toward new generation network - beyond the internet and...
TRANSCRIPT
OFC NFOEC 2009
Toward New Generation Network- Beyond the Internet and Next
Generation Network (NGN) -
March 24, 2009
Tomonori Aoyama
Keio University/NICT
Content
- The Internet, Next Generation Network and New Generation Network
- Requirement for New Generation Network
- Some important research items for New Generation Network
- Some future applications over New Generation Network
2
20202008
An image of a network evolution
Legacy Telephone NW
Cellular Phone NW
The Internet
Started NXGN(NGN)
Future NWGN
NXGN: Next Generation Network (NGN)NWGN: New Generation Network
IP-based NW
Post-IP NW
3
NXGN and NWGN(Next) (New)
NXGN (NeXt Generation Network): NGN
Replacement of legacy telephone
networks using IP-based networks to provide
triple-play/quadruple-play services
NWGN NeW Generation Network)
Clean-slate designed network architectures and main protocols different from IP-based
networks , which can be post-Internet/NGN
Industry is investing their resources to NXGN deployment.
NICT contributes to promote R&D on NWGN. 4
Can IP Protocol handle the explosion of information in 2020s ?
ApplicationsWebEmailVideo/Audio
TCP/IPAccess Networks
Ethernet (LAN)Wireless (LMDS, WLAN, Cellular)CableADSLSatellite
Applications
AccessNetworks
5
© National Institute of Information and Communications Technology
Internet Too Much Complicated
Cannot add new functions Cannot provide services for future society
Original Internet Architecture
mobility
anycast
flow-label
guaranteedservice?
universal communication?
small devices? authentication?
L2: Datalink Layer
L3: Internet Layer
L4: Transport Layer
Entrust network with your life & living ? (tele-medicine, ITS & anticrime, finance)Rich life? (connecting sensor, RFID)Safe? Secure? (spam, DDoS)Never broken? how long? (sustainable society)Flexible to future change? (nobody knows future)
Adding Functions
Individual optimum but NOT global optimumDesign from scratch has come!
Increasing Layers
dependability?
6/38
Architecture is different from that of the Internetalthough both are based on the IP protocol
Application
APINNI
Transport Stratum
Service StratumService Control Function
Transport Function
Other N
etwork
End U
ser System
UNI
-to-7
Aoyama 6
Best effort bearer function to interconnect multiple router based network
Terminal Terminal Server Terminal Terminal Server
IP router networkBest effort
IP router networkBest effort
IP router networkBest effort
Internet
NGNVersatile
Bearer functions
NGNVersatile
Bearer functions
NGNVersatile bearer functions
QoS controlled bearer function to interconnect multiple networks with clear responsibility
NGN
A comparison between The Internet and NGN
No over all network planningTCP/IP Protocol is the only common ruleBest effort based network and no clear responsibility and control rule exist among networks
User can have freedom to install applications
IP based network with network control function and with clear responsibility for the control
Qos control and security functions are installedMaintain the Internet connection function
IP router networkBest effort
IP router networkBest effort
NGNVersatile bearer function
8
R&D Marathon to realize a new communication paradigm has just started globally.
Who will win ?
Now !
9/38
Research on New Generation Networks started globally
To realize NWGN, tight collaboration among the E.C., U.S., China, korea and Japan is important.
the future Internet architecture through a clean slate approach. make a small scale, but a large number of projects converge into a fewer number of full scale architectures, and to verify using GENI A total of 1,200 million dollars to 26 projects by 2006
GENI
development. programmable
International cooperation is also included as a target.Project Bureau (GPO) is BBN. Planning to offer prototype, 1,500 million dollars in December 2007 over a two year period. A scale of construction budget is 367 million dollars,
total fields of science and technology
projects in Call 1 Call 2 is investigating
E.C.FP7FIND
main individual programs-Network of the Future
(Call 1, D bureau ) 2 billionEuros)
-FIRE(Call2, F bureau ) 4,000 million Euros
NWGN AKARI architecture development project
specification in April 2007)R&D on evaluation and establishment in
the structure of NWGN) (NICT commitment, up to 20 million Yen times 2 years plural cases)
R&D for dynamic network technology (NICT commitment, 4 years plural cases)
NWGN Research projects
JGN2plus Testbed
development2008
2010.
The E.C. Budget for GÉANT was 9,300 million Euros over 4 years since 2004. It is in the middle of preparations for upgrading and for greater capacity by transforming it into GEANT3 from 2008.
GÉANT3
The United States
NWGN Promotion Forum
November, 2007 Working in 4 WGsMinistry of International Affairs and Communications and NICT are a co-secretariat.
Japan
10/38
of communication networks !
============Telephone Network
Digitalization ISDN B-ISDN NGN over IP
IP Network TestbedARPANET / NSFNET
Commercial Internet1969
2008
R&D for beyond IP started worldwide
2006/2007
GENI/FIND, NWGN, FP7
Commercial beyond IPnetworks will start
It is time to start the second ARPANET/NSFNETtype academic networks globally for establishing beyond the Internet/NGN based on IP.
1990
1995
2020
11
World market shares for main ICT systems (2005)
Router
Server
Cellular Phone
Desk-top PC
Note PC
LSI
LCD
Japan US Korea China
Legacy TV setDigital Camera
12
3 Approaches are required to make a clean-slate designed architecture
Top-down ApproachSocial requirements and requirements for long tail applicationsIdeas to meet those requirements
Bottom-up ApproachTechnical breakthroughWhat can be realized by this breakthrough ?
New generation network architecture Design Principle
13
© National Institute of Information and Communications Technology
New Generation Social & Design Requirements in 2015~2020
Peta-bps backbone, 10G-bps FTTH100-billion devices, M2M, 1-Mega stationsCompetitive industry and user-oriented servicesMedical care, traffic control, emergency, four-ninePrivacy, financing, food tracking, anti-disasterRich society, handicapped, aged supportEarth & human monitoringBroadcasting & communication, web 2.0Economic Incentive (Business-cost model)Ecology, sustainable society Human possibility, universal communication
Design RequirementsSocial RequirementsCapacityQuantityOpennessRobustnessSafety & SecurityDiversity (Long-tail) Ubiquity (Pervasive)Converge & SimplifyNetwork ModelEnergy-savingEvolvability
New Generation Network Designed by the Architecture
14
K M G
Access frequency [page/day]
K
M
G
T
P
Yahoo300Mpage/day
InternetTV11Mpage/day
Web content
Cap
acity
of
cont
ent
SDTVDVD>GB
Web 10kB/page
MP3>MB/music
DigitalCinema> 100GB
e Commerce
IP TV
HDTV
Cine-grid
B2C
B2B
[bit]
Contents in the ubiquitous societyFrom tiny to huge
Sensor & RF ID
S2M
P2P
Both directionsTiled Display
15
AKARI Architecture Design Project Toward NewGeneration Network
Designing the future, diverse, new generation network beyond 2020- pick up techniques for the future under the principles- integrate & simplify them with design methods- Clean-slate design approach
http://akari-project.nict.go.jp/http://www.akari-project.jp/
AKARI a small light in the dark pointing to the future 16
© National Institute of Information and Communications Technology
AKARI Sustainable Architecture Principles
ID-Locator separationBi-directional authenticationTraceability
Self-* properties (emergent)Autonomic distributed controlScalableSocial Selection
KISS (Keep It Simple, Smart)Crystal synthesis Common layer (layer degeneracy)End-to-end (original Internet)
Sustainable EvolutionReliable Network Space
Diversity Inclusion
17
Overlay Network
Underlay Network
Cross-layer Control Mechanism
Application
Study Items for NWGN Architecture
Photonic NWMobile NW
Sensor NW
All the elem
ents should be redesigned.
18
Fundamental Principles
End-to-end argumentOpen, Transparent vs SecurityHandling data format
Datagram PacketFlowPath/CircuitMessage
Layering StructureIdentifier / Locator SeparationIn band vs Out of band
19/38
AKARI Architecture Focus (I)(1) AKARI Optical Switching - Path / Packet Integration -
QoS Guaranteed
Keep Both QoS Assuredness & Best Effort Efficiency=> OPS + OCS (Not OPS over OCS)
A,B,C,D E,F,G,H,I,J,K,L,M,N,O,P,Q,R,S,T
Wavelength
Pow
er S
pect
rum
Header Payload
Sustainable in Capacity
Optical Circuit Switch
Best Effort
Optical WaveSharing Optical Packet Switch
(with Fiber Delay Line)
Parallel Optical Packet Transmission
- Scalable- Energy-Saving
Sustainable in Usage - Allow Usage Diversity
Requirements:- Handle different types of traffic- Peta-bps switching capacity- Tera-bps link speed (40G x 100)- 100 billion tiny terminals
20
AKARI Architecture Focus IISeparation of Identifier & Locator
identifier
locator
identifier
identifier
locator
locator
locator
locator
Applicationlayer
Transportlayer
Networklayer
Identification sublayer
Delivery sublayer
Mapping
Destination Node is defined by Identifier
Destination Node is defined by Identifier
Destination Node is defined
by Locator
Bypass
ConventionalID/Locator Separation
Destination Node is defined by Locator
21
AKARI Architecture Focus (III)
Self-organizing Control (Concept)
nageableFast Recovery
ultaneousSim sSimultaneousSimultaneousfailuresfailure
ChainCreaction
Softwaree
Bad provisioning
Optimum
Adapted
Robust
Efficient
Base StationsBase Stations
User a, b, c
#1 #2#3 #N
WideSharedChannel
User a, b, c
- PDMA (Packet Division Multiple Access)
Free from:-Frequency Channel Allocations-Cell Design
Sustainable in Management and Capacity
StaticChannelAllocation
CSMA/CA(packet)
- Self-Organizing Mesh Network
New Network Science RequiredScale-free NetworkBio-inspired Network 22
AKARI Architecture Focus (IV)
(4) Network Virtualization (Concept)
Physical Network
Real TestBed Network
VN1 VN2
(a) Isolated Virtual Networks
Real Operational Network
VN1 VN2
(b) Transitive Virtual NetworksSelf-evolvable
Real Operational Network
VN1
VN2
(C) Overlaid Virtual Networks
Optical Path NetworkWireless Network
23
© National Institute of Information and Communications Technology
Various Virtualization Techniques
Computational VirtualizationHypervisor-Based, Host-Based, OS-Level VirtualizationHardware-support (Intel VTx, AMD-V)
Network VirtualizationI/O Virtualization (Intel VTd , AMD IOMMU) Resource ContainerNetwork Resource Allocation Management
Hardware
HostOS
Apps VM VM
GuestOSGuestOS
Apps Apps
Hardware
VMM
VM VM
GuestOSGuestOS
Apps apps
VM
MngOS
Manager
Hardware
HostOS
Apps VE VE
Apps Apps
Hypervisor-Based Host-Based OS-Level
24/38
© National Institute of Information and Communications Technology
Overlay to Network Virtualization
VMM/Hypervisor
Mgm
t S/
WVM
(Slic
e)VM
(Slic
e)
VM(S
lice)
A Slice of Computation Resources
PC Server
switch
25/38
1999
(year)
2003 2004
2005
2006
2007
2008
2009
2010
2011
2012
2013
JGN Project Mile Stone
(19999.49 -2004.3)0 ((2004.4( -2008.33)3
JGNN2pllusN lTo New Testbed
NICT2nd Mid-term Plan
(2006.4-2011.3)
CRL (NICT)1st Mid-term Plan(2001.4-2006.3)
2008.4-2011.3 2011.4-
NICT3rd Mid-term Plan
2011.4-
26
Network Research and Operation [Predicted Research Projects]
Network Operation TechnologyTraffic Measurement TechnologyNew Generation Network ArchitectureGrid NetworkOverlay NetworkAdvanced InternetVerification of Interconnection
CSTNET(China)
SingAREN(SingAREN)
JGN2 plus
Tokyo
CA*net4(Canada) SURFnet
(Netherlands)
Internet2(USA)
NLR(USA)
MREN(USA)
AARNet(Australia)
IEEAF(USA)
UKLight(UK)
GLORIAD(USA, Russia, China)
ThaiSarn(Thailand)
TWAREN(Taiwan)
HARNET(Hong Kong)
StarLight (USA)
UniNet(Thailand)
KOREN(Korea)
TransPAC2(USA)
LA
Chicago
HK
BKK
SG
CERNET(China)
GEANT2(Europe)
TEIN2(Asia, Europe)
APAN(Asia)
KR
Pacific Wave(USA)
27
Current Photonic Network R&D supported by NICT
All-optical transport1996 2005
Photonic node enabling broadband access2000 2005
Optical burst switching network2001 2005
Control plane for terabit-class network2001 2005
PHASE I
Photonic node with multi-granularity
switching capability 2005 2009
Access 2006 2010
Utility 2005 2009
Photonic RAM 2005 2009
PHASE II
NICT R & DNICT R & D29
NICT funding projects for photonic NW R&D
Photonic node with multiple granularity switching capability2005 2009
l Utility2006 2010
ll Access2006 2010
Optical RAM for all-optical packet switching2006 2010
Note: ll refers to a large-capacity and transparent information channel, identified by the wavelength value
Next phase program is discussed in PIF.
PIF: PhotonicInternet Forum
All-optical terabit-class LANAccess-to-backbone seamless networking
Establishing customer-initiated end-to-end 100Gbps optical path and QoS service provisioning between terabit-class LANs
Photonic crystal bit memory
100Tbps-Class photonic node with multiple granularity switching
30/38
Proposal for Phase III photonic Network R&D
PIF (Photonic Internet Forum ) is making the proposal for MIC
Green ICT
- Self-controllable power-minimum photonic network
100 times higher speed with the same power consumption as today
Application-oriented photonic network Network Virtualization
- Optimal resource allocation of photonic networkswith multiple QoS adapting application requirements
31/38
© National Institute of Information and Communications Technology
NWGN related R&D $ 21M
Photonic Network Projects ( 7 Groups ) $ 22 M
Security Technology Projects $ 11 M
Wireless Network Project ( 1 Group ) $ 3 M
Total about $ 60 M
NICT Funding Projects in 2008 Fiscal Year
$ 1 = ¥100
32/38
© National Institute of Information and Communications Technology
NICT Research Projects and NICT Joint Research Projects for NWGN
NWGN related R&D $ 30M
Network Testbeds $ 40M
MIC Funding Projects related to New Generation NetworkUbiquitous Network Platform ( 1 Group ) $ 15M( Ubiquitous Networked Robot Projects Proposing for 2009 Budget )
33/38
Long Tail Application
Transfer Rate100 Mb/s
100 Gb/s
Enterprise Users Researchers & Scientists
Long Tail Application
No of Users
Con
sum
er Users
Innovations have been derived from Long Tail Applications.
e.g. Internet , WWW,
34/38
Examples of long tail applications over NWGN
- Cloud Computing over photonic networks
- Connection of big tailed display for visualization of
e-science data
- Distribution of digital entertainment contents with
ultra high quality such as digital cinema & ODS*
ODS* : Other Digital Stuff
musical, opera, drama(Kabuki), sports, etc.
35/38
Large Scale Tiled Display System developed by UC San Diego
36/38
2K vs 4K: Beyond HDTV for Digital Cinema K 1K=1024 Pixel
2K=20484K=4096
=1920 1080
TV
DDiiggiittaall TTVV
Digital HDTV
DDiiggiittaall CCiinneemmaa
=2048 1080
4096 2160 4K
30 frames/s Interlace 24 frames/s Progressive
Digital Cinema standard was made by DCI and SMPTE is proceeding with the detailed specifications
37/38
Japan USA
LA
SeattleTokyo
YokosukaOsaka
Burbank
Production StudioCinema Theater
Local Distribution Center
Global Distribution
Center
CompressionEncryptionKey Center
1Gb/s IP
Link
Source: NTT
4K Pure Cinema Trial by Warner Brothers, Toho & NTT
38/38
© National Institute of Information and Communications Technology
-Theater System
In the projection room of VIRGIN TOHO CINEMAS
4K Projector
SecureMedia Box
(SMB)
TheaterControl Box
(TCB)
Corpse BrideHarry Potter 4V for VendettaDaVinci Code
PoseidenMission Impossible 3
+Tokyo Film Festival
Batman BeginsStealth
Source: NTT39/38
Requirement for Networking
2-hour 4K digital cinema contents Non-compressed : 5TBJPEG2000 Compressed : 250GB (1/20 compressed ratio)
File Transfer10 hours over 1Gb/s link Non-compressed30 minutes over 1Gb/s link 1/20 compressed
Real time streaming6Gb/s Non-compressed300Mb/s 1/20 compressed
Multicast functionDistribution of contents to theaters, halls and home theaters
QoS RequirementsPacket LossLatencySynchronization
40/38
© National Institute of Information and Communications Technology
Olympus 4K Digital Camera
42/38
Image of digital cinema and ODS distribution over a network
Projector
City hall, hospital,
Cinema Theater
Cinema Data Center School
Education
Live of Musical
Live of sports
Broadband Network
Home
43/38
Time
3 Axis for Digital Content Evolution
Content SearchContent Distribution
Content Creation Archive
Applications e-Commerce, e-Science, e-Learning, e-Medicine, Digital Library/ Musium, Digital Entertainment, Web2.0, Web3.0,
Volume
QualityFederated Structure
44/38
AMPAS ( Academy of Motion Pictures Arts and Sciences) published a report
strategic issues in archiving and accessing digital motion picture materials.
45/38
Worldwide Collaboration is indispensable
Global Collaboration from Japan point of viewsUS-Japan Joint Workshop on NWGN/FI
October31-Novbember 1,2008 at Palo Alto in 2009EU-Japan Symposium on Future Network
June 9-10, 2008 at BrusselsJuly 6-7, 2009 in Japan (Plan)
China-Japan ICT Forum (China Academy of Science NICT)July 8-9, 2008 in Shanghai
StandardizationITU-T established a focus group (FG) in Study Group 13 to discuss about future networks
ITU-T hosted the Kaleidoscope Academic Conference jointly with IEEE ComSoc.
46/38
© National Institute of Information and Communications Technology
US-Japan Joint Workshop on New Generation Network/ Future Internet
October 31-November 1, 2008 at Palo Alto
47/38
© National Institute of Information and Communications Technology
ConclusionThe Internet is expanding rapidly and creating new applicationssuch as Web2.0.NGN started in 2008 in Japan, and industry is concentrated their resources on the deployment of NGN.R&D on NWGN has started globally toward a new paradigm of communications. - Network architecture should be studied based on
requirements for ubiquitous networking and new networking technologies such as advanced Photonic networking technologies.
- Prototype of NWGN and the standard will be realized in 2015-2020 time frame.
- Network Testbeds are essential for R&D on NWGN. - Global collaboration among universities, industries and
governments is important to accelerate R&D. 48/38
Thank you for your attention !
49/38