1 Next Step in Networking: Issues and Future Dae Young KIM CNU dykim@{anf.ne.kr; cnu.ac.kr}

Post on 12-Jan-2016

215 views 1 download

Transcript of 1 Next Step in Networking: Issues and Future Dae Young KIM CNU dykim@{anf.ne.kr; cnu.ac.kr}

1

Next Step in Networking:Issues and Future

Dae Young KIMCNU

dykim@{anf.ne.kr; cnu.ac.kr}

2

Contents Part I: Network Technology Evolution

Where We Are Lessons What’s Next

Part II: Deployments and Applications Global Advanced Networks Cyber Infrastructure e-Science Asia-APAN A National Optical Networking Capability of

Internet2 NRL & HOPI

3

Part I:Network Technology

Evolution

4

Millennium Retrospect Data - Voice Comm.: Convergence? Switching: Circuit to Packet, and ...? Links: HDLC, ATM, LANs, Ethernet, ... Routing: Telecom to Data(Internet) End-to-end protocols: TCP Applications: Web service Wired vs Wireless: Copper vs Fiber Internetworking: IP

5

Where We Are (I)

By Steve Deering

email WWW phone...

SMTP HTTP RTP...

TCP UDP…

IP

ethernet PPP…

CSMA async sonet...

copper fiber radio...

email WWW phone...

SMTP HTTP RTP...

TCP UDP…

IP

ethernet PPP…

CSMA async sonet...

copper fiber radio...

email WWW phone...

SMTP HTTP RTP...

TCP UDP…

IP6

copper fiber radio...

email WWW phone...

SMTP HTTP RTP...

TCP UDP…

IP6

copper fiber radio...

IP as Ultimate Internetworking Glue IP over Everything Everything over IP Hourglass vs Wine

glass

6

Where We Are (II) Ethernet as Winning Link Technology

Ethernet Everywhere WLAN = Wireless Ethernet?

IP + Ethernet for Data Not perfect for QoS Streams(Voice, Audio, V

ideo) Just right for non-QoS(quasi-QoS or CoS) QoS possible for a few sessions

7

Lesson: Keep it Simple and Stupid

Simple is the best Internet vs OSI Ethernet vs ATM

End-to-End Argument Keep network simple/stupid and Put Intelligence at edges/ends Don’t build in the core anything that can be

built at edges/ends

8

Lesson: Flexible vs Conservative

Protocol Be flexible in what you receive Be conservative in what you send Example: TCP

Network Be flexible in what you adopt Be conservative in what you abandon Examples: Magnetic, Modem, Copper, ... Future examples: Ethernet, IP, ATM, ...

9

Lessons: More

Convergence Data + Voice Wired + Wireless Dream Not Come True?

Extremely difficult to put new functions/features in the Infra

Layer independence Need new assessment

10

What’s Next? Internetworking issue over: IP Applications, Services!

Middleware, Web Services, GRID, … Wireless: Just last hop

Mobility, Security, QoS Physical, Medium; Still evolving Long way yet to the Netopia? We’re in the low tide.

11

Issues Yet Pending

QoS Multicast; Group Communications Mobility Security Which layer?

Not IP? Link? Application? Wireless

12

QoS Myth QoS is needed in one to one connections for real time

voice and video e.g Doctor video conferencing with a patient

BUT, most Internet applications are NOT one to one real time connections, they are many to one and many to many type of connections e.g.

Doctors retrieving X-ray image from a database Multicast distribution of a movie etc Many users going to the same web site

End to end QoS is real hard if you have more than a one to one, real time connection

13

QoS(I) Different Ideas about QoS

Security, Static BW, Rigorous Definition of QoS

Reliability Throughput

Dynamic thru Quasi-dynamic Delay, (Delay) Jitter

14

QoS(II)

Back to the Basics/Principles(KISS) Circuit for QoS Packet for Data Don't mix up

Circuit over Packet? ATM, PWE, etc., …

Packet over Circuit!

15

SAND

MM

QoS

PKT

Circuit over Packet

QoS(III)

16

Hybrid Transfer Mode:HTM

A

D

C

B

Usert0

Buff er

To Remote

A B B D B B C B

Cycle1 Cycle2

Header

CL slot

CO slot

A B B B B B C B

Cycle3 Cycle4

I dle slot

t1 t2 t3

Seamless Packet over Circuit

17

And More Multicast

Not at the IP layer? At the Application Layer?

Overlay Multicast, CDN At the Link or Physical Layer?

Mobility Fast enough? Where?

Security Mature? Convenient?

18

Xcast IP packet (option) header 에 수신자의 unic

ast IP 주소들을 explicitly 포함

19

Overlay Multicast

Comparison with IP (network-layer) multicast

20

CDN(Content Delivery Network)

UnicastUnicast

21

Part II:Deployments and

Applications

22

Global Advanced Networks

23

Advanced Networks

High Performance R&E or R&D NetworkMbps -> Gbps -> Tbpsfor research, education, development Advanced Technologies on the NetworkAdvanced Applications on the Network

Network, Application 기술 개발이 미래 정보사회 구축 , 산업발전의 기초 선진국이 되기위한 필수요소

24

Basic CA*net 4 Topology

Halifax

Edmonton

Seattle

VancouverWinnipeg

Quebec City

MontrealOttawa

Chicago

Halifax

New York

Regina

Fredericton

CharlottetownVictoria

Windsor

London

Sudbury

Thunder Bay

Saskatoon

Kamloops

Buffalo

Spokane

Minneapolis

Albany

St. John's

Calgary

Toronto

Prince George

Hamilton

Kingston

CA*net 4 Node (Mini-IX)

Possible Future Breakout

Possible Future link or Option

CA*net 4 OC192

26

GEANT

27

Cyber Infrastructure

28

The Need for a Global Research and Education Network

A global R&E network is required to support true global cyberinfrastructure which will underpin global e-science

However international connections very slow compared with R&E network backbone speeds

Global connection effort not well-coordinated – dominated by bilateral thinking

29

연구 교육을 위한 특수지식환경(Grid community, e-Science community, virtual community)

교육훈련 및 프로젝트 어플리케이션

HighPerformance

Computationalservices

Base Technology: computation, storage, communication

Networking, Operating System, Middleware

Data, Information,Knowledge

Managementservices

Observation,Measurement,

Fabricationservices

Interfaces,Visualization

services

Collaborationservices

= Cyberinfrastructure : hardware, software, service, personnel, organization

Cyberinfrastructure (NSF)

30

e-Science

31

EU-GEANT, UK-eScience EU 는 유럽의 초고속연구망 (GEANT) 및 많은 Grid Project 를 지원 GEANT: 유럽의 어플리케이션 영역뿐만 아니라 ( 네트워크 포함 ) 연구 자체를 지원하는 Infrastructure

30 개국 이상 참여 , 28 개 국가 및 지역 연구교육망 포함3,000 개 이상의 연구 및 교육기관 참여9 개의 10Gbps, 11 개의 2.5Gbps

UK-eScience국제협력을 통한 차세대 연구에 대한 Infrastructure 를 제공genomics, bioscience, particle physics, astronomy, earth science & clim

atology, engineering systems, social sciences차세대 open platform standard 를 위해 노력 optimal international infrastructure 제공

32

What is eScience? The ultimate goal of e-science

to allow students and eventually members of the general public to be full participants in scientific discovery and innovation.

Using advanced high speed networks combining new concepts in distributed computing, peer to peer file sharing, Grid technology and “Third Wave of the Internet”

33

Third Wave of the Internet

The first wave text and data services such as e-mail and FTP

The second wave the web which improved ease of use and

facilitated the transfer of images, sound and videoThe third wave

integration of gridsp2p networkingopen sourcedistributed computing enabled by next generation

web services, semantic web and high speed networks

34

Today’s Network

Application

OS

Data

Application

OS

Data

Network

UserUser

The application is tightly bound to the OS

The network is a mechanism for applications to communicate with each other

The network is subservient to the computer

35

Third Wave Network

Network

Application and Data

OS

Data

OS

Data

OS

Data

OS

Data

Third Wave

OS OS

Application and data exist on the network and are uncoupled from any specific machine or location

The computer is subservient to the network

Third Wave

Third Wave Third Wave Third Wave Third Wave

36

Grids - Third Wave -Web

Computational Complexity

DataComplexity

Source: Toney Hey UK eScience Grid

37

What are Grids? Grids enable the new science Original motivation, and support, from high-end science and engineering Enable communities (“virtual organizations”) to share resources as they pursue common goals New applications enabled by the coordinated use of geographically distributed resources

E.g., distributed collaboration, data access and analysis, distributed computing, instrumentation

Persistent infrastructure for large scale computing problems

Using distributed computing resources of schools, universities and research centers

38

Site A(Kerberos)

Site B (Unix)

Site C(Kerberos)

Computer

User

Single sign-on via “grid-id”& generation of proxy cred.

Or: retrieval of proxy cred.from online repository

User ProxyProxy

credential

Computer

Storagesystem

Communication*

GSI-enabledFTP server

AuthorizeMap to local idAccess file

Remote fileaccess request*

GSI-enabledGRAM server

GSI-enabledGRAM server

Remote processcreation requests*

* With mutual authentication

Process

Kerberosticket

Restrictedproxy

Process

Restrictedproxy

Local id Local id

AuthorizeMap to local idCreate processGenerate credentials

Ditto

Grid in Action:“Create Processes at A and B that Communicate & Access Files at C”

39

Examples eResearch Grid Projects

ATLAS

Sloan Digital Sky Survey

LHC

ALMA

40

Components of CI-enabled science & engineering

CollaborationServices

Knowledge managementinstitutions for collection building

and curation of data, information,literature, digital objects

High-performance computingfor modeling, simulation, data

processing/mining

Individual &Group Interfaces& Visualization

Physical World

Humans

Facilities for activation,manipulation and

construction

Instruments forobservation andcharacterization.

GlobalConnectivity

41

Streams of Activity Converging in a CI Initiative

C ol la

b ora

torie

s

GRIDS (broadly defined)

E-science

CI-enabled Science & Engineering Research & Education

Specific disciplinary projects (not using above labels)

42

Cyberinfrastructure Opportunities

LIGO

ATLAS and CMS

NVO and ALMA

The number of nation-scale projects is growing rapidly!

Climate Change

43

Futures: The Computing Continuum

National PetascaleSystems

National PetascaleSystems

UbiquitousSensor/actuator

Networks

UbiquitousSensor/actuator

Networks

LaboratoryTerascaleSystems

LaboratoryTerascaleSystems

Ubiquitous Infosphere

CollaboratoriesCollaboratories ResponsiveEnvironmentsResponsive

EnvironmentsTerabit

Networks

ContextualAwarenessContextualAwareness

SmartObjectsSmart

Objects

Building Out

Building Up

Science, Policy and Education

PetabyteArchivesPetabyteArchives

44

The Changing Style of Observational Astronomy

The Old Way: Now: Future:

Pointed, heterogeneous

observations (~ MB - GB)

Large, homogeneous sky surveys (multi-TB, ~ 106 - 109 sources)

Multiple, federated sky surveys and archives (~ PB)

Small samples of objects (~ 101 - 103)

Archives of pointed observations (~ TB) Virtual Observatory

45

Crab Nebula in 4 spectral regionsX-ray, optical, infrared, radio

46

Four LHC Experiments: The Petabyte to Exabyte Challenge

ATLAS, CMS, ALICE, LHCBHiggs + New particles; Quark-Gluon Plasma; CP Violation

Data storedData stored ~40 Petabytes/Year and UP; ~40 Petabytes/Year and UP; CPU CPU 0.30 Petaflops and UP 0.30 Petaflops and UP

0.1 to 1 Exabyte (1 EB = 100.1 to 1 Exabyte (1 EB = 101818 Bytes) Bytes) (2007) (~2012 ?) for the LHC Experiments(2007) (~2012 ?) for the LHC Experiments

47

Virtual Observatory

http://www.us-vo.org/ Discovery process will rel

y on advanced visualization and data mining tools

Not tied to a single brick and mortar location

Will cross correlate existing multi-spectral databases petabytes in size No new telescopes or radio

dishes. Just big networks interconnecting large databases

48

Earthquake Engineering

Network for Earthquake Engineering Simulation (NEES)

$ 81.8M FY01-04 NSF support requested. Scoping study managed by NCSA; sponsored by

NSF NEES will provide a networked, national resource of ge

ographically-distributed, shared-use, next-generation, experimental research equipment installations, with tele-observation and tele-operation capabilities.

NEES will shift the emphasis of earthquake engineering research from current reliance on physical testing to integrated experimentation, computation, theory, databases, and model-based simulation using input data from EarthScope and other sources.

NEES will be a collaboratory – an integrated experimental, computational, communications, and curated repository system, developed to support collaboration in earthquake engineering research and education.

49

Ambient mic(tabletop)

Presentermic

Presentercamera

Audience camera

Grid Communities

Access Grid Collaboration Enable collaborative work

at dozens of sites worldwide, with strong sense of shared presence

Combination of commodity audio/video tech + Grid technologies for security, discovery, etc.

CRC, Sheraton and universities participating

50

Asia-APAN

51

Asia–APAN APAN Network: 한국 , 일본 , 중국 , 대만 , 싱가폴 , 호주 , 말레이시아 , 태국 , 필리핀 , 홍콩 , 베트남 , 인도네시아 , 스리랑카 , 미국 , 프랑스 , EU 연결 APAN Community: 아태지역 각국의 초고속연구교육망 제공자 및 사용자 의 모임

아직 Cyberinfrastructure 혹은 eScience 가 형성되지 않음 . 그러나 일본을 중심으로 Natural Science 분야에서 활발히 활동

Weather/Climate, Agriculture, Earth Monitoring Medical/Health Museum, Art High Energy Physics, BioInfomatics, NanoTechnology … etc

52

GOSObservation to understandthe current weather

53

GMSGMS21 March 200221 March 2002

TrajectoryTrajectory

Weather/Meteorology

54

Digital Earth

55

Forest Fire Early Detection System

ANDES

Hotspots are observed in NOAA-AVHRR and new lights are detected by DMSP-OLS.

Both data are combined and the coordinate data are stored in a file and also plotted on base images (left).

These data are sent to the related organizations in each country and also archived to be displayed on the web.

These information are automatically sent to a mobile-phone (i-mode) of the manager by e-mail every day.

12

56

HENP Major Links: Bandwidth Roadmap (Scenario) in Gbps

Year Production Experimental Remarks

2001 0.155 0.622-2.5 SONET/SDH

2002 0.622 2.5 SONET/SDH DWDM; GigE Integ.

2003 2.5 10 DWDM; 1 + 10 GigE Integration

2005 10 2-4 X 10 Switch; Provisioning

2007 2-4 X 10 ~10 X 10; 40 Gbps

1st Gen. Grids

2009 ~10 X 10 or 1-2 X 40

~5 X 40 or ~20-50 X 10

40 Gbps Switching

2011 ~5 X 40 or

~20 X 10

~25 X 40 or ~100 X 10

2nd Gen Grids Terabit Networks

2013 ~Terabit ~MultiTbps ~Fill One Fiber

Continuing the Trend: ~1000 Times Bandwidth Growth Per Decade;We are Rapidly Learning to Use and Share Multi-Gbps Networks

57

CN

SG

PERTH

GHANA

Buenos Aires/San

Paolo

St. Petersburg

Kazakhstan Uzbekistan

ChenaiNavi

Mumbai

Barcelona GreeceMD

NL

CA

Global Quilt Initiative – GMRE Initiative - 001

Global Medical Research Exchange Initiative

Bio-Medicine and Health Sciences

Layer 1 – Spoke & Hub Sites

Layer 2 – Spoke & Hub Sites

Layer 3 – Spoke & Hub Sites

Propose Global Research and Education Network for Physics

58中国农业大学植保生态智能系统技术(中国农业大学植保生态智能系统技术( IPMistIPMist ))实验室 网址:实验室 网址: http://www.ipmist.orghttp://www.ipmist.org

CASTCAST

EDUCEDUCSORGSORG

PCOMPCOM

e-Learning

Cross-organizational and international cooperation (COINCO) to Cross-organizational and international cooperation (COINCO) to forge innovative approaches to the challenge of e-learning marketforge innovative approaches to the challenge of e-learning market

APANAPAN

APRTCAPRTC

59

과학 기술 연구 환경의 획기적 변화고속 네트웍을 이용한 데이터 , 연산 능력 등 자원의 공유를 통한 연구

효율 극대화 초고속망 없이는 경쟁력 있는 첨단연구 불가 : 바이오 , 항공 , 기상 등

6T 전분야세계적 과학 기술 망 블록 등장 예상 됨 초 부처적인 사업 개념 확립 필요

e-Science, e-Education 한국의 초고속망 - KOREN, (KREN), (KREONET) 국제 : APII, TEIN 정부 - 기관 - 학교를 포함하는 이용자 그룹 : ANF

Advanced Networks and Cyberinfrastructure in Korea

60

DancingQ(I)

61

DancingQ(II)

StarTAP

U.S.

622M×2

WIDE, JGN

Japan

Tokyo-XP

Seoul

KOREN

Busan

Daegu

Suwon

TransPAC

vBNS, Abilene

APII Test-bed(Asia Pacific Information Infra.)

1G

KOREN(KOrea advanced REsearch Network)

KII Network(KII:Korea Information Infrastructure): ATM network

Daejeon

Kwangju

The National Center for Korean Traditional Performing Arts

Busan National Univ.

KII network

Kyusue

Performance Sites

62

HDTV over IP Demo(I) 270Mbps High-Definition streaming video from Portland

FukuokaBusan

Portland

Tokyo

~8000km/~5000mi

Demonstrate the performance of IP network to support extremely high rate multimedia data.

BUSAN JAPAN USA

UW J uniper M10

UWHP 4148 Switch

IEEAF/WIDE BI4K

APAN J uniper M20

Gekai XPJ uniper M10

KOREN Busan Cisco GSR

KOREN BusanCisco4006

Busan Marriotte Hotel Cisco4003

ACTACT10M100M

1 2 3 4

13 14 15 16

5 6 7 8

17 18 19 20

9 10 11 12

21 22 23 24

UPLINK

1 2 3 4 5 6 7 8 9 101112

131415161718192021222324COLCOL

PWR

SWITCH

HD D>A

HDCAM DECODER

HD DISPLAY

270 Mb HD CLIENT203.255.251.200

ACTACT10M100M

1 2 3 4

13 14 15 16

5 6 7 8

17 18 19 20

9 10 11 12

21 22 23 24

UPLINK

1 2 3 4 5 6 7 8 9 101112

131415161718192021222324COLCOL

PWR

SWITCH

ACTACT10M100M

1 2 3 4

13 14 15 16

5 6 7 8

17 18 19 20

9 10 11 12

21 22 23 24

UPLINK

1 2 3 4 5 6 7 8 9 101112

131415161718192021222324COLCOL

PWR

SWITCH

AST- HD- 1203.181.249.211

Data

WORKSTATION

KOREN 1Gbps APII

1Gbps

JGN622Mbps

WIDE IEEAF10Gbps

63

HDTV over IP Demo(II)

64

Global Advanced Networks

APAN

CII-K/ANF

Internet2CII/NSF

GEANT/EUeScience/UK

65

Advanced Network Forum (http://anf.ne.kr)

Korea

JapanTEIN

TransPAC

APII-2Link

622Mbps * 2

10Gbps

45Mbps

China

Thailand

Malaysia

Singapore

Indonesia

Philippines

Australia

Vietnam

GEANT(to

Paris)

Internet2/STAR TAP(to

Chicago)

ANF Vision of ANF Vision of Future Distributed HUBFuture Distributed HUB

ANF Vision of ANF Vision of Future Distributed HUBFuture Distributed HUB

Myanmar

nGbps

nMbps

APII, A3I Links

66

Distributed Cluster - Proposal

Access PointExchange Point

Australia

KoreaJapan

China

Thailand

Malaysia

Singapore Indonesia

USA

PhilippinesVietnam

Hong Kong

Sri Lanka

Taiwan

South East Asia Cluster(MY, TH,…)

North America

North East Asia Cluster(JP, KR,…)

Europe

Oceania Cluster(AU, ,…)

67

IEEAF: APAN Opportunities

IEEAF: 622 Mbps POS +10 Gbps

AARNet

68

A National Optical Networking Capability of

Internet2

69

Abilene Focus(’03~’04)

High perfomance , native advanced services: Multicast, IPv6, Large Flows End-to-End support

Dedicated Capability Experimentation 10-Gbps optical upgrade

TeraGrid experiment:best-effort virtual circuit

Advanced Restoration Techniques

70

Abilene Restoration

Abilene has a partial mesh of unprotected DWDM circuits replacing protected SONET circuits

VoIP and other real-time applications are becoming more important

Graceful restart for IS-IS and BGP

71

Why a national optical facility?

Expansion capability (λ’s) at marginal cost

New technology: 10Gigabit Ethernet in place of SONET

Means for introducing interdomain optical switching

Influencing development of new protocols at IP/optical interface

72

Abilene Network 10-Gpbs Optical Upgrade –(’02~’03)

73

NRL(National RambdaRail)

74

What is NLR(I)

a consortium of leading U.S. research universities and private sector technology companies

NLR aims to reenergize innovative research and development into next generation network technologies, protocols, services and applications

75

What is NLR(II)

combine new optical circuit technologies and existing high performance Internet services to develop a next generation of advanced networking capabilities.

intend to offer national experimental service over a λ–‘lambda grid’ deployment initially

76

Features

Largest optical networking & research facility in the world ~10,000 route-miles of dark fiber Four 10-Gbps λ’s provisioned at outset

Use of high speed Ethernet for WAN 10 Gigabit Ethernet LAN PHY is primary inte

rface

77

Internet2 and NLR Intend to offer national experimental ser

vice over a single λ for first 5 years of operation –lambda grid

Corporate partners Cisco(optronics/switching/routing) Level 3(fiber) Strong interest by other optronics companie

s Budget: $83-100M over 5 years

78

National LambdaRail Architecture

79

NLR’s ‘Virtuous Circles’ and the Vital Role of Dark Fiber

80

HOPI(Hybrid Optical/Packet Infrastructure

81

Outline

Assembling the vital ingredients High-peformance national IP network –Abilene Regional Optical Networks(RONs) National optical capabilities –NLR

Hybrid networking – next steps Plans for the NLR λ dedicated to Internet2 Steps towards developing a Hybrid Optical Packet I

nfrastructure(HOPI)

82

10-Gbps λ over full NLR footprint

Details 5-year commitment Likely 10GigE framing(in lieu of

OC192c SONET) Expect some type of ‘TDM’

infrastructure to be provisioned by Internet2 in collaboration with NLR

83

Hybrid Optical Networking Includes both IP packet and circuit capabilities Provides new opportunities for demanding applic

ations and network experimentation Does not obviate security and performance issu

es Requires interoperability and varying degrees of

on-demand resource allocation Depends on interplay of national, regional, and

metropolitan efforts Examples: National LambdaRail, regional optical

networks

84

NLR-Internet2 relationship

85

References ANF – http://anf.ne.kr APAN - http://apan.net Canarie – http://www.canarie.ca Internet2 – http://www.internet2.edu Geant – http://www.geant.net National LambdaRail -http://www.getlight.net/

86

Conclusion

IP, Optical, Wireless Low Tide in Networking Research Yet Long Way to the Netopia Chances for the Innovative Information Infrastructure The Third Wave