Agenda

1
Agenda

• History of IP Backbones
• The Emerging Two Layer Network
• Network Platforms
• Standards and Forums
• GMPLS

2

• The Emerging Networking Paradigm  
• The change in the fundamental character of
  backbone network traffic, as demonstrated by the
  current shift in the telecommunications industry
  from traditional voice-centric TDM/SONET
  circuit-switched networking paradigm to
  data-centric packet-optimized networking
  paradigm, is leading to revolutionary changes in
  the traditional concepts of how networks are
  constructed.
• The primary reason is the nature of traffic
  crossing today's long-haul backbones.
• Internet Protocol (1P) applications are the
  fastest growing segment of a service provider's
  network traffic..

3
Typical IP Backbone (Late 1990s)

• Data piggybacked over traditional voice/TDM
  transport

4
(No Transcript)
5
Why So Many Layers?

• Router
• Packet switching
• Multiplexing and statistical gain
• Any-to-any connections
• Restoration (several seconds)
• ATM/Frame switches
• Hardware forwarding
• Traffic engineering
• Restoration (sub-second)

• MUX
• Speed match router/ switch interfaces to
  transmission network
• SONET/SDH
• Time division multiplexing (TDM)
• Fault isolation
• Restoration (50mSeconds)
• DWDM
• Raw bandwidth
• Defer new construction

• Result
• More vendor integration
• Multiple NM Systems
• Increased capital and operational costs

6
IP Backbone Evolution
Core Router (IP/MPLS)

• MUX becomes redundant
• IP trunk requirements reach SDH aggregate
  levels
• Next generation routers include high speed
  SONET/SDH interfaces

FR/ATM Switch
MUX
SONET/SDH
DWDM (Maybe)
7
IP Backbone Evolution
Core Router (IP/MPLS)

• Removal of ATM Layer
• Next generation routers provide trunk speeds
• Multi-protocol Label Switching (MPLS) on routers
  provides traffic engineering

FR/ATM Switch
MUX
SONET/SDH
DWDM (Maybe)
8
Removing the ATM Layer
Logical Topology

• Why Remove ATM?
• Two networks to manage - IP and ATM
• Cell tax
• Lack of high-speed SAR interfaces
• High density of virtual circuits
• IP routing protocol stress

9
Collapsing Into Two Layers
IP Service (Routers)
Optical Core
Optical Transport (OXCs, WDMs, SONET ?)
10
Collapsing Into Two Layers
IP Service (Routers)
Optical Core
Optical Transport (OXCs, WDMs, SONET ?)

• IP router layer functions
• Service creation
• Multiplexing and statistical gain
• Any-to-any connections
• Traffic engineering
• Restoration (10s ms)
• Subscriber to transport speed matching
• Delay bandwidth buffering and congestion control
• Internet scalability

11
Collapsing Into Two Layers
IP Service (Routers)
Optical Core
Optical Transport (OXCs, WDMs, SONET ?)

• Optical transport layer functions
• TDM and standard framing format
• Fault isolation and sectioning
• Restoration (10s ms)
• Survivability
• Cost efficient transport of massive bandwidth
  (DWDM)
• Long haul transmission distances
• Metro transmission distances ????

12
The Emerging Two-Layer Network
Data Layer
Routers
IP Services
Transport Layer
OXCs
TDMs
WDMs
LH Transport

• Reduced complexity
• Network more scalable
• Uniform admin management of IP and transport
  layers

• Reduced cost
• Transport layer visible to IP Services
• Transport layer signaling is an open standard
  (RSVP CR-LDP)

13
SONET/SDH Benefits

• Rapid and predictable restoration
• 10s of ms depends on ring size
• Simple to engineer
• Standard framing and multiplexing (Time Division
  Multiplexing TDM)
• Maintainability
• Performance monitoring
• Fault isolation and sectioning
• Bandwidth management
• Network management
• Transparency
• Voice, video or data traffic
• Challenge
• Remove complexity and keep benefits

14
SONET/SDH Benefits

• Rapid and predictable restoration
• 10s of ms depends on ring size
• Simple to engineer
• Standard framing and multiplexing (Time Division
  Multiplexing TDM)
• Maintainability
• Performance monitoring
• Fault isolation and sectioning
• Bandwidth management
• Network management
• Transparency
• Voice, video or data traffic
• Challenge
• Remove complexity and keep benefits

TrafficQuickly ReroutedAfter Failure
15
SONET/SDH Limitations

• Traditional SONET/SDH-based networks
• Engineered for voice, not data
• Slow to provision
• Planning complexity
• Grooming complexity
• Delivery measured in weeks
• Expensive to scale
• Space, power, one wavelength per chassis
• Inflexible
• Static not dynamic bandwidth
• Granularity why not 5.5Gbps ?
• Little interoperability at control plane
• Customers forced to buy from one vendor
• Stifles best-in-class deployment
• Packet layer no visibility into optical layer

16
What is an IP Router?
A Device Which Moves IP Datagrams Across an
Internetwork From Source to Destination

• Minimum qualifications
• Capable of switching IP datagrams L3 forwarding
• Symmetric any-port-to-any-port switching speed
• Delay-bandwidth buffering, plus congestion
  control
• Internet scale IS-IS, OSPF, MPLS, BGP4
• Todays benchmark
• Wire-rate forwarding on all ports for 40-byte
  packets
• Performance independent of load
• Support of CoS queuing, shaping, and policing
• Traffic engineering
• Classification and filtering at wire rate

ISO 7 Layer Model
7 - Application
6 - Presentation
5 - Session
4 - Transport
3 - Network
2 - Datalink
1 - Physical
17
What is an IP Router?
Routing Algorithm Goals

• Optimal routes
• Calculate and select the best routes many
  methods
• Simplicity
• Functional efficiency with low routing protocol
  overhead
• Robust and stable
• Predictable and correct functionality in a
  variable environment (hardware failure, high
  load, topology changes)
• Rapid convergence
• Slow route calculations cause loops and drops in
  service
• Flexibility
• Speed accuracy to adapt to network changes
  (bandwidth, delays, queues, traffic levels, etc.)

ISO 7 Layer Model
7 - Application
6 - Presentation
5 - Session
4 - Transport
3 - Network
2 - Datalink
1 - Physical
18
What is an IP Router?
IP Service Creation

• Any-to-any connectivity
• Internet scale routing allows anyone to connect
  to anyone (within or outside of own company)
• Applications
• Processing granularity to differentiate HTML from
  FTP
• Multicast
• Not possible with voice circuit switching
  technology
• Internet radio, video on demand, push Web
• Content sites
• Directing Web traffic
• Complementing cache servers
• Security

ISO 7 Layer Model
7 - Application
6 - Presentation
5 - Session
4 - Transport
3 - Network
2 - Datalink
1 - Physical
19
Optical Cross-connects (OEO)
SONET/SDH Digital Cross-connect (DXC) Also known
as DigitalCross-connect Switch (DCS)
DXC/DCS
20
Optical Cross-connects (OEO)
SONET/SDH Digital Cross-connect (DXC) Also known
as DigitalCross-connect Switch (DCS)
Electrical Switch Matrix
21
All Optical Cross-connects (OOO)
All Optical Cross-connect (OXC) Also known as
PhotonicCross-connect (PXC)
OXC/PXC
22
All Optical Cross-connects (OOO)
All Optical Cross-connect (OXC) Also known as
PhotonicCross-connect (PXC)
Optical Switch Fabric
23
What is an OpticalCross-connect?

• Connects one port (l) to another port
• Add/Drop function with certain l
• Delivers high bandwidth
• Quick to provision bandwidth

ISO 7 Layer Model
7 - Application
6 - Presentation
5 - Session
4 - Transport
3 - Network
l1
l2
2 - Datalink
1 - Physical
l2
l1
24
OXC/PXC Switching Mechanisms

• Micro-electrical Mechanical Systems
• MEMs
• Used for many other applications
• From Lucent, Corning, Xros (Nortel), and others
• Currently 8 x 8 OXC
• 256 mirrors, long-term goal 1,024
• OXC
• ADM uses seesaw MEMS
• Electrical controls
• Voltage applied to mirror tilts on 2 axis or
  6 degrees
• Switch times typically 10 to 25 ms

25
OXC/PXC Switching Mechanisms

• Liquid Crystal Light Valves
• From Spectra Switch and Chorumtechnologies
• Switch speed sub-millisecond
• Future switch speed in nanosecond
• 1 x 2 port switch
• 2 x 2 Add/Drop
• Electrical controls

Liquid Crystal Cell
ON
Output 1
Input
Polarizing Beam Splitter
Polarizing Beam Splitter
Liquid Crystal Cell
26
OXC/PXC Switching Mechanisms

• Liquid Crystal Light Valves
• From Spectra Switch and Chorumtechnologies
• Switch speed sub-millisecond
• Future switch speed in nanosecond
• 1 x 2 port switch
• 2 x 2 Add/Drop
• Electrical controls

Liquid Crystal Cell
Input
Polarizing Beam Splitter
Polarizing Beam Splitter
Output 2
OFF
Liquid Crystal Cell
27
OXC/PXC Switching Mechanisms

• Bubbles
• From Agilent
• 32 x 32 or dual 16 x 32 ports
• Suitable for
• Wavelength Interchange Cross-connect (WIXC)
• Wavelength Selective Cross-connect (WSXC)
• Optical Add/Drop Multiplexers (OADM)
• Inkjet Silica Planar Lightwave Circuitry
• Electrical controls
• Bubbles created by heating index matching fluid
• Switch times under 10 ms

28
Developing an All OpticalPacket Router

• Needs
• How do you read a photonic header?
• The pipeline approach?
• Switching and logic
• Current technology not fast enough
• Lithium Niobate devices have speed, but too much
  crosstalk
• Photonic Bandgap Devices (optical equivalent to
  transistor)
• Buffering/Memory
• Optical buffers (fixed loop delays) exist, but
  are insufficient
• Bi-stable lasers
• Holographic memories
• SEEDS (Self Electro-optic Effect Devices)

29
Operational ApproachesOverlay and Peer Models

• Overlay model
• Two independent control planes
• IP/MPLS routing
• Optical domain routing
• Router is client of optical domain
• Optical topology invisible to routers
• Routing protocol stress scaling issues
• Does this look familiar?
• Peer model
• Single integrated control plane
• Router and optical switches are peers
• Optical topology is visible to routers
• Similar to IP/MPLS model

?
30
Operational ApproachesThe Hybrid Model

• Hybrid model
• Combines peer Overlay
• Middle ground of 2 extremes
• Benefits of both models
• Multi admin domain support
• Derived from overlay model
• Multiple technologies within domain
• Derived from peer model

31
Standards and Industry Forums

• Optical Internetworking Forum (OIF)
• Industry forum
• Kick-off meeting May 1998
• Standard OIF UNI based on IETF work
  (CR-LDP/RSVP)
• Internet Engineering Task Force (IETF)
• Driving GMPLS standards development
• Initial application was MPlambdaS
• Peer model and Hybrid model
• Extend MPLS traffic engineering to the optical
  control plane
• Rapid provisioning
• Efficient restoration
• ITU-T
• Study Group 13
• Study Group 15

32
IETF

• GMPLS now Hosted by CCAMP WG
• Common Control And Measurement Plane
• MPLS WG revised charter (without GMPLS)
• Eleven main GMPLS building blocks
• Internet Drafts
• Current work includes extending existing control
  protocols (example, OSPF ISIS)
• New future extensions considered
• BGP4
• For cross AS, and Carrier of Carriers
  applications
• LCAS
• Link Capacity Adjustment Scheme protocol for
  SONET
• SONET Virtual Concatenation (dynamic TDM circuit
  control)
• Intent to submit work to ITU-T

33
ITU-T

• Study Group 13 (SG13)
• Focus Multi-protocol IP-based networks their
  inter-working
• Study Group 15 (SG15)
• Focus Optical other transport networks
• G.ASON Automatically Switched Optical Network
• Addresses the control layer for intelligent
  optical networks
• Ambition to reference IETF standards

34
OIF Optical UNI Signaling
IETF-GMPLS
OIF-UNI
UNI
UNI
UNI
UNI
Optical Transmission Network
UNI
UNI

• Uses procedures and messages defined for MPLS
  traffic engineering and GMPLS
• Features
• Runs in UNI-only mode (overlay model)
• Optical path creation, modification, and deletion
• Optical path status inquiry and response
• Allows one protocol to support two different
  applications
• OIF UNI client bandwidth requests (hide optical
  topology)
• GMPLS service provider provisioning (expose
  optical topology)

35
Traditional MPLS Applications
Traffic Engineering
Source
Destination
Layer 3 Routing
Traffic Engineered LSP
VPNs
PE
PE
CPE
CPE
FT/VRF
FT/VRF
P
Site 1
Site 3
FT/VRF
CPE
CPE
P
Site 2
Site 2
P
P
CPE
CPE
FT/VRF
Site 1
Site 3
P
FT/VRS
FT/VRF
FT/VRF
PE
PE
36
Generalized MPLS (GMPLS)

• Traditional MPLS supports packet cell switching
  
• Extends MPLS to support multiple switching types
• TDM switching (SDH/SONET)
• Wavelength switching (Lambda)
• Physical port switching (Fiber)
• Peer model
• Uses existing and evolving technology
• Facilitates parallel evolution in the IP and
  optical transmission domains
• Enhances service provider revenues
• New service creation
• Faster provisioning
• Operational efficiencies

37
GMPLS Mechanisms

• IGP extensions
• Forwarding adjacency
• LSP hierarchy
• Constraint-based routing
• Signaling extensions
• Link Management Protocol (LMP)
• Link bundling

38
IGP Extensions

• OSPF and IS-IS extensions
• Flood topology information among IP routers and
  OXCs
• New link types
• Normal link (packet)
• Non-packet link (TDM, l, or fiber)
• Forwarding adjacency (FA-LSP)

39
IGP Extensions

• OSPF and IS-IS extensions
• Flood topology information among IP routers and
  OXCs
• New link types
• Normal link (packet)
• Non-packet link (TDM, l, or fiber)
• Forwarding adjacency (FA-LSP)

40
IGP Extensions

• New Link Type sub-TLVs
• Link protection
• Protection capability
• Attributes
• None, 11, 1N, or ring
• Priority for a working channel

11 Protection
13 Protection
41
IGP Extensions

• New Link Type sub-TLVs
• Link descriptor
• Characteristics of the link
• Selected attributes
• Link type
• SONET, SDH, clear, Gig E, 10 Gig E
• Minimum reservable bandwidth
• Maximum reservable bandwidth
• Attributes change over time
• Provides a new constraint for LSP calculation
• Shared Risk Link Group (SRLG)
• List of the links SRLGs
• Does not change over time

42
Forwarding Adjacency
SONET/SDH ADM
SONET/SDH ADM
FA-LSP
ATM Switch
ATM Switch

• A node can advertise an LSP into the IGP
• Establishes LSP using RSVP/CR-LDP signaling
• IGP floods FA-LSP
• Link state database maintains conventional links
  and FA-LSPs
• A second node wanting to create an LSP can use an
  FA-LSP as alinkin the path for a new, lower
  order LSP
• The second node uses RSVP/CR-LDP to establish
  label bindings for the lower order LSP

43
Forwarding Adjacency
SONET/SDH ADM
SONET/SDH ADM
FA-LSP
Ingress Node (High Order LSP)
Egress Node (High Order LSP)
ATM Switch
ATM Switch

• IGP attributes describing a forwarding adjacency
• Local (ingress) and remote (egress) interface IP
  addresses
• Traffic engineering metric
• Maximum reservable bandwidth
• Unreserved bandwidth
• Resource class/color (administrative groups)
• Link multiplexing capability (packet, TDM, l , or
  fiber)
• Path information (similar to an ERO)

44
LSP Hierarchy
LSC Cloud
TDM Cloud
PSC Cloud
LSC Cloud
TDM Cloud
PSC Cloud
FSC Cloud
Fiber 1
Bundle
Fiber n
FA-PSC
FA-TDM
l LSPs
l LSPs
FA-LSC
Explicit Label LSPs
Time-slot LSPs
Explicit Label LSPs
Time-slot LSPs
Fiber LSPs
(Multiplex Low-order LSPs)
(Demultiplex Low-order LSPs)

• Nesting LSPs enhances system scalability
• LSPs always start and terminate on similar
  interface types
• LSP interface hierarchy
• Fiber Switch Capable (FSC) Highest
• Lambda Switch Capable (LSC)
• TDM Capable
• Packet Switch Capable (PSC) Lowest

45
Constraint-based Routing

• Reduces the level of manual configuration
• Input to CSPF
• Path performance constraints
• Resource availability
• Topology information(including FA-LSPs)
• Output
• Explicit route for GMPLS signaling

46
GMPLS Signaling Extensions

• Label Related Formats (Generalized Labels)
• Generalized label request
• Link protection type (none, 11, 1N, or ring)
• LSP encoding type (packets, SONET, SDH, clear,
  DS-0, DS-1, )
• Generalized label object
• Packet (explicit in-band labels)
• Time slots (TDM)
• Wavelengths (lambdas)
• Space Division Multiplexing (fiber)
• Suggested label
• Label can be suggested by the upstream node
• Speeds LSP setup times
• Label set
• Restrict range of labels selected by downstream
  nodes
• Required in operational networks

47
GMPLS Signaling Extensions

• Bi-directional LSPs
• Resource contention experienced by reciprocal LSP
  using separate signaling sessions
• Simplifying failure restoration in the non-PSC
  case
• Lower setup latency
• RSVP notification messages
• Notify message informs non-adjacent nodes of LSP
  events
• Notify-ACK message supports reliable delivery
• Egress control
• Terminate LSP at a specific output interface of
  egress LSR

48
Link Management Protocol
Control Channel
Bearer Channel

• The link between two nodes consists of
• An in-band or out-of-band control channel
• One or more bearer channels
• Link Management Protocol (LMP)
• Automates link provisioning and fault isolation
• Assumes the bi-directional control channel is
  always available
• Control channel is used to exchange
• Link provisioning and fault isolation messages
  (LMP)
• Path management and label distribution messages
  (RSVP or CR-LDP)
• Topology information messages (OSPF or IS-IS)

49
Link Management Protocol
Services Provided by LMP

• Control channel management
• Lightweight keep-alive mechanism (Hello protocol)
• Reacts to control channel failures
• Verify physical connectivity of bearer channels
• Ping test messages sent across each bearer
  channel
• Contains senders label (fiber, ?) pair object
  for channel
• Eliminates human cabling errors
• Link property correlation
• Maintains a list of local label to remote label
  mappings
• Maintains list of protection labels for each
  channel
• Fault isolation
• Loss of light is detected at the physical
  (optical) layer
• Operates across both opaque (DXC) and transparent
  (PXC) network nodes

50
Link Bundling
Bundled Link 1
Bundled Link 2

• Multiple parallel links between nodes can be
  advertisedas a single link into the IGP
• Enhances IGP and traffic engineering scalability
• Component links must have the same
• Link type
• Traffic engineering metric
• Set of resource classes
• Link multiplex capability (packet, TDM, ?, port)
• (Max bandwidth request) ? (bandwidth of a
  component link)
• Link granularity can be as small as a ?

51
GMPLS Benefits

• Open standards allow selection of best-in-class
  equipment
• Routers have visibility into the
  transmissionnetwork topology
• Eliminates N2 meshes of links scaling issue
• Reduces routing protocol stress
• Optical paths span an intermix of routers and
  OXCs to deliver provisioning-on-demand networking
• Leverages operational experience with MPLS-TE
• No need to reinvent a new class of control
  protocols
• Promotes parallel evolution of UNI and NNI
  standards
• Enables rapid development deployment of new OXCs

52
GMPLS Modern Thinking for Modern Times

• Aligns with the way that the next generation
  network needs to be built and managed
• 20th Century Transmission network was dominant
• Voice ran over the transmission network
• ATM/Frame Relay delivered private data services
• Internet was just one among many services
• Transmission network created subscriber services
• 21st Century Internet is dominant
• Routers create the services that matter ()
• Network must be optimized for IP/Internet
• OC-48/OC-192 make routers the largest consumers
  of bandwidth
• New architecture is driven by routers subsuming
  functions previously performed by the
  transmission network
• The transmission network must evolve in a way
  that is most beneficial to the creation of
  Internet services