1 Photonic Network Ken-ichi Sato sato . kenichi @lab. ntt .co. jp Satoru Okamoto [email protected]. jp Tutorial October 1, 2003 APNOMS 2003, Fukuoka, Japan NTT Network Innovation Laboratories 2 Part I : Photonic Network -Why It’s So Important?- Progress of Broadband Internet Access P2P Communication Broadband Services Advances in Transport Network Technologies Fundamentals of Photonic Network Photonic MPLS Router Part II : Photonic Network Control and Management Photonic Network Control and Management Overview Photonic Network Architecture Overview IP over Photonic Network Architecture Overview MPLS, GMPLS (MPLambdaS), and ASON GMPLS protocols GMPLS management Interoperability Test Events of Photonic Network Control Outline
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Part I : Photonic Network -Why It’s So Important?-Progress of Broadband Internet AccessP2P CommunicationBroadband ServicesAdvances in Transport Network TechnologiesFundamentals of Photonic NetworkPhotonic MPLS Router
Part II : Photonic Network Control and ManagementPhotonic Network Control and Management OverviewPhotonic Network Architecture OverviewIP over Photonic Network Architecture OverviewMPLS, GMPLS (MPLambdaS), and ASONGMPLS protocolsGMPLS managementInteroperability Test Events of Photonic Network Control
Data from Ministry of Public Management, Home Affairs, Posts and Telecommunications.
6
Leased Line Monthly Charge (34-45 Mbps)
Distance (km)
1$ = 122 yen1U= 109 yen
Deutsche Telekom (Digital 34MS)France Telecom (Transfix HD; 34 Mbps)British Telecom (Mega Stream34)AT&T (ACCUNET T45)NTT East (ATM Mega-Link)
Prices as of March-December 2002.
0
1
2
3
10 100
Tokyo
ParisLondon
New YorkDusseldorf
AsiaEurope
USA
Mill
ion
Yen
per
Mon
th
Tokyo
Paris
London
New York
Dusseldorf
Data from Ministry of Public Management, Home Affairs, Posts and Telecommunications.
7
New Z
eala
nd
Nether
lands
0
1
2
3
4
Switz
erla
nd
Japan
Icel
and
USANorw
ayCan
ada
Denm
ark
Belgiu
m
UK
Hong Kong
Singap
oreSw
eden
Taiw
anAust
riaFin
land
Korea
Austra
lia
0
5
10
15
20
25
Japan
Korea
Belgiu
mHong K
ongTai
wan
New Z
eala
ndSin
gaporeUSA
Canad
aAust
ralia
Nether
lands
Norway UK
Icel
and
Swed
enAust
ria
Switz
erla
ndDen
mar
kFin
land
16
40
80
Price of Typical BB Connection As a % ofMonthly Household Income, April 2003
Price Per 100 kbps of Data per Month, US $,April 2003
T. Reynolds, ITU-T Promoting Broadband Workshop, Geneva, April, 2003
%$/
100
kb/s
Price of Broadband Connection
1(0.18)
1.6(0.29)
120(21.2)
8
Reasons for ADSL High-Penetration in Japan?
1. Low Subscriber Charges Stemming From:
● Competitive Environment- Local loop unbundling for ADSL and FTTH- Collocation
● Many ADSL Operators: about 50
2. Higher Speed Systems (up to 24 Mbps)
N.B. No Financial Support from Government
9
Why is FTTH so Cheap in Japan?
1. Low Installation Cost• Ratio of Apartment Inhabitants is High; about
40%.• High population density; 70% of the land is
covered with mountain• Aerial cable is used for the last mile (no digging)
2. Strong Competition• Competition between ADSL and FTTH• Competition among FTTH operators
3. Cost-effective Systems using Ether-based MediaConverter and B-PON.
10
Bandwidth Requirement
DSCQS Value (%)
108642 3
01020304050607080
・TV Quality: > 10 Mbps・ITU-R Broadcast TV Quality: > 8 Mbps・VHS Video Quality: > 4 Mbps
Quality and Necessary bandwidth
Broadcast TV Quality
VHS Video QualityWorse than VHSVideo, but acceptable
Worse than VHSVideo, tolerable limit
ITU-R Broadcast TV Spec.
Required Bandwidth for75 % of Video Samples Mbps
11
Available Speed vs. Transmission Loss for ADSLM
axim
um
Do
wn
stre
am S
pee
d 12
10
8
6
4
2
0706050403020100
Mb/s 12 M type
8M type
1.5M type
25 dB
Transmission Loss between Subscriber and SP building (dB)
12
Flets ADSL 8M Type From http://www.ntt-east.co.jp/flets/misc/adspeed.html
Measured Downstream Speed for 8 Mb/s ADSL
Mea
sure
d D
ow
nst
ream
Sp
eed
Mb/s
0 10 20 30 40 50 60 Transmission Loss between Subscriber and SP building (dB)
・: # of data is 1○: # of data is 2-4●: # of data is >5
Total # of data >4,200
13
1. More than 8-10 Mbps will be required to get TV qualitystreaming video services stably, and much higher isneeded to support SHD video quality services.
2. ADSL provides several mega bit per second only forlimited users who are very near (less than about onekilometer) the service provider’s office. If the price isthe same for all subscribers, unfairness in terms ofavailable bandwidth exists. This unfairness willbecome more and more tangible when bandwidthdemanding services proliferate.
Why FTTH ?
14
P2P Communication
15Data from White Paper 2002, Ministry of Public Management, Home Affairs, Posts and Telecommunications.
Trends in Internet Usage by Purpose
: 2000
: 2002
330%170%
+260%
230%
260%
Multiple Answers Permitted, Excerpt
16
120
100
80
60
40
20
0
Mb/s
Outgoing Traffic from the University of Wisconsin to the Internet
: Napster: Others
1999 May 1999 Sept. 2000 May 2000 Sept.
17
● About 3.4 % of Internet Users (about 99 mill.)● Application Used for File Exchange (multiple answers permitted)
WinMX : 82.4 %
Winny : 22.8 %
Napster : 22.5 %
● Average Number of Downloaded Files per User is 162,
of Which 32.5% were Video.
● 45.1% of them uploaded; Published 124 Files on the
Average
Increase in P2P Traffic
P2P File Exchange Statistics in Japan Data are at June 2003.
Copyright and Moral Issues Need to be Resolved before P2P canBecome More Widely Spread. Increase in Users Will Cause SignificantNetwork Traffic Load.
18
Lessons Learned form P2P
Abrupt Generation of Large Volume of Traffic
Increase of Large File-size Traffic
Different Traffic Patterns(Up-/Down Traffic )
- Instead of 10:1(down:up), just 3(or 2):1
Number or Users Increased with the AvailableSubscriber Access Speed
Usage Increased as HD Storage Capacity withAlways-on Connections Increased (Local Storeand Replay Enhances Usage)
Peer-to-Peer is More Than Sharing Music/Video Files;It Might Create the Way to a New Business Scheme.
• FedEx is already a Terabit network– thousands of disks and tapes shipped daily– jitter and delay is pretty poor– cost for shipping tape approx. 0.000001¢/byte– current cost of sending data over fiber 0.001¢/byte
(1998 CANARIE Inc.)
Photonic Network
・Light speed・Low cost(nighttime)
Transport of package media
24
Photonic Backbone Network
・immediate・low cost
Digital Cinema
QUALCOMM’s Approach Expectation for Digital Cinema - 31,000 screens (USA) - 231 major titles (1997) - Cost for printing, $700 mill. ( $22,400/screen)
Digital Cinema will Spur - increase in smaller theaters - increase in number of copies - world-wide simultaneous release
25
Resolution comparisonHDTV-1080p/24 vs. 4K SHD Digital Cinema with 8M pixels
HDTV-1080p 1920 x 1080 pixels
4Kp-SHD 3840 x 2048 pixels
Standard TV720 x 480
26
Format comparison of Digital Cinema
Polaroid film
35mm film
3024
60mm filmLaser Printer
PrintingMedicine
2000
1000
0
35mmMovie
HDTV1080P720P Standard
TV(480i)
HDTV(1080i)
TV
Still Images Motion Picture
0
3000
SHD Digital Cinema with8M(3840x2048) pixels
Temporalresolution(Frame/sec.)
Spatial resolution(lines)
Legacy media
Horizontal resolution is used for motion pictures. 4k motion picture means= 2000 scanning lines 2k motion picture means = 1000 scanning lines (@HDTV)
4K
2K
27
Digital Cinema Architecture Empowered by Photonic Technologies
Streaming Trials at Internet2TCP/IP based streaming experiments over large-scale, high-speed networks (1) Stable Streaming of SHD Digital Cinema (cope with long delay)
(2) Arbitration of Streaming Traffic by using MXQ (MaXimal Queuing) Mechanism
(3) Wide-Area Multipoint Live Streaming (Flexcast): Connecting 3 sites using GEMnet and Abilene (Internet2)
Yokosuka Japan (NTT Lab.), Chicago, (UIC/EVL,NWU/StarLight), Los Angeles (USC)
SHD Digital Cinema Streaming : From Chicago to L.A (3,000km) 300Mb/s
GbE
OC12/ATM
Abilene
LightReef Z4
LightReef Z4
Cisco SL6509
Traffic Monitor
NTT Server
Application Traffic MonitorDecoderUSC/ Zemekis Center
StarLightNew Route 66
Route 66
Abilene 10 GbE
<20 MbpsSunny Vale
PacificLink
NTT
GEMnet
Cisco 12404
29
Demonstration of SHD Digital Cinema in cooperationwith European Digital Cinema Forum (EDCF)
CineCitta in Rome–Famous Cinema Studio–Ministry of Culture is pushing
4K Digital Cinema in Italy
National Film Theater in London
–Testbed of EDCF is located atNational Film Theater (6/24)
–Support of DTI
30
Advances in TransportNetwork Technologies
31
• Less Service Dependent than Circuit/Flow• Grouped Circuits/Flows Serving as a Unit of Network Operation,
Design and Provisioning, Including Traffic Engineering• Object to Be Rearranged in Node and Transmission
Line/System Failure Restoration
Switch Switch
Cross-connect Path
Optical Fiber/Radio Wave
Optical Fiber/Radio Wave
Circuit
Cross-connect Path
Circuit
Role of Path
32
Switch
Higher-Order Electrical Path Cross-connect
Optical Fiber
Circuit
OpticalFiber
Higher-Order Electrical Path
Lower-Order Electrical Path Cross-Connect
Lower-orderElectrical Path
Hierarchical SDH Path Structure
ServiceAccess
Trans-Access
33
Service 1
(a) Media management by VPs
Transmission Link
Media 1
(c) QoS management by VPs
QoS 1QoS 2QoS 3
(b) Network service management by VPs
Transmission Link
Transmission Link
Media 2Media 3
Service 2Service 3
VP
VP
VP
8 7 6 5 4 3 2 1VPI
VPI VCIVCI
VCI PT CLPHEC
123456
53
OctetBit
Header
Information field
Header structure at NNI
Header Information field
5 octets 48 octets
53 octets
Cell structure
ATM Virtual Path
34
Virtual Path Benefits - Compared to Digital Path in STM -
Simplification of Interface and Node Structure
Simplification of Path Layer Architecture
Simplification of Path Accommodation DesignIn Terms of Path Hierarchy
Network Flexibility Enhancement
35
Hierarchical Path Structure
Switch/RouterFunction
Optical Path Cross-connect
Optical Fiber
Circuit/Flow
Optical Fiber
Optical Path
Electrical PathCross-ConnectFunction
Electrical Path(VP, LSP, DP)
36
Optical PathWP (Wavelength Path) and VWP(Virtual Wavelength Path)
WP 1 WP 2 WP 3 VWP 1 VWP 3VWP 2
VWP 4WP 4
37
1. Enhanced Transmission Capacity With WDM
2. Enhanced Cross-Connect Node Throughput with Wavelength Routing
3. Flexible and Progressive Transport Capacity Enhancement
4. Provision of Transport Platform(Different Degree of Transparency Can be Utilized)
5. Effective Network Protection/Restoration
Advantages of Optical Path
38
Multiplexing and Path Realization Technologies
MultiplexingTechnologies
Path Technologies
Path Identification(# of Paths/Link)
Soft/HardState Routing
PDHSDH
ATM
Packet
WDM
Digital Path(VC-1n, VC-3/4)
VP
LSP
OpticalPath
Time Position inthe TDM Frame
(< 192)
Cell header (VPI)(< 4096; NNI< 128; UNI)
Sim Label(<220)
Wavelength(<1,000)
Hard
Hard
Soft
Soft
Store-&-ForwardElectrical Processing
+Space Switch
Waveguide Router (Self-Routing)
and/orSpace Switch
Time SlotInterchange
+Space Switch
Store-&-ForwardElectrical Processing
+Space Switch
39
Label-Switching Router (LSR)
MPLS
Label Switched Path (LSP)
Egress LSRIngress LSRMPLS domain
IP Router
IP packet isencapsulated.
Label is swapped.
MPLS integrates IP and data-link layer technologies.
Removes theMPLS header.
(LabelEdgeRouter)
(Label EdgeRouter)
(LabelSwitchRouter)
40
The initial MPLS effort will be focused on IPv4. However, the coretechnology will be extendible to multiple network layer protocols(e.g., Ipv6, IPX, Appletalk, DECnet, CLNP). MPLS is not confinedto any specific link layer technology, it can work with any mediaover which Network Layer packets can be passed between networklayer entities.
MPLS provides connection-oriented (label based) switching basedon IP routing and control protocols. MPLS may be likened to a'shim-layer' which is used to provide connection services to IP andwhich itself makes use of link-layer services from L2 (e.g. PPP,ATM, Ethernet).
MPLS
A Framework for Multiprotocol Label Switching <draft-ietf-mpls-framework-05.txt> September 1999
41
Layer 3
Layer 2
IP Header
ATM
IP
VPI/VCI DLCI Shim Label
Frame Relay Ethernet
Shim Label
PPP
MPLS Labels; Label stacking and support of various media
MPLS is intended to run over multiple link layersSpecifications for the following link layers currently exist:
ATM: label contained in VCI/VPI field of ATM headerFrame Relay: label contained in DLCI field in FRheaderPPP/LAN: uses ‘shim’ header inserted between L2and L3 headers
42
MPLS Header
Layer 2 Header MPLS Header IP Packet
Label (20 bits) EXp (3) S (1) TTL (8 bits)
Label; 20 bits (4 bits for indicating how to handle labels, 16 bits for indicating FEC)Experimental (was CoS, class of service); 3 bitsStacking bit; 1 bit (indidcates the presence of a label stack; 1=last entry in label stack)Time to Live; 8 bits (same functionality as IP TTL; used to through away looping packets)
4 Bytes
★ short and fixed length★ associate to Forward Equivalent Class (FEC)
Comparison of IP over ATM and MPLS (IP+ATM)IP over ATM MPLS
}Structure
Mapping ofIP Packetinto Layer 1
IP Router ATM-SWIP Router PartMPLS-SW Part
MPLS-Router
IP Packet
Segmentationinto ATM Cell
Accommodation into SDH Frame
IP PacketLabel AddedPPP/HDLC Processing
Accommodation into SDH Frame(Ethernet transmission is also possible w/oPPP/HDLC processing)
Characteristics
NetworkOperation
- VPI(/VCI) is swapped link by link- SAR Burden- Meshed VP connection between routers (Limited network expandability)
- Label swapping (link by link)- Variable length packet processing at MPLS-SW- Hierarchical LSPs with label stacking
- Separate (independent) IP and ATM layer operation (ATM will be common transport platform to other services than IP)
- Integrated IP and MPLS layer operation- Operation on different label switched path realization techniques- Standardized MPLS signaling protocols
45
☆Traffic Engineering (Path oriented)- Possibility to set-up other paths than “shortest paths”- Multiple paths between two points: Load sharing,
1 + 1, 1 : N, M : N Protection☆Very flexible because of de-coupling of forwarding and routing (forwarding decision is separated from routing
process)☆Support capabilities of VPNs and new service provisioning
- by allowing the forwarding infrastructure to remain the same while new services are built through the assignment of packets to an LSP
☆Enahanced QoS
Why MPLS?● MPLS helps scaling pubic IP networks and enhance network performances● MPLS provides network providers with means that can differentiate their services from others.
46
2
Routing G-bit Networking Tera/Peta-bit Networking
IP
ATM
etc
.SD
H
Elec
trica
lO
ptic
al
Wav
e-le
ngth
Routers basedon softwarerouting
Routers basedon hardwarerouting(ASIC)
Tbit Routers(IP v6,
HierarchicalAddressing etc.)
Routing isdone withIP only(RouterMulti-hop)
Laye
r 2La
yer 1
PhotonicMPLS
Node ThroughputEnhancement
TrafficEngineering(QoSguarantee)
IP over Optical Path
IP over ATMIntroduction ofUnderlyingTransferMechanismwhich enableseffective trafficengineeringMesh-likeconnectionis possible(Routersingle hop)
MPLS: Multi Protocol Label Switching
MPLS
Router Throughput IncreaseIP over SDH IP over WDM (SDH)
Laye
r 3
Enhancement of Networking Function
Evolution of IP transport mechanism
47
MPLS and Photonic MPLSMPLS Router
Ingress Egress
Label Switch
Ingress Egress
Photonic Router
MPLS
Photonic MPLS
l Label is added to each packet.
l Wavelength label is added to each layer 1 stream.
WP approachVWP approachWavelength Label
Optical Label Switch
Labeled PacketLabeled Packet
Label
IP Packet IP Packet
IP PacketIP Packet
48
MPLS and Photonic MPLS Cont..
IP packet
MPLS Router
IP packet
Ingress
Labeled Packet
IP packet
Egress
IP packet
Ingress Egress
Photonic Router
MPLS
Photonic MPLSl Label merge function can be realized.
l Label merge is difficult. Label stack is also difficult.
IP packet
IP packet
Labeled Packet
49
MPLS and Photonic MPLSLSP 1
LSP 2
LSP 3
LSP 1
LSP 2 OLSP 1
MPLS-Router
Photonic MPLS-Router
OLSP 1
LSP1 and LSP 2 are accommodated within LSP 3.
LSP1 and LSP 2 are accommodated within OLSP1.
MPLS-Router MPLS-router pert multiplexes LSP1 and
LSP2 and connects to OLSP1.
50
Comparison of Electrical MPLS and Photonic MPLS
Electrical MPLS Photonic MPLS
Path Label Switched Path(Label is attached to each packet)
Hit-less RouteChange Yes (Make-before-break) No (possible only at electrical level)
51
PTSPhotonic (MPLS) Plane
IP (MPLS) Router
Photonic (MPLS) Router
MPLS Plane
AdministrationDomain
Photonic MPLS and MPLS hierarchical architecture
52
Best Effort(Connectionless) Engineered QoS
(Connectionless on Connection Oriented)
Guaranteed QoS(Connection Oriented)
ATM XC/Switch
MPLS RouterIP Router
Big IP RouterBig MPLS
Router
Photonic MPLS Router
Thro
ughp
ut (N
etwo
rk S
cala
bility
)
Photonic MPLS Router
IP over ATM
OXC/OADM
Digital XC
53
OXCs and Photonic MPLS routers co-exist?
Yes, each has its own application.
Photonic MPLS routers are regarded as the ultimate inintegrated router systems.OXCs create optical platform on which different transfermode services can be provided.
Criteria of the Selection - Services (IP only or IP and other services, ex. Lambda service?) - Ownership (Are IP routers and OXCs and Transmission Equipment owned by the same provider?) - Segmentation of Network Management
54
This document describes extensions to MPLS signalingrequired to support Generalized MPLS. GeneralizedMPLS extends MPLS to encompass time-division (e.g.SONET ADMs), wavelength (optical lambdas) andspatial switching (e.g. incoming port or fiber to outgoingport or fiber). This document presents a functionaldescription of the extensions.
Enhancement of MPLS for GMPLSSupport different types of switches; packet-switchcapable, layer 2-switch capable, TDM capable,lambda-switch capable, fiber-switch capable.
A new Link Management Protocol (LMP) that runsbetween adjacent nodes and is used for both linkprovisioning and fault isolation.
Enhancement of OSPF/IS-IS routing protocols toadvertise availability of optical (and other)resources in the network.
Enhancements to RSVP/CR-LDP signaling protocolsto allow a label-switched path to explicitly specifiedacross the optical core.
Evolution of Photonic NetworkTransport ofInformation Block withWavelength Label
Distributed Controlnetwork withPhotonic MPLS
CentralizedControl Mesh-type Networkwith OXC’S
OADM RingNetwork
Point-to-Point WDMTransmission
PhotonicMPLS Router
IP Routing
WavelengthRouting
IP Packets aremapped withinWavelengthLabeled BitStream
Photonic Router
Dyna
mic
Con
trol o
f Wav
elen
gth
Stat
ic W
avel
engt
hO
pera
tion
80
Photons and Electrons are Very Different
Both Photons and Electrons have the nature of a particleand of a wave, but they are very different.
Photons usually behave like a wave. - Photon has no mass. - Photon has no charge.Electrons usually behave like a particle. - Electron has mass. - Electron has charge.
A crucial difference is the degree of susceptibility inregard to interacting with others.There is no basic optical functional device equivalent tothe transistor.
81
Characteristics of Photons and Electrons
Is it possible to bend the direction of photon propagation in free space?
Yes, but huge mass is necessary: The mass of the sun can bend light by 1.74second; Einstein’s Theory of Relativity.
1.74’’
If dielectric constant is changed, direction of light propagation can be changed.
Dielectric constant can be changed and controlled by using, - mirror and mechanical means - electro-optic effect - thermo-optic effect, etc.
82
Characteristics of Photons and ElectronsIs it possible to confine/store photon?
Yes, but huge mass -40 million times that of the sun- is necessary to create aBlack Hole. But can not get photon back out of black hole.
No Optical Capacitor.
Is it possible to slow down photon propagation speed?
Yes, but at almost zero degree.
300,000 km/sec at normal temperature
~0 km/sec at almost 0 Kelvin
83
0.5 mm5 mm
10 mm
40 k cell/chip1 cell @ 10 Gb/s
8.5 m
v=6x10-4 mm3/cell
V=100 mm3/cell >105 v
Bare fiberStraight Line
Bare fiberCoiled 6 cm
125 mmf
V= 16000 mm3/cell >107 v
Semiconductor Memory(S-RAM)
Optical Fiber Delay Line
Volume of Memory
Si Memory Chip
84
10
100
1,000
10,000
100,000
1994
1995
1996
1997
1998
1999
2000 2001
10,000
100,000
1,000,000
155
620
2500Ce
ll Bu
ffer p
er 1
50 M
b/s
Path
Spe
ed
Cell
Buffe
r per
Lin
k
Year Link Speed (Mb/s)
A
B
(Not a technical limit)
Available Cell Buffer per Link in ATM Systems
85
IP over ATM
ATM XC WDM LT
MPLS (IP plus ATM)WDM LT
IP over Optical Path
WDM LTOXC
Photonic MPLS
WDM LT
IP over Digital Path
Scalability Global address IP signaling (distributed control)
Traffic engineering(Introduction of paths)Node throughputenhancement (Cut-through)
8 input /output pairs (fiber port canbe added one by one)
256 x 256 channels
T h e n um ber of av a ilable o p t i c a lchannels is expandable up to 256, with 8wavelengths’ modularity (each switchmodule accommodates 8 wavelengths.)
Maximum number of available POSinterface is 128. Consists of one totwenty MPLS routers.
Page 99
Part II : Photonic Network Control and Management
• Photonic Network Control and Management Overview
• Photonic Network Architecture Overview• IP over Photonic Network Architecture
Overview• MPLS, GMPLS (MPLambdaS), and ASON• GMPLS protocols• GMPLS management• Interoperability Test Events of Photonic
Network Control
Page 100
Outline
• Photonic Network Control and Management Overview– Target Applications– Standardization Partner-ship
• Photonic Network Architecture Overview• IP over Photonic Network Architecture Overview• MPLS, GMPLS (MPLambdaS), and ASON• GMPLS protocols• GMPLS management• Interoperability Test Events of Photonic Network
Control
Page 101
Target Photonic Network Applications
• Traditional Photonic Network(for Telephone Network Backbone)
– Provisioning Base• Centralized NMS base
– Centralized Control Plane (C-Plane)– Centralized Management (M-Plane)
– Routing/Topology Dissemination• Network topology and resource availability
– Connection Signaling• Automated provisioning and failure recovery
• Concepts endorsed by every standards body– ITU-T, IETF and OIF
Page 107
Standardization Bodies and Forums
ITUITU International Telecommunication UnionInternational Telecommunication UnionIETFIETF Internet Engineering Task ForceInternet Engineering Task ForceOIFOIF Optical Internetworking ForumOptical Internetworking ForumTMFTMF TeleManagementTeleManagement ForumForum
ASONASON MPLS, MPLS, GMPLSGMPLS
Optical UNI/NNIOptical UNI/NNI
Focused in this tutorial
NMSNMS
Page 108
• Different focus– ITU focuses on architecture
• ASON architecture– IETF focuses on building blocks
• GMPLS protocol specs.– OIF focuses on applications and interoperability
• Implementation Agreements (IAs) and Interoperability Test Events
• Common goal: better photonic networking • Recognized need for coordination
ITU vs. IETF vs. OIF
L. Ong, “Optical Control Plane Activities in IETF and OIF”, ITU-T Workshop on IP/Optical Chitose, Japan, 9-11 July 2002
Page 109
Major Work Areas of ITU (ITU-T SG15)
• Optical Transport Network (OTN) structure – Automatic Switched Optical Network (ASON)– Architecture and interfaces for the OTN– Optical (Photonic) Cross-Connect and Switch functions– Network management and control
• OTN technology (terrestrial and submarine)– Coarse and dense WDM, STM-256 (40Gb/s) signal
channels– Optical components & amplifiers (e.g. tunable filters)– Fiber characteristics, more channels/fiber– Transmission technology (Soliton/RZ), long reach
Page 110
OTN Technology Standardization Work Plan
• Standardization areas covered– OTN Technologies (variety of aspects)– SDH & SONET
• Next Generation data centric transport network– Generic Framing Procedure (GFP)– Virtual Concatenation (VCAT)– Link Capacity Adjustment Scheme (LCAS)
– OTN Transport Plane (D-Plane)– ASON Control Plane (C-Plane)
Page 111
• IETF’s Traditional Focus– The Internet: IP and IP Services – routing,
transport, applications, security & management• Sub-IP Area
– Coordinates activities below the IP layer, esp. MPLS/GMPLS and IP over Optical.
IETF Optical Standards
IESGIESG
Transport RoutingApp., General, Internet, OAM,
SecuritySub-IP
sigtran pwe3 ospf idr ipo ccamp
ipo : IP over Opticalccamp : Common Control and
Measurement Plane
IESG : The Internet Engineering Steering Group
Areas
WGs
Page 112
IETF GMPLS: History
• How did GMPLS start?– Outgrowth of MPLS - IP traffic engineering work – “Generalized” protocols for label-switched path
creation• Fiber switching• Wavelength/Waveband switching• Time slot switching (SDH/SONET)
• Possible Architectures – Flat network – routers and optical systems fully
peered (Peer Model)– Hierarchical network – routers are optical clients
(Overlay Model)• Scope
– Support of IP networking over optical transport– Non-IP-related use of GMPLS is out-of-scope
Page 113
Optical Internetworking Forum (OIF)
• The only industry group bringing together professionals from the data and optical worlds.
• Open forum– international– carriers– component and systems vendors– testing and software companies
• Launched in April of 1998 with an objective to foster development of low-cost and scaleable internet using optical technologies.
• Mission: To foster the development and deployment of interoperable products and services for data switching and routing using optical networking technologies.
VC Virtual ContainerSTM Synchronous Transport Module
Page 118
SDH Standardization
Network Architecture(G.803, G.805)
Structures and Mappings(G.707)
Physical Layer(G.957, G.691)
Equipment Functional Spec.(G.783, G.806)
Equipment Management(G.784, G.7710)
Information Model(G.774 Series)
Protection Switching(G.gps, G.841, G.842)
Laser Safety(G.664)
Data and Signaling Communications Network
(G.7712)
Jitter and Wander Perf.(G.825)
Error Performance(G.826-829)
Ghani Abbas and Stephen Trowbridge, “OTN Equipment and Deployment in Today’s Transport Networks”, ITU-T Workshop on IP/Optical Chitose, Japan, 9-11 July 2002
Page 119
Optical Transport Network (OTN)
• Metro Access• SDH metro ring applications• Multi-Service Provisioning Nodes - combining data and SDH
• Metro Core• SDH ADM metro ring and mesh application• Optical add/drop multiplexers (proprietary)
• Long Haul/Ultra Long haul• SDH ADM rings and line systems• DWDM line systems (proprietary)
TodayToday’’s Transport Network Environments Transport Network Environment
New Generation Transport Network is required New Generation Transport Network is required OTN
Page 120
OTN Requirements
• Functionality as that offered by SDH or better• Transparent transport of SDH and other payloads• Stronger FEC (forward error correction)
ITU-T G.709 is the answerG.709 defines the interfaces for the OTN
New Transport Network Technologies are called OTN.New Transport Network Technologies are called OTN.So, old fashioned SDH/SONET are called PreSo, old fashioned SDH/SONET are called Pre--OTN.OTN.
Page 121
Optical Transport Network (OTN) Standardization
Network Architecture(G.872)
Structures and Mappings(G.709)
Physical Layer(G.692, G.959.1, G.694.x)
Equipment Functional Spec.(G.798, G.806)
Equipment Management(G.874, G.7710)
Information Model(G.874.1, G.875)
Protection Switching(G.gps, G.otnprot)
Laser Safety(G.664)
Data and Signaling Communications Network
(G.7712)
Jitter and Wander Perf.(G.8251)
Error Performance(G.optperf)
Ghani Abbas and Stephen Trowbridge, “OTN Equipment and Deployment in Today’s Transport Networks”, ITU-T Workshop on IP/Optical Chitose, Japan, 9-11 July 2002
GMPLS protocols are commonly used in IETF, ITU-T and OIF.
Page 129
Press Release (4 December 2001)
The 40Gigabit per Second Phone Call:
Global Standards for Automatically Switched Optical Networks Enable New Market Services
Page 130
Drive to Automatically Switched Network
• Make the network intelligent• On-demand bandwidth to the edge of the
network• New applications
– Disaster Recovery – Distributed SAN– Data warehousing
• Backup Bunkers (no more tapes)– Big Pipes on Demand
• Download movies to movie theaters• Site replication
– Optical VPN – Grid Computing
Page 131
• Different focus– ITU focuses on architecture
• ASON architecture– IETF focuses on building blocks
• GMPLS protocol specs to ASON.– OIF focuses on applications and
interoperability• Implementation Agreements (IAs) and
Interoperability Test Events to ASON.
Reminder (ITU vs. IETF vs. OIF)
Page 132
IP, MPLS and GMPLS
1. IP: Shortest Path takes all packets
2. MPLS: Traffic Engineering allows flows to be mapped to different paths for better utilization
3. GMPLS: MPLS control protocols could also set up connections in a circuit network
L. Ong, “Optical Control Plane Activities in IETF and OIF”, ITU-T Workshop on IP/Optical Chitose, Japan, 9-11 July 2002
Page 133
MPLambdaS (MPλS) and GMPLS
• Multi-Protocol Lambda Switching: October 1999– Combining MPLS Traffic Engineering Control With Optical
Crossconnects (OXCs)1. provide a framework for real-time provisioning of optical channels in
automatically switched optical networks2. foster the expedited development and deployment of a new class of
versatile OXCs3. allow the use of uniform semantics for network management and
operations control in hybrid networks consisting of OXCs and labelswitching routers (LSRs).
– The proposed approach is particularly advantageous for OXCsintended for data-centric optical internetworking systems.
• This concept can be extent to other circuit switch systems.– Attempt to SDH control: March 2000– Generalized Multi-protocol Label Switching (GMPLS): June 2000
Ayan Banerjee, et al., “Generalized Multiprotocol Label Switching: An Overview of Signaling Enhancements andRecovery Techniques”IEEE Communication Magazine, Vol. 39, No. 7, July 2001.
Packet (PSC) LSP1
λ (LSC) LSP3
Fiber (FSC) LSP4
Time slot (TDM) LSP2
LSC
FSC
Page 138
Hierarchical LSP and Signaling (2/2)
Ayan Banerjee, et al., “Generalized Multiprotocol Label Switching: An Overview of Signaling Enhancements andRecovery Techniques”IEEE Communication Magazine, Vol. 39, No. 7, July 2001.
PSC LSP
TDM LSP
LSC LSP
FSC LSP
GMPLS RSVP-TE signaling
Page 139
OIF OUNI and IETF GMPLS Inter-working
OUNI Resv
ResvConf
ResvConf+ MESSAGE_ID_ACK
ACK
UNI Transport Connection EstablishedSource UNI-C may start transmitting
Destination UNI-C may start transmitting
ACK
R1 XC2XC1
OUNI Path
ACK S=R1, D=R2
S=R1, D=R2
I-NNI Path
OUNI Resv+ MESSAGE_ID_ACK
OUNI Path
ACK
S=R1, D=R2
I-NNI ResvMESSAGE_ID
MESSAGE_IDI-NNI ResvConf
R2OUNI OUNIGMPLS
Page 140
UNI - E-NNI Inter-working
Path
Path
Rev
Resv+ MESSAGE_ID_ACK
ACK
NNI Transport Connection EstablishedSource NNI may start transmitting
ACK
Source UNI-C Destination UNI-CUNI-N I-NNI
ResvConf
ResvConf+ MESSAGE_ID_ACK
ACK
Destination NNI may start transmitting
PathACK
ACKResv
ACKResvConf
I-NNI E-NNI E-NNI I-NNI I-NNI UNI-N
Vendor Vendor AA Vendor Vendor BB
Page 141
GMPLS Management (MIBs)
• Link Management Protocol Management Information Base– draft-ietf-ccamp-lmp-mib-06.txt
*1 - Not enough time to complete testing*2 - Proxy or test device supports control plane only.H = Another client was used to form Hetrogeneous testResults scrambled to protect individual companiesTested, Signaling & Transport success
ONE Device
Clie
nt D
evic
e
Color KeyNot Interoperable
Tested, Signaling success
•Over 196 Tests Conducted
99.5% Signaling Inter-working
68.6% Signaling & Transport Inter-working
Page 145
OFC2003 UNI/E-NNI demo
• 12 vendors, 15 systems.• UNI GMPLS E-NNI GMPLS UNI
Page 146
GMPLS Interoperability demonstration
• Held at UNH Interoperability Lab• Staging for GMPLS demo at NGN 2002• Organized by The MPLS Forum• Participants