D.Mynbaev TCET4102,Module12,Spring2008 1
NEW YORK CITY COLLEGE of TECHNOLOGYTHE CITY UNIVERSITY OF NEW YORK
DEPARTMENT OF ELECTRICAL AND TELECOMMUNICATIONS
ENGINEERING TECHNOLOGY
Course : TCET 4102 Fiber-optic communicationsModule 12: Optical networks - Power budget, GMPLS, protocols and standards
Prepared by: Professor Djafar K. MynbaevSpring 2008
D.Mynbaev TCET4102,Module12,Spring2008 2
• Module outline:
– Power budget of a fiber-optic communications link
– Optical networks: Architecture and protocols
• Architectures
– The Internet and optical networks
– Generalized multiprotocol label switching, GMPLS
– Core and edge networks
– Control plane and data plane
– Overlay and peer models of control plane
• Layered architecture of transport networks
– Protocols and standards
• Standardization organizations
• Development of control plane standards
Textbook: Djafar K. Mynbaev and Lowell L. Scheiner, Fiber-Optic Communications Technology, Prentice Hall, 2001, ISBN 0-13-962069-9.
Notes:
The figure numbers in these modules are the same as in the textbook. New figures are not numbered.
Always see examples in the textbook
Key words
• Power budget and time budget
• IP over optical networks
• Generalized multiprotocol label switching, GMPLS
• Core and edge networks
• Control plane and data plane
• Layered architecture of transport networks
• Protocols and standards
• ITU-T, IETF and OIF
• ASTN – Automatically Switched Transport Network
• ASON - Automatically Switched Optical Network
• OTN – Optical Transport Network
Module 12: Optical networks – power budget,
GMPLS and protocols and standards
D.Mynbaev TCET4102,Module12,Spring2008 3
Power budget
• A fiber-optic network is the most widespread network in the world and it continues to grow. The hottest area of growth for fiber-optic-network installations is optical access networks, specifically, PONs, as discussed in Module 11.
• On of the most important problems in installation of optical networks is their design. Design is always more art than technique; however, some general recommendations can be provided. Also, regardless the type of optical network—long-haul, metro-or access—there are several points that are common to design of this networks.
• The first step in designing an optical network is considering their physical layers. Physical layout of fiber cables, the types of cables, the types of fibers, the types of connectorization, and the hardware --making the right decisions from among all these choices is the subject of this design.
• From a logical standpoint, design is the first step in network installation, but the designer has to be familiar as well with installation systems and components. It is not our intention to make you a professional network designer; rather, our purpose here is to give you, a future professional user of these systems, some insight into the area of design.
• It should be noted that there are some standards and recommendations that help a designer in his or her work. For example, at the outset that Regulation TIA/EIA 568A, the ―Commercial Building Telecommunications Cabling Standard,‖ is widely accepted as the industry standard for LAN installation. We’ll refer often to this standard in this section. To start with, TIA/EIA 568A accepts only two types of fiber -- 62.5/125 μm graded-index multimode and a singlemode -- as standard transmission media for local-area networks.
D.Mynbaev TCET4102,Module12,Spring2008 4
• Link consideration -- power budget and rise- time budget (bandwidth)
• Power budget
• Physically, a network is a collection of nodes connected by links. Let’s consider an individual fiber link that may include splices, connectors, and some other passive components. Because of the attenuation introduced by these components, a receiver gets much less light power than was transmitted to it. The question is this: Does the light signal arriving at the destination point of a link have enough power to be detected by the receiver? This is what power budget is about. Fig. 8.21 demonstrates the concept: Power- budget consideration allows us to calculate power at the receiving end and know the loss allocations along the link.
• An example shown in the graph in Fig. 8.21 illustrates the power-budget concept: The power-vs.-distance curve shows the light power at each point along the fiber link. When a transmitter radiates a light signal with the power of -10 dBm (0.1 mW), this is called initial power. A connector coupling light from a transmitter to a fiber causes an 0.2 dB attenuation; hence, notice the first decline in this curve. A patch-cord cable has an attenuation of 1.0 dB/km for 62.5 fiber at 1310 nm. This is the negative slope of the curve. Thus, after traveling 10 m, light attenuation is 1.0 dB/km x 0.01 km = 0.01 dB. At this point, a patch panel is used. The typical loss of a PC connector is 0.3 dB; hence, the curve drops at this loss level. From the patch panel, a regular fiber cable is used whose attenuation is 1.0 dB/km . Thus, the curve develops with a slope of -1.0 dB/km along the transmission distance. From the patch panel to the nearest fusion splicing point, the cable’s loss is 1.0 dB/km x 0.49 km = 0.49 dB. The attenuation introduced by fusion splicing is typically 0.2 dB; thus, the appropriate curve drop is shown in Fig. 8.21. As you continue to go along the link, you need to take into account all sources of attenuation with respect to their locations. This is shown in Fig. 8.21, and an analysis of this figure will help you understand this idea.
• (You should recall the three basic terms used here: light power is measured in dBm and it is mostly less than 1 mW; hence, it is negative; loss is measured in dB as the difference between two dBm values of light power; attenuation is measured in dB/km as loss per distance and it is positive by convention.)
Power budget
D.Mynbaev TCET4102,Module12,Spring2008 5
Po
wer
budget
Figure 8.21
D.Mynbaev TCET4102,Module12,Spring2008 6
Power-budget calculation can be made as follows:
Cable loss 1.0 dB/km x 2.0 km = ..........2.0 dB
Splicing loss
Fusion 0.2 dB x 2 = 0.4 dB
Mechanical = 0.3 dB
Total splicing loss............=...........0.7 dB
Connector loss
0.3 dB x 2 = 0.6 dB
0.2 dB x 2 = 0.4 dB
Total connector loss .........=.........1.0 dB
Total loss.............................=................................................. 3.7 dB
Transmitter power launched into a fiber, Pin......................... - 10 dBm
Power at the receiving end, Pout = Pin - total loss =.................. -13.7 dBm
Receiver sensitivity, Prec......................................................... -20 dBm
Power margin = Pout - PRS = ................................................... 6.3 dB
Again, these calculations introduce the of idea how power budget works for network design. It allows you to calculate the maximum distance you can achieve with a given fiber cable and the number of splices and connectors you can afford to run along your link. The loss allocation shown in Fig. 8.21 helps you to find critical points and achieve a better design .
The first piece of advice for a designer: Always keep reasonable power margins because your network will grow; therefore, you’ll need to connect new users and include new components, steps that are accompanied by power loss.
Power budget
D.Mynbaev TCET4102,Module12,Spring2008 7
• Example:
Problem: A fiber-optic communications link includes five splices at 0.02 dB/splice, four connectors at 0.2 dB/connector, transmitter power of - 10 dBm, and receiver sensitivity of -25 dBm. What is the maximum transmission distance of this link if a singlemode fiber cable with attenuation of 0.3 dB/km is used?
Solution: If we analyze the power-budget calculations given previously, we can readily write the following formulas:
Total loss (dB) = Pout (dBm) – Pin (dBm), where Pin = Prs.
On the other hand,
Total loss (dB) = Fiber loss (dB) + Component loss (dB)
Since we need to find a distance, we need to isolate the fiber loss, which is the only member of these equations that containdistance as a parameter. Hence,
Fiber loss (dB) = Total loss (dB) – Component loss (dB)
Now the maximum length of a fiber-optic link can be compute as
Maximum transmission distance (reach) = -Fiber loss (dB)/Attenuation (dB/km).
Let’s plug the given numbers into this formulas:
Pin = -10 dBm and Pin = Prs = -25 dBm.
Therefore,
Total loss = -25 (dBm) – (-10 (dBm)) = -15 dB.
Component loss = 5 x -0.02 dB + 4 x -0.2 dB = -0.9 dB.
Thus,
Fiber loss = -15 dB – (-0.9 (dB)) = -14.1 dB.
Finally,
Maximum transmission distance (reach) = 14.1 dB/0.3 dB/km = 47 km..
Power budget
D.Mynbaev TCET4102,Module12,Spring2008 8
As optical communications has migrated from point-to-point links to optical networks,
switching and routing problems
one approach:
optical packet
switching
Data packets
Optical (fiber/lambda)
circuit switching
Optical
packet switching
Electronic
packet switching
O/E
conversion
Data packets
Internet
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 9
• Internet today and tomorrow: Delivering traffic occurs in the optical domain, while processing the signals mostly takes place in electronic domain:
• There are optical switches that provide circuit-switching operations, including fiber switching and lambda switching not efficient way to deliver the Internet traffic;
• Delivering Internet traffic requires routing packets, which is done by electronics O/E/O conversions costly and slow the strong need for optical processing:– There is a lot of research aimed at developing optical packet-switching
technology, but there is no commercially available technology.
• Another emerging approach is using labels for data transport through an optical core transport, thus leaving electronic processing to the edge IP routers.
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 10
Data
Optical (fiber/lambda)
circuit switching
Electronic
packet switching
O/E
conversion
Internet
Internet
IP router/
MPLS
Data
/GMPLS
IP router/
MPLS
/GMPLS
Optical core
Label-based approach to
delivering the Internet traffic.
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 11
Overview of optical networks
Architectures and protocols: Protocols and standardsGMPLS – how it works (after [31], [35], [36]).
Data
O/E
conversion
InternetIP router/
MPLS
/GMPLS
Optical coreLabels over packet headers
Label switched path (LSP)
Generalized labels
Generalized LSP
D.Mynbaev TCET4102,Module12,Spring2008 12
Overview of optical networks
Architectures and protocols: Protocols and standards
GMPLS main features:
•It extends the MPLS concept of label switching;
•It allows for switching Layer 0 (WDM), Layer 1 (SONET/SDH), Layer 2 (ATM) and
Layer 3 (IP) traffic;
•It allows for automatic provisioning bi-directional optical path through a core
network;
•GMPLS nodes can operate with links with one or more of the following switching
capabilities: fiber-switched capable (FSC), lambda-switched capable (LSC), TDM-
switched capable (TSC), and packet-switched capable (PSC);
•GMPLS is based on extension of signaling (RSVP-TE and CR-LDP) and routing
(OSPF-TE and ISIS-TE) protocols. New LMP protocol has been introduced.
•It provides separation data and control planes, which makes data plane protocol
independent;
•GMPLS is still draft (final!); however, there are commercially available products.
GMPLS – how it works (after [31], [35], [36]).
D.Mynbaev TCET4102,Module12,Spring2008 13
• The state of the field today can be considered evolutionary.
• Optical networks have become the main transport ―facilities‖ for all telecommunications traffic (from ultra-long-haul to local transmissions, where the ―last mile‖ problem is being resolved by installation of PON).
• The shift in the paradigm of optical transmission from a simply ―piping‖ function to performing the logical upper-layer function.
• Packet routing is still done in the electronic domain.
• Optical packet switching (OPS), optical burst switching (OBS) and optical label switching (OLS) approaches are emerging.
• All these developments and research efforts are quickly melding into new cutting-edge technology that we should see being implemented commercially.
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 14
Core network
Edge
End user
End user
Edge
End user
End user
End user
Core and edge networks
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 15
• Core network transmits traffic from edge to edge.
• Edge networks prepare traffic for transmission; e.g., mapping T1
• and T3 traffic into SONET frames from one user while wrapping
• ATM traffic into SONET format from the other.
• Thus, the function of a core network is transport traffic from edge
• to edge while the function of a core network is to aggregate traffic
• from end users for core transmission.
• Core-edge classification can be applied to the networks of any
• scale: long-haul, metro, and access.
Optical networks
Architectures and protocols
D.Mynbaev TCET4102,Module12,Spring2008 16
Overview of optical networks
Architectures and protocols
Optical networks - general picture
Optical networks include transport network, control
plane, and client networks. Clients are edge IP/MPLS
networks while transport are core WDM networks. Client
networks request for services and transport networks
provide these services in form of fixed bandwidth
(lightpaths). Control plane is responsible for signaling and
routing, that is, enables transport core to deliver traffic.
(See next slide.)
D.Mynbaev TCET4102,Module12,Spring2008 17
OXCOXC
OXC
OXC
Optical transport network
IP router
IP router
Client
(IP/MPLS
edge network)
Control plane
Client
D.Mynbaev TCET4102,Module12,Spring2008 18
Overview of optical networks
Architectures and protocols: Architectures
Optical transport network (OTN) needs a
well-defined control plane that can handle
various service requirements. Therefore,
development of control plane is a critical
issue toward building the intelligent optical
networks.Control and data planes of an individual network node are
presented in the next slide.
D.Mynbaev TCET4102,Module12,Spring2008 19
Overview of optical networks
Architectures and protocols: Architectures
Control plane and data plane: Separation of control and transport functions
Data plane
(OXC)
Control plane
(router)
Control
messages
Control
messages
TrafficTraffic
Interface
Optical network node
D.Mynbaev TCET4102,Module12,Spring2008 20
Overview of optical networks
Architectures and protocols: Architectures
Control plane and data plane (continued)
General remarks:
Control plane: A set of software and/or hardware residing in a
network node that execute control and management functions.
Implementation of control plane depends on protocols. An example of
hardware is a router, such as label edge router, LER, or label switched router,
LSR, for GMPLS control plane. Examples of control plane protocols include
signaling system seven (SS7) protocol stack in voice transmission, open shortest
path first (OSPF) routing protocol in data transmission and generalized
multiprotocol label switching (GMPLS) protocol.
Data (information, or forwarding) plane: A set of hardware and
software that provides transportation of voice, data, and video
traffic. An example of hardware is an OXC and an example of protocols is IP
suite.
Modified!
D.Mynbaev TCET4102,Module12,Spring2008 21
Overview of optical networks
Architectures and protocols: Architectures
Control plane and data plane (continued)
General remarks
Control and data plane interaction: Control plane at a node generates
routing and label tables and exchange this information with peers.
This information is used by data (forwarding in IP routers) plane for
transportation [19]. In other words, control plane protocols (OSPF
and others) enable IP to forward traffic correctly [15]. Separation of
control and data planes makes data plane protocol-independent.
Today, control plane interact with data (forwarding) plane through
open interface, which constitute the third (current) generation of the
network element architecture [19].
D.Mynbaev TCET4102,Module12,Spring2008 22
Overview of optical networks
Architectures and protocols: Architectures
Control plane and data plane (continued)
Control domain 1
Control domain 2
NNI
Subnetwork 1
Subnetwork 2
NNI – network-network interface
D.Mynbaev TCET4102,Module12,Spring2008 23
Overview of optical networks
Architectures and protocols: ArchitecturesControl plane and data plane (continued)
Control plane - a general view: Control planes residing in nodes of
any given subnetwork make up a control domain of this subnetwork.
Control planes enable traffic transportation within and between their
subnetworks.The main functions of an optical control plane are targeted solving the problem of ‖find,
route, and connect,‖ which requires the follows [17], [18]:
•Naming and addressing scheme (find)
•A routing process to handle the network resources usage and route calculation (route),
including routing and wavelength assignment (RWA) and topology and resources discovery
•A signaling network that provides communication between entities requesting services and
those provision these services
•A signaling protocol for the setup, maintenance, and tear down of optical trails, including
lightpath signaling and maintenance
In addition, control plane has to support network survivability based on fault monitoring
and protection and restoration.
D.Mynbaev TCET4102,Module12,Spring2008 24Source: [21]
Control plane and data plane (continued)
D.Mynbaev TCET4102,Module12,Spring2008 25
Overview of optical networks
Architectures and protocols: Architectures
Control plane – UNI and NNI: Control domains of different
subnetworks of transport network communicate through network-to-
network (NNI) interface, while client networks get access to the
transport network through user-to network (UNI) interface.
When GMPLS is used, NNI provides interface between two
network-side label-switched routers (LSRs).
UNI are subdivided into client UNI (UNI-C) and network UNI (UNI-
N) interfaces. UNI-C provides a signaling termination function,
whereas UNI-N provides pass-through and internetworking function,
circuit routing, and reachability information. They communicate to
one another in IP packets. When GMPLS is used, UNI provides
interface between edge GMPLS node and network-side LSR [20].
Control plane and data plane (continued)
D.Mynbaev TCET4102,Module12,Spring2008 26
Overview of optical networks
Architectures and protocols: Architectures
Development of an User-Network Interface (UNI) by the Optical Internetworking
Forum (OIF) [22]:
•UNI enables client user devices to dynamically request services from the optical
transport network
•Client devices include IP routers, ATM switches, SONET/SDH network elements,
etc.
•Main functions of an UNI:
•Signaling (in-fiber and out-of fiber)
•Addressing (use a special address space do not reveal topology, resources, and
addressing of the optical network to the clients)
•Link management (used, in particular, to create, delete, and query the status of
connections).
UNI/NNI interoperability was demonstrated at OFC’03 and following conferences.
•UNI determines the visibility that a client (IP router) has into the optical transport
network. The service model that uses UNI concept is called overlay model.
Control plane and data plane (continued)
D.Mynbaev TCET4102,Module12,Spring2008 27
Control plane models.
Overview of optical networks
Architectures and protocols: Architectures
Control plane overlay model:
Client control is independent of the
optical layer OTN is opaque to IP
network two
control planes.
After [24], [25], and [26].
D.Mynbaev TCET4102,Module12,Spring2008 28
Control plane models
Overview of optical networks
Architectures and protocols: Architectures
Control plane peer model: Single control for both client and
optical domains OTN is transparent to IP network one control
plane.
After [24], [25], and [26].
Client (IP) domainOptical (transport)
domainClient (IP) domain
D.Mynbaev TCET4102,Module12,Spring2008 29
Overview of optical networks
Architectures and protocols: Architectures
Control plane and data plane (continued)
Problems in developing control planes [16], [20], [21], [22]:
• Optical control plane is distributed in contrast to traditional centralized plane
multiplicity of control planes control planes of optical networks must be able
to work with other control planes (multiple operator networks; multiple vendors;
different technologies; multiple administrative domains, e.g., core-metro
application) interoperability between equipment from different vendors.
• Control plane has to work with all-optical networks change in technology will
require new developments in control plane.
• Control planes have to operate over disparate transport technologies (IP, ATM,
SONET/SDH, and WDM). (This problem is resolved by developing unified control
plane (UCP) that represents a common set of control functions and interconnection
mechanisms that allow unified communication, routing, and control across all these
transport technologies).
•Testing (Some platforms have been developed by Agilent Technologies).
D.Mynbaev TCET4102,Module12,Spring2008 30
Overview of optical networks
Architectures and protocols: Architectures
Definition of control plane is the critical issue because this
plane is responsible for routing and signaling processes in
optical networks. In other words, it is a control plane that
adds intelligence to the optical networks.
• There are two main approaches to control plane definition: GMPLS
based (widely accepted) and PNNI based (proposed [25],[26]).(GMPLS – generalized multiprotocol label switching. ; PNNI – private network-
network interface, the control plane developed and widely used in ATM networks.)
• Development of GMPLS-based control plane interacts with the
development of ASON-based optical network architecture. [30] (ASON
–automatic switched optical network.)
Control plane and data plane (continued)
Modified!
D.Mynbaev TCET4102,Module12,Spring2008 31
Overview of optical networks
Architectures and protocols: Architectures
Layered architecture of optical transport networks
• Yesterday and today:
– Initially optical fibers were used as pipes to transport large volume of traffic while all processing (intelligent) work was relegated to electronics. Thus, multiplexing, switching and routing was done in electronic domain. Optical transport was simply sets of point-to-point links.
– Today optical networks have reached the point where the need arise for execution of all transport tasks in optical domain.
– Now we are in transition stage.
D.Mynbaev TCET4102,Module12,Spring2008 32
Layered architecture of optical transport networks
Overview of optical networks
Architectures and protocols: Architectures
IP
WDM optical layer
SONET
ATM
Data services
Bandwidth utilization and QoS
Transport and network resilience
Optical transmission capacity
Optical networks:Typical today architecture.
Best effort IP
QoS data
Voice and
leased line
D.Mynbaev TCET4102,Module12,Spring2008 33
Overview of optical networks
Architectures and protocols: ArchitecturesNEW!
IP router IP router
ATM switch
SONET switch
ATM switch
SONET switch
WDM network WDM network
Transport
Data Data
Layered architecture of optical transport networks – physical connections
D.Mynbaev TCET4102,Module12,Spring2008 34
Overview of optical networks
Architectures and protocols: Architectures
Modern multi-layer network architecture is translated into physical
implementation as shown in the previous slide:
Data packets from IP routers go to ATM switches; the ATM switches
connect to SONET switches; and the SONET switches connect to
DWDM network. Optical network transport traffic and at the
destination point the reverse process takes place.
This type of data transmission is ineffective and costly because:
•Each layer has its own management and control
•Each layer is managed separately by different types of service
providers
•Interfacing between layers requires manual provision.
NEW!
Layered architecture of optical transport networks – physical connections
D.Mynbaev TCET4102,Module12,Spring2008 35
IP/MPLS
GMPLS-based intelligent
optical layer
Data services
Data transport with
flexible bandwidth,
QoS, and network
resilience
Optical networks: Architecture of today and tomorrow.
Overview of optical networks
Architectures and protocols: Architectures
Layered architecture of optical transport networks
Best effort IP,
QoS data, voice,
private lines
Optical Ethernet
and circuit-oriented
applications
After [22], [27].
D.Mynbaev TCET4102,Module12,Spring2008 36
Layered architecture of optical transport networks –
physical connections
Overview of optical networks
Architectures and protocols: Architectures
NEW!
As the network migrates, intermediate layers will begin to disappear:
•ATM will be eliminated by using MPLS
•SONET migrates to next-gen SONET with GMPLS
•DWDM will migrate to optical intelligent network with switching
In this architecture, data packets will be transported directly by
optical layer.
D.Mynbaev TCET4102,Module12,Spring2008 37
Layered architecture of optical transport networks – migration to next-gen network
Overview of optical networks
Architectures and protocols: Architectures
IP
WDM optical layer
SONET
ATM
Today
New!
IP/MPLS
GMPLS-based intelligent
optical layer
Today and tomorrow
MPLS
VoIP
GMPLS
DWDM/
CWDM
D.Mynbaev TCET4102,Module12,Spring2008 38
Layered architecture of optical transport networks – migration to next-gen network
Overview of optical networks
Architectures and protocols: ArchitecturesNew!
Telecommunications technology is undergoing two significant changes (we
could call them "revolutions"): rapid migration to MPLS and aggressive
deployment of VoIP. VoIP is considered as a complement technology (and
the future replacement) to the traditional TDM-based voice circuits.
At the same time, optical communications technology, bolstered by the
relatively mature DWDM technology and the rapidly developing CWDM, is
migrating from simple point-to-point "pipes" to intelligent networks. This
migration is based on the development of optical control plane, where
GMPLS- and ASON-based approaches are emerging to provide signaling
and routing in these new networks.
All these developments constitute a new step in making optical intelligent
networks a reality.
D.Mynbaev TCET4102,Module12,Spring2008 39
Overview of optical networks
Architectures and protocols: ArchitecturesNEW!
Layered architecture of optical transport networks – physical connections and
interfaces
Source: http://www.convergedigest.com/tutorials/mpls6/page3.asp
UNI – user network interface, O-UNI – optical UNI, LMP – link management protocol.
D.Mynbaev TCET4102,Module12,Spring2008 40
Overview of optical networks
Architectures and protocols: Architectures
Layered architecture of optical transport networks
IP router mesh
ATM mesh
SONET rings
Point-to-point WDM
Topology view
at the traditional
multi-layer
architecture
D.Mynbaev TCET4102,Module12,Spring2008 41
Overview of optical networks
Architectures and protocols: Architectures
Layered architecture of optical transport networks
IP router mesh
Mesh and ring WDM
Client layer
Optical layer
Topology view
at the emerging
collapsed-layer
architecture
After [22].
D.Mynbaev TCET4102,Module12,Spring2008 42
Optical network overlay model
Overview of optical networks
Architectures and protocols: Architectures
Optical transport network
IP
network
IP
network
UNIUNI
D.Mynbaev TCET4102,Module12,Spring2008 43
Overview of optical networks
Architectures and protocols: Architectures
IP edge router requests a service (for instance, a label
switched path connection) from an adjacent OTN router. An
optical network interconnects its core nodes to provide the
requested service but doesn’t inform IP network. Thus, OTN
offers high-bandwidth connectivity in the form of lightpaths.
No routing or other type of information from the optical
network is available to the IP networks; that is, OTN
(service provider) is opaque to the the IP network (service
requester). (After [28]). See also overlay model of a control plane.
Optical network overlay model
D.Mynbaev TCET4102,Module12,Spring2008 44
Overview of optical networks
Architectures and protocols: ArchitecturesNew!
Optical network overlay model: individual connection with GMPLS
OXC
OXC
Optical transport network (OTN)
IP router = label-
switched router
(LSR)
UNI
UNI
Two separate control planes: One in OTN and the other in the LSRs.
Advantage: It is easily to deploy since transport and clients are
independent. Disadvantage: data and control traffic are combined
limited number of LSRs can participate in the network [34].
IP router = label-
switched router (LSR)
D.Mynbaev TCET4102,Module12,Spring2008 45
Overview of optical networks
Architectures and protocols: Architectures
IP
networkIP
network
Optical transport network
IP and optical domain
Overview of optical networks
Architectures and protocols: Architectures
Optical network peer model
Modified!
D.Mynbaev TCET4102,Module12,Spring2008 46
Overview of optical networks
Architectures and protocols: Architectures
Optical network peer model
IP and OTN are treated as a single integrated network. Here,
an OXC is treated like another router as far as the control
plane is concerned. From routing and signaling points of view
there is no difference between a router-to-OXC interface and
OXC-to-OXC interface. Once a lightpath is established
across the OTN, it can be considered as a virtual link between
edge routers. Therefore, edge IP routers are involved in
optical transmission OTN is transparent to IP.After [28]).
D.Mynbaev TCET4102,Module12,Spring2008 47
Overview of optical networks
Architectures and protocols: ArchitecturesNew!
Optical network peer model: individual connections with GMPLS
OXC3
OXC1
OTN
LSR3
LSR1LSR2
OXC2
D.Mynbaev TCET4102,Module12,Spring2008 48
Overview of optical networks
Architectures and protocols: Architectures
Optical network peer model: individual connections with GMPLS
New!
Main features:
•Lightpath is used exclusively for data forwarding
•A control plane (signaling and routing) spans across
both OTN and edge IP routers (LSRs)
•More flexible and effective than an overlay model
•Difficult to deploy because it requires interoperability
between edge and core networks long-term
perspective.
D.Mynbaev TCET4102,Module12,Spring2008 49
Overview of optical networks
Architectures and protocols: Protocols and standards
Next generation of optical networks must be able to handle
a larger set of applications. This is why new networking
architectures and protocols need to be developed.
In this subsection we will review optical network protocols
and standards. We will concentrate only on optical
standardization issues, while leaving outside of our
discussion generic Internet and other data-networking
protocols and standards.
D.Mynbaev TCET4102,Module12,Spring2008 50
Overview of optical networks
Architectures and protocols: Protocols and standardsStandardization organizations and areas of their activities:
Organization Main activity in optical networking Comments
ITU-TInternational
Telecommunication Union –
Telecommunication
Standardization Sector
Automatically Switched Optical
Network (ASON), Automatically
Switched Transport Network (ASTN),
Optical Transport Network (OTN)
Architecture and
framework of optical
control plane and
optical transport plane
IETFInternet Engineering Task
Force
Generalized Multiprotocol Label
Switching (GMPLS), Common Control
and Measurement Plane (CCAMP), IP
over Optical (IPO)
Optical control plane
OIFOptical Internetworking
Forum
User-to-Network Interface (UNI),
Network-to-Network Interface (NNI)
Optical control plane
D.Mynbaev TCET4102,Module12,Spring2008 51
Overview of optical networks
Architectures and protocols: Protocols and standardsNew!
Standardization organizations, main areas of their activities and their
relationship: (after [30].)
ITU-T
ASON
IETF
GMPLS
OIF
UNI/NNI
D.Mynbaev TCET4102,Module12,Spring2008 52
Overview of optical networks
Architectures and protocols: Protocols and standards
Standardization efforts in the optical networking area are concentrated today mostly
on development of optical control plane standards; these standards are necessary to
(1) allow next-generation optical networks to be built out of devices from a
mixture of vendors and
(2) they specify the minimum set of features that these devices must support [30].
Two organizations—ITU-T and IETF—are the most active in this area. Mapping
between their standardization efforts are shown in the next slide.
In addition, OIF (non-profit organization with more than 300 members, including
many of the world’s leading carriers and vendors) is active in the development and
deployment of interoperable product and services for data switching and routing
using optical networking technologies [29]. The main outcome of the OIF efforts is
UNI 1.0 specification. OIF has started NNI standardization project.
IEEE is active in developing standards for optical Ethernet.
Standardization organizations and areas of their activities:
D.Mynbaev TCET4102,Module12,Spring2008 53
Overview of optical networks
Architectures and protocols: Protocols and standards
ITU-T generic architectures
Development of control plane standards
ASTN G.8070
ASON G.8080 OTN G.872
Sig
nal
ing
Ro
uti
ng
Dis
cover
y
Lin
k m
a--
nag
emen
t
CA
C
Man
age-
men
t
of
OT
N
Res
tora
tio
n
Tra
nsp
ort
IETF protocolsGMPLS architecture
GMPLS signaling functional
RSVP-TE and CR-LDP SDH/SONET
GMPLS routing functional
OSPF-TE and IS-IS SDH/SONET Link management
After [29].
D.Mynbaev TCET4102,Module12,Spring2008 54
Legend:
ASTN – Automatically Switched Transport Network
ASON - Automatically Switched Optical Network
OTN – Optical Transport Network
CAC – Connection Admission Control
GMPLS – Generalized Multiprotocol Label Switching
RSVP – Resource Reservation Protocol
TE – Traffic Engineering
LDP – Label Distribution Protocol
CR-LDP – constraint-based LDP
LMP – Link Management Protocol
OSPF – Open Shortest Path First
IS-IS – Intermediate System-to-Intermediate-System After [29].
Overview of optical networks
Architectures and protocols: Protocols and standards
D.Mynbaev TCET4102,Module12,Spring2008 55
Overview of optical networks
Architectures and protocols: Protocols and standards
ITU-T standardization efforts are concentrated on developing
optical network architectures. The result is ASTN, which is an
optical transport network architecture with the dynamic connection
capabilities. This architecture includes management, control, and
transport planes. Control plane architecture is determined by ASON
that defines components in the optical control plane and interaction
among those components. Transport plane is determined by OTN
that enables optical transmission of various types of client signals
through the use of Forward Error Correction (FEC) bytes. (After [28]).
This architecture is shown in the following slide.
D.Mynbaev TCET4102,Module12,Spring2008 56
Clients:
IP, ATM,
TDM,
etc.
Clients:
IP, ATM,
TDM,
etc.
switch
Clients:
IP, ATM,
TDM,
etc.
OCC
OCC
OCC
OCC
switch
switch
switch
Management
plane
ASON control plane
Optical transport network
UNI
E-NNIE-NNI
ITU-T ASTN architecture (after [28]).
Overview of optical networks
Architectures and protocols: Protocols and standards
OCC –Optical connection controller
CCI – Connection controller interface
UNI – User-network interface
I-NNI – Interior network-network interface
E-NNI – Exterior network-network interface
D.Mynbaev TCET4102,Module12,Spring2008 57
Overview of optical networks
Architectures and protocols: Protocols and standardsThe following communications models help to summarize discussion on
standardization activities (after [31]).
Su
bn
etw
ork
1
Su
bn
etw
ork
2
IP layer IP layer
UNI UNI
NNI
Subnet
wo
rk 1
Su
bn
etw
ork
2
IP layer IP layer
IET
F m
odel UNI UNI
GMPLS
Optical layer
OIF
mo
del
Subnetwork 1 Subnetwork 1Client Client
UNI UNINNI
ITU
-T m
odel
GMPLS model
ASTN/ASON model
D.Mynbaev TCET4102,Module12,Spring2008 58
• Network paradigm is shifting: While we still want
from network more and more capacity (raw
bandwidth), main focus has shifted to bandwidth-
on-demand paradigm (―pay as you grow‖) -- that
is, flexibility and scalability based on virtually
unlimited bandwidth. Components will have to
meet this new reality.
• Bottom line for network operators: Reduce cost
per bit of transmitted signal and meet customer
requirements!
Conclusion
D.Mynbaev TCET4102,Module12,Spring2008 59
• Critical issues for tomorrow’s networks:
– Standards on optical control plane (GMPLS-
based, others?)
– Standards on transmission technology (Circuit
and packet switching SONET-based
protocols vs. IP-based protocols)
– Components (more functionality, cost
effective).
Conclusion
D.Mynbaev TCET4102,Module12,Spring2008 60
Conclusion
In spite of all the problems, optical
communications keeps growing and continues
to be the linchpin of modern
telecommunications.
What would we have without optical
communications?
D.Mynbaev TCET4102,Module12,Spring2008 61
References
References from Module 9 related to Module 12:
1. John R. Vacca, Optical Networking Best Practices Handbook, Hoboken, N.J.: Wiley –Interscience, 2007.
4. Djafar K. Mynbaev, The physical layer of the optical networks: Devices and subsystems, IEEE Communications Society, online tutorial, 2006.
5. Djafar K. Mynbaev and Lowell L. Scheiner, Fiber-Optic Communications Technology, Prentice Hall, 2001.
6. Djafar K. Mynbaev, ―Next-generation optical networks from network layer and physical layer perspectives,‖ Tutorial presented at the 11th International Conference on Telecommunications, Fortaleza, Brazil, August 2004.
7. Manasi Deval et al, ―Distributed Control Plane Architecture for Network Elements,‖ Intel Technology Journal, November 14, 2003, pp.51-63.
8. Uyless Black, Optical Networks, Prentice Hall, 2002.
9. Yinghua Ye and Sudhir Dixit, ―Surviavibility in IP-over-WDM Networks,‖ in IP over WDMedited by Sudhir Dixit, Hoboken, N.J.: Wiley – Interscience, 2003.
10. Rajiv Ramaswami and Kumar N. Sivarajan, Optical Networks – A Practical Perspective, 2nd ed., San Francisco: Morgan Kaufmann, 2002.
11. Vivek Alwayn, Optical Network Design and Implementation, San Jose, CA: Cisco Press, 2004.
12. Arun K. Somani, Survivability and Traffic Grooming in WDM optical Networks, New York: Cambridge University Press, 2006.
D.Mynbaev TCET4102,Module12,Spring2008 62
References fro Module 12:1. Djafar K. Mynbaev and Lowell L. Scheiner, Fiber-Optic Communications Technology,
Upper Saddle River, N.J.: Prentice Hall, 2001.
2. Benyuan Zhu, ―Ultra high density and long haul transmissions,‖ OFC’04, ThE1.
3. B. Mikkelsen et al, ―Deployment of 40 Gb/s systems: Technical and cost issues,‖ OFC’04, ThE6.
4. Jeff Hecht, ―Speeding up transmission rates with slower signals,‖ Laser Focus World, November 2002, pp. 91-96.
5. H.S. Chung, Y.G. Jang, and Y.C. Chung, ―Directly modulated CWDM/DWDM System using Negative Dispersion Fiber for Metro Network Application,‖ OFC’04, WG5.
6. Stamatis Kartalopoulos, ―The Flexibility of DWDM in Handling Continually Increasing Bandwidth Demands for Future Optical-Fiber Communication Networks,‖ IEEE LEOS Newsletter, April 2002, pp 15-19.
7. John Ceske et al, ―CWDM vertical-cavity surface-emitting laser array spanning 140 nm of the C, S, and L fiber transmission bands,‖ OFC’04, TuE7.
8. Ken-ichi Sato, ―Key Enabling Technologies for Future Networks,‖ Optics & Photonics News, May 2004, pp. 34-39.
D.Mynbaev TCET4102,Module12,Spring2008 63
References (continued):
9. Fei Xue and S.J. Ben Yoo, ―High-Capacity Multiservice Optical Label Switching for the Next-Generation Internet.,‖ IEEE Communications Magazine, May 2004,ppS16-S22.
10. S.J. Ben Yoo, ―Optical-label switching, MPLS, MPLambdaS, and GMPLS,‖ Optical Network, May/June 2003, pp. 17-31.
11. Botaro Hirosaki et al, ―Next-generation Optical Networks as a Value Creation Platform,‖ IEEE Communications Magazine, September 2003, pp. 65-71.
12. Ken-Ichi Kitayama and Masayuki Murata, ―Versatile Optical Code-Based MPLS for Circuit, Burst, and Packet Switchings,‖ Journal of Lightwave Technology, November 2003, pp. 2753-2764. 1.
13. Nasir Ghani, Sudhir Dixit and Ti-Shiang Wang, ―On IP-WDM Integration: A Retrospective,‖ IEEE Communications Magazine, September 2003, pp. 42-45.
14. Monir Hamdi and Chumming Qiao, ―Engineering the Next-Generation Optical Internet,‖ Optical Networks, November/December 2003, pp. 5-6.
15. Uyless Black, Optical Networks, Prentice Hall, 2002.
16. Angela L. Chiu and John Strand, ―Control plane considerations for all-optical and multi-domain networks and their status in OIF and IETF‖ Optical Networks, January/February 2003, pp.26-35.
D.Mynbaev TCET4102,Module12,Spring2008 64
References (continued):
17. Chinsheng Xin et al, ―On an IP-Centric Control Plane,‖ IEEE Communications Magazine, September 2001, pp. 88-93.
18. Alan McGuire, Shehzad Mirza, and Dareen Freeland, ―Application of Control Plane Technology to Dynamic Configuration Management,‖ IEEE Communications Magazine, September 2001, pp. 94-99.
19. Manasi Deval et al, ―Distributed Control Plane Architecture for Network Elements,‖ Intel Technology Journal, November 14, 2003, pp.51-63.
20. ―Driving Optical Network Evolution,‖ www.iec.org/online/tutorial/drive_opt/index/html.
21. ―OIF UNI/NNI Interoperability Demo,‖ www.oiforum.com/public/downloads/UNI-NNI.ppt.
22. Jim Jones, “User-Network Interface (UNI) 1.0,” Optical Networks, March/April 2003, 2003, pp. 85-93.
23. Daryl Eigen, Shibin Jiang, and Ike Song, ―Metro Amplifiers Provide Gain Without Pain,‖ Optical Networks, March/April 2003, pp. 95-97.
24. Peter Tomsu and Christian Schmultzer, Next Generation Optical Networks, Prentice Hall PTR, 2002.
D.Mynbaev TCET4102,Module12,Spring2008 65
References (continued):
25 Josep Sole-Pareta et al, ―Some Open Issues in the Optical Networks Control Plane,‖http://pcsostres.ac.upc.es/doctorat/papers/200306-172/icton2003-xmasip.pdf.
26 Sergio Sanchez-Lopez et al, ―PNNI-based Ciontrol Plane for Automatically Switched Optical Networks,‖ Journal of Lightwave Technology, November 2003, pp. 2673-2681.
27 Lu Shen and Byrav Ramamurthy, ―Provisioning and restoration in the next-generation optical core,‖ Optical Networks, March/April 2003, pp. 32-45.
28 Soo-Hyuan Choi et al, ―Standardization efforts in optical networking focused on architecture and signaling issues,‖ Optical Networks, May/June 2003, pp. 32-48.
29 Wesam Alangar, ―Optical networking evolution in ITU-T and IETF – A reality check,‖ OFC’04, FH1.
30 Nic Larkein, ―ASON and GMPLS – The Battle of the Optical Control Plane,‖ Data Connection Limited, Enfield, UK, August 2002.
31 Thomas DiMicelli, ―Emerging Control Plane Standards and the Impact on Optical Layer Services,‖ www.oiforum.com/public/downloads/DiMicelli2.ppt
32 Djafar Mynbaev, ―PON performance improves with multiplexing,‖ WDM Solutions, January 2003, pp.18-20.
D.Mynbaev TCET4102,Module12,Spring2008 66
References (continued):
33. Ari Tervonen, ―Optical Enabling Technologies for WDM Systems,‖ in IP over WDMedited by Sudhir Dixit, Hoboken, N.J.: Wiley – Interscience, 2003.
34. Yinghua Ye and Sudhir Dixit, ―Surviavibility in IP-over-WDM Networks,‖ in IP over WDM edited by Sudhir Dixit, Hoboken, N.J.: Wiley – Interscience, 2003.
35. ―Generalized Multiprotocol Label S witching (GMPLS),‖ Web ProForum Tutorials, The International Engineering Consortium (IEC), 2004.
36. ― GMPLS Overview,‖ in JUNOS 5.5 Internet Software Configuration Guide: MPLS Applications, Juniper Networks, 2003/2004