Electrical Level: OTN (G.709/Digital Wrapper) STM-N, 10GbEther, GbEther SDH/SONET Extensions: NG-SDH/ SONET (GFP, VCAT, LCAS) 10/100BaseT, 100BaseFX and GbEther with RPR, etc. Convergent Packet Platform MPLS-TP (formerly T-MPLS) VLAN Cross-Connect Provider Backbone Transport (Mac-in-Mac tunneling) Programmable Transport Line Module STM-1/OC-3, STM-4/OC-12, STM-16/OC-48, GbEther, Fiber Channel, FastEther, ESCON, etc.

Transport

ASON/GMPLS Automated Connection Provisioning

Next Generation Network Fixed Mobile Convergence Using 3GPP IMS and SIP

Control/Architecture

2015 Integration Technologies

5. Hierarchical Optical Path Networks and Technologies

4. Bottleneck of Present IP-based Technologies Posed to Future Networks

3. Advances in video Technologies

2. Access Network Development

1. Overview of Transport Network Evolution

Outline


We.3.A.1

Vol. 6 - 95

Headend CMTS / QAM

HFC

0

10

20

1.5 M type (2000.12-)

8 M type (2001.12-)

12 M type (2002.11-)

24 M type (2003.7-)

40 M type (2003.12-)

47 M type (2004.8-)

30

Source: Cisco

40

50

60

70

Vol. 6 - 96


Source: M. Ninomiya, presented at BWFA, 2005

Transmission Loss between User’s House and Service Provider’s Office (dB)

4

8

12

16

20

24

28

32

36

40

44

48

Down-Stream Bit Rate of High-Speed ADSL


6. Conclusions

ADSL

ONU (6 - 12 HH)

Fiber node (500 - 1500 HH)

500ft

Coax (6MHz DS channels, N*6MHz in future)

Cable Modem + set-top + home GW

Today ~700- 800 MHz DS / ~40 MHz US In the future ~900MHz DS / ~80MHz US

Fiber Node (50 HH)

ONT + set-top + home GW

DSL modem + set-top + home GW

DSL modem + set-top + home GW

DSL modem

Customer Premise

VDSL2 ADSL2+

ADSL2+ / VDSL2

Up to 5000 feet of copper

BPON / GPON + optional RF overlay

DSLAM (100s HH)

Up to 12000 feet of copper

Curb

0

1000

2000

3000

4000

5000

Cumulative Percentage of Users against Loop Length 100 90 80 70 60 50 40 30 20 10 0 6000 7000

Loop Length between Subscriber and SP Building (m)


" ! """

0

5

10

15

20

25

Mb/s

Measured Downstream Speed for 26 Mb/s ADSL


Copper

Node (Neighborhood)

Passive optical splitters / combiners in outside plant

Aggregation

DSLAM or DSL from DLC

RT

DLC: Digital Loop Carrier, DSLAM: Digital Subscriber Line Access Multiplexer, RT: Remote Terminal, VDSL: Very high data rate Digital Subscriber Line, GW: Gate Way, OLT: Optical Line Termination, ONT: Optical Network Terminal, ONU: Optical Network Unit, HFC: Hybrid Fiber-Coaxial

OLT

FTTH


Down-Stream Bit Rate on Ideal Condition (Mbps)

Fiber

Aggregation

FTTC


Measured Downstream Speed


Aggregation

Current telco DSLAM / ATM aggregation broadband

CO

Access Architecture in USA

FTTN



Outline


We.3.A.1

Ratio (%)


20-25

10-15

4-10

2-4

1-2

0-1

Measured Downstream Speed Mbps

15-20

11

Number of FTTH Subscribers in the World


Source: Acca Networks

Measured Downstream Speed Mbps

0

50

75

25

Cumulative (%)

100

#!"

FTTx

will be Needed

"""!

Population Density'(/km2)

$

$

10

Vol. 6 - 97

T. Yamada, FOE 2006, 2006.


0

0

0.5

0.1

1.5

2

1

Fiscal Year

Start of BFlet’s Service

0.2

0.3

0.4

2.5

NTT East

12

All Areas Urban Areas with more than 100,000 Inhabitants

NTT West

Fiscal Year

Start of BFlet’s Service

Source: NTT West Data Book

20

40

60

80

100

Investment in Optical Access Network Development and Opticalization to the Distribution Point by NTT

"%

%

#"

"

$"&

Cheaper FTTx Deployment Cost

Source: ITIF (The Information & Innovation Foundation), May 2008.

be Applied

xDSL may


Investment (Trillion Yen)

Average Local Loop Length'(Km)

Average Loop Length and Population Density

Cumulative Investment (Trillion Yen)

978-1-4244-2229-6/08/$25.00 ©2008 IEEE Fiber Installation Ratio (%)

Distribution of Downstream Speed for 26 Mb/s ADSL (10,000 data)


We.3.A.1

30

2003.3

2005.3

All SPs NTT East + NTT West

USA（Pass）

USA

Japan

Vol. 6 - 98


Source: MIC, Japan and FTTH Council

10

100

1,000

10,000

Year

2007.3

8.77

12.2

2009.3

2011.3

as of April 2008

2.9 mil

11.8 mil

12.2 mil

Target Reset on Nov. 11, 2007

Number of FTTH Subscribers


Source: M. Ii, OFC/NFOEC 2006, March, 2006

0

10

20

FTTH Subscribers (million)

Number of Subscribers（thousand）

Target Set on Nov. 8, 2004

Number of NTT's FTTH Subscribers toward 20 Million

2007

2008

2009

2010


2011

NTT

Situation varies in each country Four countries make up more than 80 %

Strong public support. City- or Municipality-wide optical networks. Private initiative. FastWeb. Power utility company, Hafslund (Oslo municipality owns the dominant share.) Very high population density. Competition between cable pushes higher speed service. Electrical power company plays a dominant role.

50,000 100,000 150,000 200,000 250,000 300,000 350,000

FTTH Subscribers, End 2007 0

Source: FTTH-Council Europe/IDATE 2008

Czech Republic

Poland

Slovenia

Finland

France

Denmark

Netherland

Norway

Italy

Sweden

2006

NTT

Forecast Household Pass Forecast Homes Connected

FTTH in Europe

2005


Source: IGI, 2007

0

2,000,000

4,000,000

6,000,000

8,000,000

10,000,000

12,000,000

14,000,000

16,000,000

16,000,000

20,000,000

Number of Verizon's FTTH Subscribers


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

Vol. 6 - 99


6. Conclusions






Outline

Source: The Broadband Fact Book, Internet Innovation Alliance


1.4

1.6

1.8

2.0

2.2

Annual Growth

Internet Traffic Growth as Seen by AT&T

1998

1999

2000

2001

2002

2003

2004

2005

2006

2007


# !% $ $!

% $ $!!"

% $ $!

$

"

% $ $!

HDTV-1080p/24 vs. 4K-SHD Digital Cinema with 8M pixels vs. Super Hi-Vision (EHRI-3) with 32M pixels

Total Broadband Subscriber Traffic of 6 ISPs Exchanged Peak Traffic among all ISPs at Major IXs Exchanged Traffic among all ISPs at Major IXs Exchanged Traffic among 6 ISPs at Major IXs

Total Traffic Downloaded by Broadband Users in Japan

1.4 times increase per year

Resolution comparison


Source: Ministry of Internal Affairs and Communications

1997

Monthly average of daily average traffic (Gb/s)

Internet Traffic Growth in Japan


We.3.A.1

EHRI-0

3840x2160

EHRI-1 5760x3240

EHRI-2

Vol. 6 - 100

30 40 50

100

Level 3 (720P)

200

Level 1 (2kP)

300 400 500

Level 2 (1080P)

Screen Size (Inch)

Level 4 (480P)

5: Very Good 4: Good 3: Fair 2: Poor 1: Very Poor


1

2

3

4

5

MOS

Screen Size and Available Quality


23

7680x4320

EHRI-3

Courtesy of Dr. Yoshihiro Fujita, Presented at NAB2006 Broadcast Engineering Conference

• Cable transmission of LSDI

ITU-T SG9

No. of Pixels 1920x1080

Hierarchies

• LSDI : Large Screen Digital Imagery, TG6/9 – Draft New Recommendation ‘Parameter values for an expanded hierarchy of LSDI image formats for production and international programe exchange’ • EHRI: Extremely High-Resolution Imagery, WP6J • Multi-channel sound system for LSDI, TG6/9

ITU-R SG6

Standardization in ITU

0

1000

2000

3000

4000

Polaroid film

35mm film

Printing Medicine Laser Printer 60mm film

0

Still Images

HDT V108 0P 720P

35mm Movie

60 Temporal resolution (Frame or Field/sec)22


150 PDP presented at 2008 International CES Expected to be commercially available in 2009

2006, 102 inch, Samsung and LG 2007, 103-inch plasma TV, Panasonic (commercially available) 2007, 108-inch LCD TV, Sharp (commercially available) 2007, 110-inch Projection TV, Victor (commercially available) 2008, 150-inch plasma TV, Panasonic (presented at 2008 CES, expected to be available in 2008)

Large Screen TV, 100 is already available, 150 has already been presented

Screen Size is Getting Bigger

24

Standard TV(480i)

1.5 Gbps

HDTV(1080i)

72 Gbps (24 Gbps; tentative)

Super Hi-Vision (4,320p)

Television

Legacy media

6 Gbps

4k Digital Cinema with 8M4096x2160 pixels

Motion Picture

Horizontal resolution is used for motion pictures. 4k motion picture means= 2000 scanning lines 2k motion picture means = 1000 scanning lines (@HDTV)

2K

4K

8K

Spatial resolution lines

Comparison of Resolution


We.3.A.1

32"

26"

37"

40-43" 37"

LCD Plasma Panel

50"

-20 ~ -30%/year

year

40"

Vol. 6 - 101

0.4 M


MPEG-2

Original

2.1 M

HDTV

Original

8.8 M

4k Cinema

8k Super Hi-Vision 33.2 M

JPEG2000

Original

MPEG-4 (H.264) MPEG-2 MPEG-4 (H.264)

Original

Standard TV

Spatial Resolution (pixel/frame)

1M

10M

100M

1G

10G

100G

Bit Rate of Different Video Format


Bit Rate (bit/s)

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

White Paper, Ministry of Internal Affairs and Communications, 2007

$

Price Trend of LCD and Plasma Panel for High Definition TV (in Japan)

36.0

40.2

23.8


http://www.dlp.com/cinema/ http://www.christiedigital.com/AMEN/EntertainmentSolutions/DigitallyReleasedMovies/ http://www.aboutprojectors.com/Sony-SRX-R220-projector.html http://www.ntt-west.co.jp/news/0703/070316a_1.html

Trial presentation started in October 2005 in Japan. Twelve films have been shown in Japan at six theatres as of May 2007.

4K Cinema (40962160)

About 400 theatres have adopted DLP Cinema as of December 2005. That number now exceeds 3,000. The adoption rate should rise; 5,000 of the 100 thousand screens around the world are renewed every year.

28

Christie CP2000 WORLD'S MOST DEPLOYED DIGITAL CINEMA PROJECTOR

DLP Cinema developed by TIBarco, Christie, and NEC are now licensed

SONY "4K" DIGITAL CINEMA PROJECTOR, SRX-R220, • 4096 x 2160 pixel resolution • 18000 ANSI Lumens • 2000:1 Contrast Ratio

36.8

35.8

27.4

The first film presented by 2K Cinema was "Star Wars Episode I" directed by George Lucas, shown in June 1999 in USA. Digital Cinema Projector System that uses DLP (Digital Light Processing) Cinema developed by TI. More than 190 movies have been shot using DLP Cinema.

2K Cinema (20481080)

Digital Cinema

Source: Japan Electronics and Information Technology Industries Association

40.4

42.3

17.3


Shipping of LC-TV (16:9) in Japan

Screen Size of TV (Japan)


We.3.A.1

Distribution Center

Theater

NTT

Itabashi MusashiMurayama

Theater

NTT

(NTT East)

Roppongi Encryption, Compression, Odaiba

Film Studio

Film Studio

Burbank

USA

! 01/*&$3/1 -"+& )3 0/22)#,& 3/ 1&",)9& )32 &731&-&,8 ()'( 7 0)7&,2 1&2/,43)/. 6()$( )2 23)04,"3&% #8 3(& )')3", ).&-" .)3)"3)5&2 29

Digital Cinema Server

MPEG-2

Original

0.4 M

2.1 M

HDTV

Original

8.8 M

4k Cinema

8k Super Hi-Vision 33.2 M

JPEG2000

Original

MPEG-4 (H.264) MPEG-2 MPEG-4 (H.264)

Original

Standard TV

Spatial Resolution (pixel/frame)

1M

10M

100M

1G

10G

100G

Bit Rate of Different Video Format


Film Studio

Film Studio

1 Gbps

Los Angels

Seattle

Color Control, Key Control,

Quality Control

Theater

http://www.ntt-west.co.jp/news/0703/070316a_2.html http://www.watch.impress.co.jp/av/docs/20061025/sony.htm http://www.sony.jp/CorporateCruise/Press/200704/07-0425/

Bit Rate (bit/s)

1 Gbps

Yokosuka

Distribution Distribution Center Center

Tokyo

(NTT West)

Tkatsuki Nanba

Key Distribution

Key Center

Osaka

Tokyo Japan 200 Mbps

Trial System Configuration of 4K Pure Cinema Distribution

1960

1970

1980

1990

Penetration of TV Households

2010

2020

15 years to reach 80% penetration

HD enabled TV sets

4K TV 8K TV

半球状 Hemispherical Screen スクリーン

150mm 150mm

視線高調整台視線高調整台

Observer 観察観者察者

広視野画像投影系 Wide Vision 850mm 広視野画像投影系 850mm Projection System

Courtesy of Dr. Yoshihiro Fujita, Presented at the 11th Optical Technology Symposium, March 4, 2008, AIST, Tokyo.

Measurement of Body Inclination (movement of Center of Gravity)

Tilt still picture

Measurement of Psychologically Induced Effect Using Hemispherical Screen

Sight Angle and Degree of Psychological Effect (Induced Effects)


2000

Colour TV

Source: P. Revillon, European HDTV Conference, Luxemburg, 2005

0% 1950

80% penetration

21 years to reach 80% penetration

Black & White TV

10% 25 years to reach

20%

30%

40%

50%

60%

70%

80%

90%

100%

The Third Generation of TV Sets in Europe will be Flat and HD


We.3.A.1

Vol. 6 - 102

Hi-Vision

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

*

*

1980

*

*

1990

*

*

*

2000

1990

Expo (2005)

Broadcast format (2008)

*

2010 2020

(2020)

SHV broadcast * (2025) * First exhibit * broadcast Experimental (2002) (2015) Demo.*at Aichi Home-use* AV machinery

Study start (1995)

*

2000

Experimental broadcast Study start (1989) PDP practical model (1964) First exhibit (1996) (1969) Satellite digital broadcast (2000) Broadcast format 1125lines (1976) Terrestrial digital broadcast Demo. at Tsukuba (2003) Expo (1985)

*

1970

Vol. 6 - 103

UHDTV

SMPTE 2036

16:9

16:9

Aspect ratio

3840x2160

7680x4320

3840x2160

7680x4320

Pixel count

50,60*

24*,25,30*, 50,60*

Frame frequency

Sampling Bit depth

Rec.709

Rec.1361

Colorimetry

*divided by 1.001 are also specified

4:4:4, 10, 12 4:2:2, 4:2:0

4:4:4, 10, 12 4:2:2, 4:2:0

structure


Expanded LSDI

Application

ITU Rec. 1769

Standard

Ultra high definition television – Image parameter values for program production

SMPTE 2036 (2007)

Parameter values for an expanded hierarchy of LSDI image formats for production and international programm exchange

ITU-R Recommendation BT.1769 (2006)

Tiered image formats based on 1920x1080

Extremely high resolution imagery

ITU-R Recommendation BT.1201 (1995-2004)

Standardization of UHDTV

Courtesy of Dr. Yoshihiro Fujita, Presented at NAB2006 Broadcast Engineering Conference

Super Hi-Vision

Road Map of Super Hi-Vision

Free Viewpoint TV


2.5" 33M-pixel CMOS imager

・Between camera and CCU: Optical wavelength multiple in one HD camera cable, 1,000 m ・Weight of head: 40 kg, Power consumption: 150 W

・5x power zoom lens

・1.25" 8M-pixel CMOS, Four-chip imaging

・Weight of head: 80 kg, Power consumption: 600 W

・Between camera and CCU: HD-SDI (optical, coaxial) x 16, 300 m

・50 mm single-focus lens


36

Ref. Masayuki Tanimoto, "FTV (Free viewpoint TV) and Creation of Ray-Based Image Engineering", ECTI Transaction on Electrical Engineering, Electronics and Communications, Vol. 6, No. 1, pp.3-14, February 2008.

2.5" 8M-pixel CCD, Four-chip imaging


We.3.A.1

Vol. 6 - 104


39

Prof. Masayuki Tanimoto, Nagoya University, Japan, [email protected]

2008/07FDIS (Final Draft International Standard) of MVC issued and MVC finalized

2008/05Test sequences for 3DV determined

2008/013DV (3D Video) that targets 3D display application of FTV started as the first focus of FTV

2007/04： MPEG FTV started for standardization of entire FTV

2006/07： Activity moves from MPEG to JVT

2006/01： Proposals evaluated

2005/07： Call for Proposals on MVC issued

2005/04： Test sequences for MVC determined

2005/01： Start of MVC decided for compression of FTV

2004/10： Call for Evidence on MVC (Multi-View Video Coding) issued

2004/03： FTV got strong support from industry in response to the call

2003/10： Call for Comments on 3DAV issued

2002/07： FTV proposed to MPEG 3DAV

FTV Standardization Activity


37

captured scene


Circular alignment of 100 HDTV cameras

100-Camera Capture System of FTV

Outline


6. Conclusions








38

Ray Reproducing 360-Degree Display: The SeeLinder


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE Ethernet FR ATM PPP FDDI OTN SDH Satellite Wireless Coax Fiber Pair Cable

IP

TELNET SMTP SNMP POP FTP HTTP TCP UDP

DNS

Fiscal Year (end of March)

Vol. 6 - 105


Source: Ministry of Internal Affairs and Communications, Japan, March 2007.

108 kWh/year

4.8 %

% of total electricity sold in Japan

Carrier (Fixed)

User Communication Terminal(Fixed)

Carrier (Mobile)

PC (Business)

Server (Middle Range)

Router

Hub

5.8 %

Prediction of Electricity Consumption on IT (Japan)


Throughput Bottleneck of IP Transport

Energy Bottleneck of Internet

Bottleneck of TCP/IP-based Internet

2000

2005

2010

10 TWh

Fiscal Year

Target for Reduction

5 %

100 GW

1 Mbps


World Population: 6 Billion Broadband Access Take Rate: 33 %

Percent of World Power Supply

Power Consumption

Access Rate

58 %

1 TW

10 Mbps

The energy bottleneck could eventually limit network growth.

Power Consumption will Limit Internet Growth


Source: NTT West

Power Dissipation by NTT Group in Fiscal Year 2003 was 7.4 Tera Wh, One Percent of the Total Electrical Power Purchased in Japan.

1990

3.4 TWh

Electrical Power Purchased by NTT Group

Consumption of Electricity by Telecommunication Carriers is Rapidly Increasing


We.3.A.1

¥ 100,000,000 ¥ 150,000,000

Vol. 6 - 106

1

0.1 W


2

0.5 W

1W

10 W

10 Device Cost

5W

20 W

100

( Yen ) ($)

Device Cost Dominant

47

Electricity Cost Dominant

Device Lifetime = 5 years Total Power = 2.2 x Device Power Consumption Electricity Charge = 24 yen/kWh

Device Cost and Lifetime Electricity Cost


Source: HP Homepage

Power consumption: 1 kW Electricity charge:¥ 25,000/kVA/Month (including air conditioning)

DLS380 G5/2CPU (maximum configuration) Retail Price ¥ 525,000 Electricity charge ¥ 600,000 (2 years)

Floor rental fee: ¥ 12,000/m2 /year x 700 m2 (in downtown area, it's more expensive) Electricity price: ¥ 25,000/kVA/Month (including air conditioning) (number of racks: 250, number of servers: 550)

Floor rental fee/year Electricity charge/year

0.5

0.25 1

1 0 46

Flets-ISDN Flets-ADSL 64 kbps 12 Mbps Copper


Source: NTT Group CSR Report 2005

0

25

50

75

100

125

B-Flets 100 Mbps Fiber (FTTH)

0.1

1

10

100

1000

10000

48

Application of Optical Technologies Contribute to Lowering Energy Used not only for Core/Metro Networking but also for Access Networking

Access Technology and Environmental Load


2.23

0.53

Device

Package

PSU Efficiency

Equipment

Room Cooling System Power Supply Efficiency UPS+PDU Blowers

Center

Power Consumption of Each Device will more than DOUBLE.

Environmental Load (kg-CO2)

In running dater center, electricity charge exceeds floor rental fee or equipment price. Significant part of the ownership cost is now electricity cost.

Normalized Ownership Cost (Total cost/Device Cost)

Device Power Consumption will Multiply

Environmental Efficiency (kbps/kg-CO2)

Ownership Cost is Changing from Hardware to Electricity Cost


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE


Throughput Bottleneck of IP Transport

Energy Bottleneck of Internet

Ethernet FR ATM PPP FDDI OTN SDH Satellite Wireless Coax Fiber Pair Cable

IP

TELNET SMTP SNMP POP FTP HTTP TCP UDP

DNS

Bottleneck of TCP/IP-based Internet


Source: H. Shinohara, ITU Symposium on ICTs and Climate Change, April, 2008, Kyoto.

CO2 Reduction Attained by PON Architecture

Source: Mr. Michael Scharf, University of Stuttgart, Presented at ECOC 2007 Workshop "Future Internet Design,” Berlin, September 16, 2007.


C.Lange, M. Braune, and N. Gieschen,” OFC/NFOEC 2008, JWA105, February, 2008.

FTTB Nnetworks Consume More Energy than FTTH Networks


We.3.A.1

Vol. 6 - 107

Entities to be controlled

(Local Arrangement)

Queuing Mgt.

Local/Global Arrangement

Priority queueing, Weighted fair queueing

Remarks

Traffic Redirecting

Vol. 6 - 108


Regarding traffic engineering, the ability to provide maximum services is the key.

IPv4 DSCP byte/IPv6 trafficclass byte, Aggregate perDiffServ Class (CoS) hop behavior -In band signaling along data path -Evaluation of QoS demand and Flow per flow feedback by transit flow router Flow Router -TCP jump start depending on (dynamic) available bandwidth -Signaling and resource handling Traffic (reservation) In Band Flow Engineering -Problems stemming from RSVP Signaling (connection (Connection requiring periodic refreshing of (IETF IntServ) oriented) /Flow the reservations of all flows (not applicable to large networks) Oriented; More Global -Connection admission control NGN based on static information Connection Arrange -Class base QoS control (like ment) (SIP) DiffServe) -Constraint-based routing Class -No capacity admission control -3 bit of MPLS QoS header (Forwarding MPLS per LSP -Network bandwidth Equivalent Class) partitioning with Layer2/ MPLS connection by -Time required to Out Band Signaling Connection connection establish/release connection Queuing mgt. and traffic engineering will compliment each other. 55

Buffer Management Packet/cell

Mechanism

IP Traffic Management Mechanisms

Source: Mr. Michael Scharf, University of Stuttgart, Presented at ECOC 2007 Workshop "Future Internet Design,” Berlin, September 16, 2007.

Outline


6. Conclusions







ITU-T Recommendation Y.1541, "Network performance objectives for IP-based services," February 2006.

The values of all objectives are provisional and they need not be met by networks until they are revised (up or down) based on real operational experience.

Class 6, 7: QoS class for video services

Table 3/Y.1541 Provisional IP network QoS class definitions and network performance objectives

IPTD: IP Packet Transfer Delay IPDV: IP packet Delay Variation IPLR: IP packet Loss Ratio IPER: P packet Error Ratio IPRR: IP Packet Reordering Ratio

Class 0, 1: For real-time service including VoIP

Table 1/Y.1541 IP network QoS class definitions and network performance objectives

Provisional QoS Class for IP Video Transmission


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

R. Tucker, "Optical Packet-Switched WDM Networks: a Cost and Energy Perspective," OFC/NFOEC 2008, OMG1, San Diego, February 2008.

NTT

AT&T

KDDI

ALU

15

16

17

18

OECC2008, PDP-6

OFC2008, PDP7

OFC2008, PDP3

OFC2008, PDP2

OFC2008, PDP1

ECOC2007, PD1.7

OFC2007, PDP22

OFC2007, PDP20

ECOC2006, Th4.1.3

Conference

Vol. 6 - 109


Source: NTT Network Innovation Laboratories

1, 2, 4: real time transmission experiments Others: off-line experiment

NTT

CoreOptics

14

20

NTT

13

Melbourune Univ.

Lucent

12

19

Organization

Ref.

1

0.1

16.4

1

0.8

3

1

20.4

1

Capacity (Tb/s)

2100 (DSF)

1000

2550 (DMF)

1000

640

1300

2375 (SMF)

240

2000(NZDF)

Distance (km)

111

107

111

121.9

114

100

111

111

107

Line Rate (Gb/s)

10

5

ROADM node

PDM-OFDMQPSK

PDM-OFDMQPSK

PDM-QPSK

PDM OFDM-8QAM

PDM-RZ-8PSK

OFDM-DQPSK

PDM-RZ-QPSK

CSRZ-DQPSK

RZ-DQPSK

Modulation

Challenges on 100-Gb/s long-haul DWDM Transmission for 100GbE Transport


http://www.caida.org/outreach/presentations/nanog9806/

Internet Hop Distance Distribution

~ 1 Tb/s Electrical Router

Optical Processing

Electrical Processing

Optical Path

Ethernet, etc.

IP

~ 10 Tb/s Photonic Router

IP Layer Cut-through on Optical Path

L2/L1

L2

SDH/SONET

L3

Optical Fast Path/Circuit Switching

New Protocol

~10 years ~ 100 Tb/s Photonic Router

~5 years

(1) Introduction of HO-OP (Waveband) (2) Introduction of Optical Fast Circuit/Path Switching (Burst Switching)

HO-OP(Waveband)

Optical Path

Ethernet, etc.

L2 L2/L1

IP

Future Networks L3

New IP Networksafter 2004


L2/L1

L2

L3

Traditional IP Networks

Direction towards Network Throughput Expansion and Total Power Reduction


Photonic MPLS

Traffic Jam at Node

Photonic Network

Existing Network

WDM + (electrical) IP Router

Wavelength Routing on Photonic Superhighway


We.3.A.1

IP (Two way signaling)

Burst

!!"% ""!"" # ""! #"!

~3/8

Electrical Router CutThrough on Optical Paths

Vol. 6 - 110


~1/20

Introduction of Optical Fast Circuit Switching and Hierarchical Optical Paths

Single Shelf Router Throughput: 0.64 Tb/s (WAN Count), Total Throughput: 18.6 Tb/s (WAN Count) Node Terminating Traffic Ratio: 30 %, Service Traffic Ratio for IP/OCS: 1/4

Power Consumption

Electrical Router

BXC

(M+(L+1)+M)/ Mx(L+1)

L hop


L hop

Band Utilization =

M: # of Optical Paths per Waveband

Multi-Layer Optical Path Network

Single Layer Optical Path Network

Example: M=8, L=4, =0.9 R=0.58, M=16, L=4, =0.9 R=0.51, M=16, L=6, =0.9 R=0.39

Ratio of total cross-connect switch port (Multi-layerSingle Layer) =

Comparison of Cross-Connect Switch Port Number


Node cost reduction

•Reduction of necessary number of switch ports •Reduction of switch size

WXC

Optical Fiber

WaveBand

…

Reduction in Power Consumption


Optical Fast PS/CS

% !! " #"!"!! " ##"$! "#" ! !! %

IP

New Protocol

Merits • Large Capacity Optical Path is Realized by Multiplexing Multiple Optical Paths •Routing is done as a WaveBand; cut through of wavelength level routing processing

…

&!!'

Ethernet FRATM PPP FDDIOTN SDH Wireless Satellite Coax Fiber Pair Cable

IP

TCP UDP

DNS TELNET SNMP SMTP FTP HTTP POP

WaveBand Grouped optical path to be treated as a higher order path

Hierarchical Optical Path Networks

… …

Introduction of Fast Optical Path/Circuit Switching and OBS - Solving IP Bottleneck? -


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE Waveband multiplexers/demultiplexers

Waveband conversion technologies

Vol. 6 - 111


Several algorithms based on heuristics relaxation have been developed; they are categorized as follows.

The number of binary variables in the combinatorial optimization problem explosively increases with network size. This characteristic makes the problem computationally impossible to accurately solve for large networks.

The design aims at minimizing cost functions subject to waveband and wavelength continuity constraint.


More detailed explanations are given in Ref [1].

Cluster-search method in a source-destination Cartesian product space [21], [48]

Relaxation-based methods [41], [44]

On demand waveband assignment [42], [43]

Waveband tunnel construction first [45]

Grooming of wavelength paths having common source/ destination or partially shared routes [46], [47].


WXC: Wavelength path cross-connect BXC: Waveband cross-connect

Network control protocol (Extension of ASON/GMPLS)

The waveband routing and waveband assignment problem of multi-granular optical networks is a generalization of the single-granular optical network design problem.

BXC

WXC

Multi-layer optical path cross-connect switch architectures

Hierarchical Optical Path Network Design (2)

,#&"( . . .

Hierarchical OP network design algorithms

Hierarchical Optical Path Network Technologies

Hierarchical Optical Path Network Design (1)

th f WB Pa umber o

op N

H Average


Ref. [2, 34]

) %%%)"%') *#'

) %%))%)"$*#'%(+ )&%')( $)$)+%'!( / *") "-' $"-'0

Reduction in Switch Port Number


We.3.A.1

S-space

Node pair Point Traffic Demand Value

D-space

Normalized Network Cost

Vol. 6 - 112

0

) BPHT : X.Cao et al., IEEE J-SAC ,2003

1 2 3 4 5 6 7 Average Optical Path Demand between Nodes

WB Bandwidth=8

Single Layer Optical Path NW

end-to-end BPHT () Proposed

Network Cost Comparison


0

0.5

1.0

1.5

2.0

Ref. [33, 34]

Multilayer Optical NW/ Single Layer Optical NW


Ref. [33, 34]

8

The proposed method searches for the clusters in the space and places them into wavebands.

A Cartesian product space is introduces, in which nearby traffic demands are classfied as clusters of points.

Original Network

Traffic Demand

S-D Cartesian product space

Introduction of S-D Cartesian Product Space

d3

D

S s3

d1

s1

d2

0

td

Ref. [35]


D-loop

…….. Loop number

td

Loop number

S-loop = ls

Accommodate remaining wavelength paths

S-loop

backup

primary path Wavelength demands

3. Adapt wavelength paths to waveband chain (= concatenated loops)

Establish primary and backup waveband paths routing by Suurballe’s algorithm

2. Establish waveband chains

1. Search for S-D loop pairs

Setup Waveband Chains between S-D loop pairs

Proposed Algorithm for Protected Network Design

Ref. [33, 34] c Copyright 2008, Nagoya university

Remaining Wavelength Path Accommodation

s2

s-d Cartesian product space

3. Add the wavelength paths in the cluster

routing of the waveband path by Dijkstra’s Algorithm

2. Establish waveband path

find a set of wavelength paths to be accommodated within a waveband

1. Search for clusters

Waveband Setup

Network Design Algorithm


We.3.A.1

978-1-4244-2229-6/08/$25.00 ©2008 IEEE

Ref. [35]

Vol. 6 - 113

Hierarchical

Single layer


Ref. [36, 37]

9x9

Numerical Results (2)


5x5

Numerical Results

Hierarchical with protection Single layer w/o protection

topology

N

of fiber / waveband

N


WXC: Wavelength path cross-connect BXC: Waveband cross-connect

Waveband conversion technologies

BXC

WXC

Multi-layer optical path cross-connect switch architectures

Waveband multiplexers/demultiplexers

Network control protocol (Extension of ASON/GMPLS)

Hierarchical OP network design algorithms

Hierarchical Optical Path Network Technologies


Ref. [35]

–8 wavebands per fiber –8 wavelengths per waveband

Capacity

–randomly distributed wavelength paths

demands

wavelength converter

Traffic

NO

–9x9 polygrid (81 nodes, N=9)

Physical

Simulation Parameters


We.3.A.1

Output Fiber

Output Slab Waveguide

Vol. 6 - 114

AWG 2

• WB MUX/DEMUX can be developed using concatenated conventional AWGs; they are connected via waveguides that offer special connection rules.


AWG 1

Newly Developed WB MUX/DEMUX using Concatenated AWGs


• When input optical fiber is shifted by one port, the output port of each channel will shift by one port.

• Input optical channels are demultiplexed and output from each output port channel by channel

Input Slab Waveguide

Silicon Substrate

AWG

Output Waveguides

Waveguide Cross-Section View

Input Waveguides

Input Fiber

Si Substrate

Silica Layer

Waveguide Core

Arrayed-Waveguide Grating (AWG)

2, 3,

1,

WB k

WB3

~ , l l+1, l+2, l+3, ~ , 2l 2l+1, 2l+2, 2l+3, ~ , 3l

....

..... .... 2, 3,

....

WB2

~ , l l+1, l+2, l+3, ~ , 2l 2l+1, 2l+2, 2l+3, ~ , 3l

....

WB1 WB2

1,


Ref. [40, 41]

(b) Interleaved WB Arrangement

(a) Continuous WB Arrangement

WB1

Waveband Arrangement


It retains multi/demultiplexing granularity at the individual wavelength channel level while outputting WBs at different ports. It can accommodate multiple input fibers simultaneously and demultiplex each band to different output ports. [40]-[43]

Two concatenated conventional individual wavelength channel AWGs

8-skip-0 band configuration with a total of 40 channels and 100-GHz spacing [39]

Specially designed two concatenated arrayed-waveguide gratings (AWG)

8-skip-0 band operation supporting a total of 32 channels at 100-GHz spacing -409 layers, non-linear dispersion at the band edges [38]

A thin-film filter

Waveband Multi/Demultiplexer


We.3.A.1

( ( (

( ( ( ( ( ( ( ( ( ( (

C B A 15 11 , 7 , 3 ) D C B A 16 12 , 8 , 4 ) D C B A 1 13 , 9 , 5 ) D C B A 2 14 , 10 , 6 ) D C B A 3 15 , 11 , 7 ) D C, B, A) 4 16 12 8 D C B A 5 1 , 13 , 9 ) D C, B, A) 6 2 14 10 D C B A 7 3 , 15 , 11 ) D C B A 8 4 , 16 , 12 ) D C, B, A) 9 5 1 13 D C, B, A) 10 6 2 14 D C B A 11 7 , 3 , 15 ) D C, B, A) 12 8 4 16

D

( 13D 9 C, 5 B, 1 A) ( 14D 10C, 6 B, 2 A)

AWG Y y1 Y1 y2 Y2 y3 Y3 y4 Y4 y5 Y5 y6 Y6 y7 Y7 y8 Y8 y9 Y9 y10 Y10 y11 Y11 y12 Y12 y13 Y13 y14 Y14 y15 Y15 y16 Y16 y17 Y17 y18 Y18 y19 Y19 y20 Y20

....

13C14C15C16C 9 C10C11C12C 5 C6 C7 C8 C 1 C2 C3 C4 C 13D14D15D16D 9 D10D11D12D 5 D6 D7 D8 D 1 D2 D3 D4 D

WB4C WB3C WB2C WB1C WB4D WB3D WB2D WB1D WB4A 13A14A15A16A WB3A 9 A10A11A12A WB2A 5 A6 A7 A8 A

13B14B15B16B 9 B10B11B12B 5 B6 B7 B8 B 1 B2 B3 B4 B

WB4B WB3B WB2B WB1B

WB1A 1 A2 A3 A4 A

....

WB2

WB3

Vol. 6 - 115

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20


Ref. [41]

116

116

116

116

7 cm

WB 4 WB 3 WB 2

WB 4 WB 3 WB 2 WB 1

WB 4 WB 3 WB 2 WB 1

WB 4 WB 3 WB 2 WB 1

Loss: