1

Connection-Oriented Networks - Harry Perros 1

Chapter 9: Wavelength Routing
Optical Networks
TOPICS
Wavelength routing networks
Protection schemes
G.709 - The digital wrapper
Control plane architectures
GMPLS
OIF UNI
Connection-Oriented Networks - Harry Perros 2
A wavelength routing network
This is an optical network that consists of OXCs
interconnected by WDM fibers, with each fiber
consisting of W wavelengths
DCS
Input
WDM
fibers
Output
WDM
fibers
Switch fabric
. . .

2
Connection-Oriented Networks - Harry Perros 3
OXC functionality
It switches optically all the incoming
wavelengths of the input fibers to the
outgoing wavelengths of the output fibers.
For instance, it can switch the optical signal
on incoming wavelength !
i
of input fiber k
to the outgoing wavelength !
i
of output
fiber m.
Connection-Oriented Networks - Harry Perros 4
Converters:
If it is equipped with converters, it can also
switch the optical signal of the incoming
wavelength !
i
of input fiber k to another
outgoing wavelength !
j
of the output fiber
m.
This happens when the wavelength !
i
of the
output fiber m is in use.
3
Connection-Oriented Networks - Harry Perros 5
Optical add/drop multiplexer (OADM):
An OXC can also be used as an OADM.
That is, it can terminate the optical signal of
a number of incoming wavelengths and
insert new optical signals on the same
wavelengths in an output port.
The remaining incoming wavelengths are
switched through as described above.
Connection-Oriented Networks - Harry Perros 6
Transparent and Opaque Switches
Transparent switch:
The incoming wavelengths are switched to the output
fibers optically, without having to convert them to the
electrical domain.
Opaque switch:
The input optical signals are converted to electrical signals,
from where the packets are extracted. Packets are switched
using a packet switch, and then they are transmitted out of
the switch in the optical domain.
4
Connection-Oriented Networks - Harry Perros 7
Lightpaths
Wavelength routing networks are circuit-
switched networks.
In order for a user to send data to a another
user, a connection has to be first setup.
This connection is a circuit-switched
connection and it is established by
allocating a wavelength on each hop along
the connections path
Connection-Oriented Networks - Harry Perros 8
An example of a lightpath
!1,..,!W !1,..,!W !1,..,!W !1,..,!W
OXC 1 OXC 2 OXC 3 Router A Router B
!1 !1
OXC 1 OXC 2 OXC 3 Router A Router B
!1 !1
A three-node wavelength routing network
A lightpath between routers A and B
5
Connection-Oriented Networks - Harry Perros 9
The wavelength continuity constraint
When establishing a lightpath over a wavelength
routing network, the same wavelength has to be
used on every hop along the path.
If the required wavelength is not available at the
outgoing fiber of an OXC through which the
lightpath has to be routed, then the establishment
of the lightpath is blocked, and a notification
message is sent back to the user.
Connection-Oriented Networks - Harry Perros 10
Converters
In order to decrease the probability that a
lightpath is blocked, the OXC can be
equipped with converters.
A converter can transform the optical signal
transmitted over a wavelength to another
wavelength.
6
Connection-Oriented Networks - Harry Perros 11
In an OXC, for each output fiber with W
wavelengths, there may be c converters, where 0 !
c !W.
No conversion: c=0
Partial conversion: 0 < c <W
Full conversion: c=W
A converter can only transform a signal on a
wavelength ! to another wavelength which is
within a few nm from wavelength !.
Connection-Oriented Networks - Harry Perros 12
An example of different lightpaths
O
E
Router B
Router
A
Router D
Router
C
E O
OXC 3
!
3
O E
E
O
OXC 2
OXC 1
!
1
!
1
!
1
!
1 !
1
!
1
!
2
!
3
Lightpaths
A -> C: !
1
B -> D:!
1
and !
2
C -> D:!
3
and !
1
OXCs 1 and 2: no converters
OXC 3 has converters
7
Connection-Oriented Networks - Harry Perros 13
Traffic grooming
A lightpath is exclusively used by a single
client.
Often the bandwidth a client requires is
significantly less than the wavelengths
bandwidth. This means that part of the
lightpaths bandwidth is unused. Also, the
user pays for more bandwidth than required.
Traffic grooming permits mane user to
share the same lightpath.
Connection-Oriented Networks - Harry Perros 14
Sub-rate units
The bandwidth of a lightpath is divided into sub-
rate units so that it can carry traffic streams
transmitted at lower rates.
For instance a 2.5 Gbps (OC-48) bandwidth can
be available in sub-rate units of 50 Mbps (OC-1)
A client can request one or more of these sub-rate
units. This improves wavelength utilization and
lowers users costs.
8
Connection-Oriented Networks - Harry Perros 15
An example of traffic grooming
Established lightpaths:
OXC 1 to OXC 3
OXC 3 to OXC 4
Wavelength capacity: 2.488 Gbps (OC-48/STM-16)
16 sub-rate units of 155 Mbps (OC3/STM-1)
OXC 1
!
1
!
1
OXC 5 OXC 6
OXC 4
OXC 3 OXC 2
!
2
!
2
Connection-Oriented Networks - Harry Perros 16
A user attached to OXC 1 that wants to
transmit data to a user attached to OXC 3,
can request any integer number of OC-
3/STM-1 sub-rate units up to a total of 16.
Additional lightpaths can be established
between OXCs 1 and 3, if the traffic
between these two OXCs exceeds 2.488
Gbps.
9
Connection-Oriented Networks - Harry Perros 17
Traversing more than one lightpath:
Let us consider a user attached to OXC 1
who requests a connection to a user attached
to OXC 4 for four sub-rate units.
In this case, a new lightpath has to be
established between OXCs 1 and 4, say,
over OXCs 6 and 5.
Connection-Oriented Networks - Harry Perros 18
Alternatively, the connection can be routed
through the two lightpaths (OXC 1 -> OXC
3 and OXC 3 -> OXC 4).
Provided that there is free capacity on each
lightpath and OXC 3 is equipped with a
SONET/SDH DCS which permits it to
extract the data stream from the incoming
SONET/SDH frames on the first lightpath
and place it into the SONET/SDH frames of
the second lightpath.
10
Connection-Oriented Networks - Harry Perros 19
Protection schemes
Optical networks will be used by
telecommunications companies and other
network providers, which typically require a
carrier grade reliability.
That is, the network has to be available
99.999% of the time, which translates to an
average downtime for the network of 6
minutes per year!
Connection-Oriented Networks - Harry Perros 20
Types of failures:
Link failures are very common and they occur
when a fiber cable is accidentally cut.
A link can also fail if an amplifier that boosts the
multiplexed signal of all the wavelengths on the
fiber fails.
An individual wavelength within a fiber may also
fail if its transmitter or receiver fails.
Finally, an OXC can fail, but this is quite rare due
to built-in redundancies.
11
Connection-Oriented Networks - Harry Perros 21
Path and link protection
Protection can be performed at the level of
an individual lightpath or at the level of a
single fiber.
Path protection denotes schemes for the
restoration of a lightpath, and
Link protection denotes schemes for the
restoration of a single fiber, whereby all the
wavelengths are restored simultaneously.
Connection-Oriented Networks - Harry Perros 22
Point-to-point links
The simplest optical network is a point-to-
point WDM link that connects two nodes.
Link protection can be done in a
dedicated 1+1 manner, or in a
non-dedicated 1:1 or 1:N manner
12
Connection-Oriented Networks - Harry Perros 23
Dedicated 1+1 scheme:
the signal is transmitted simultaneously over
two separate fibers which are preferably
diversely routed.
The receiver monitors the quality of the two
signals and selects the best of the two.
If one fiber fails, then the receiver continues to
receive data on the other fiber.
Connection-Oriented Networks - Harry Perros 24
1:1 scheme:
There are still two diversely routed fibers, a working
fiber and a protection fiber.
The signal is transmitted over the working fiber, and if
this fiber fails, the source and destination switch to the
protection fiber.
Shared 1:N scheme:
This is a generalization of the 1:1 scheme, where N
working fibers are protected by a single protection
fiber. (Only one working fiber can be protected at any
time. )
13
Connection-Oriented Networks - Harry Perros 25
WDM optical rings
WDM optical rings can be seen as an extension of
the SONET/SDH rings in the WDM domain.
Many different WDM ring architectures have been
proposed. We examine the following rings:
optical unidirectional path sharing ring (OUPSR),
two-fiber optical bidirectional link sharing ring (2F-
OBLSR)
four-fiber optical bidirectional link sharing ring (4F-
OBLSR).
Connection-Oriented Networks - Harry Perros 26
An optical unidirectional
path sharing ring (OUPSR)
Working
fiber
Protection
fiber
A
B
14
Connection-Oriented Networks - Harry Perros 27
Features
It consists of a working and a protection ring
transmitting in opposite directions
It used as a metro edge ring, and it connects a
small number of nodes, such as access
networks and customer premises, to a hub node,
which is attached to a metro core ring.
The traffic transmitted on the ring is static and
it exhibits hub behavior. That is, it is directed
from the nodes to the hub and from the hub to
the nodes. Static lightpaths are used.
Connection-Oriented Networks - Harry Perros 28
Features
Transmission is unidirectional.
The 1+1 protection scheme is used to
implement a simple path protection scheme.
That is, a lightpath, is split at the source node
and it is transmitted over the working and
protection ring.
The destination selects the best signal.
15
Connection-Oriented Networks - Harry Perros 29
2F-OBLSR and 4F -OBLSR
The two-fiber and four-fiber optical bidirectional
link shared rings are used in the metro core where
the traffic patterns dynamically change.
A signaling protocol is used to establish and tear
down lightpaths.
Protection schemes are implemented using a real-
time distributed protection signaling protocol
known as the optical automatic protection
switching (optical APS)
Connection-Oriented Networks - Harry Perros 30
The two-fiber optical
bidirectional link shared ring
It utilizes two rings transmitting in opposite
direction as in the OUPSR.
The wavelengths is each fiber are grouped
into two sets: one for working wavelengths
and one for protection wavelengths.
If a fiber fails, the traffic is re-routed onto
the protection wavelengths of the other
fiber.
16
Connection-Oriented Networks - Harry Perros 31
The four-fiber optical
bidirectional link shared ring
Working
fibers
Protection
fibers
Node 1 Node 2
Node 3 Node 4
A
B
Span
switching
Ring
switching
Connection-Oriented Networks - Harry Perros 32
Features
It utilizes two working fibers and two protection
fibers.
Protection can be done at both the fiber level or at
the lightpath level.
Fiber protection switching is used to restore a
network failure caused by a fiber cut or a failure
of an optical amplifier. Lightpath protection
switching is used to restore a lightpath that failed
due to a transmitter or receiver failure.
17
Connection-Oriented Networks - Harry Perros 33
Span switching
If the working fiber from node 2 to 3 fails, then all
the lightpaths will be switched onto its protection
fiber from node 2 to 3.
Ring switching
If all four fibers are cut between nodes 2 and 3,
then the traffic will be diverted to the working
fibers in the opposite direction.
In this case, the lightpath from A to B will be
routed back to node 1, and then to node 3 through
node 4.
Connection-Oriented Networks - Harry Perros 34
Mesh optical networks
Both path and link protection can be
implemented in a mesh network.
Link protection can be implemented using
the point-to-point 1+1, 1:1, and 1:N schemes
Path protection is achieved by using
dedicated or shared back-up paths.
18
Connection-Oriented Networks - Harry Perros 35
1+1 path protection
The user signal is split into two copies and each
copy is transmitted simultaneously over two
separate diversely routed lightpaths.
The receiver monitors the quality of the two
signals and selects the best of the two. If one
lightpath fails, then the receiver continues to
receive data on the other lightpath.
Connection-Oriented Networks - Harry Perros 36
1:1 path protection
In the case of the 1:1 path protection, the
user signal is carried over a working
lightpath. The back-up protection lightpath
has also been established, but it is not used.
If the working lightpath fails, the source and
destination switches to the protection
lightpath.
19
Connection-Oriented Networks - Harry Perros 37
1:N path protection
This is a generalization of the 1:1 path
protection, where N different working
lightpaths share the same protection path.
Obviously, only one working lightpath can
be protected at any time
Connection-Oriented Networks - Harry Perros 38
Shared risk link group (SRLG)
An SRLG is a group of links that share the same
physical resource, such as a cable, a conduit, and
an OXC.
Failure of this physical resource will cause failure
of all the links.
When setting up a working and a protection
lightpath, care is taken so that the two lightpaths
are not routed through the same SRLG.
20
Connection-Oriented Networks - Harry Perros 39
An example
The working lightpath from OXC 1 to OXC 2 uses links
{1,6,11} and its protection lightpath uses links {3,8,13}.
That is, they are SRLG-disjoint.
1
2
3
6
7
8
11
12
13
4
5
9
10
OXC 1
OXC 2
Connection-Oriented Networks - Harry Perros 40
The concept of SRLG can also be used in the 1:N
shared protection scheme.
The two working lightpaths {1,6,11} and {2,7,12}
from OXC 1 to OXC 2 are SRLG-disjoint.
Therefore, it makes sense that they both use the
same SRLG-disjoint protection lightpath {3,8,13}.
1
2
3
6
7
8
11
12
13
4
5
9
10
OXC 1
OXC 2
21
Connection-Oriented Networks - Harry Perros 41
The ITU-T G.709
(The Digital Wrapper)
Information is typically transmitted over a
wavelength using SONET/SDH framing and also
Ethernet framing.
In the future, it will be transmitted using the new
ITU-T G.709 standard, otherwise known as the
digital wrapper
Connection-Oriented Networks - Harry Perros 42
Types of traffic:
The standard permits the transmission of
different types of traffic, such as:
IP packets and Gb Ethernet frames using GFP
ATM cells
SONET/SDH synchronous data.
Features of the G.709 standard
22
Connection-Oriented Networks - Harry Perros 43
Bit-rate granularity:
G.709 provides for three bit-rate
granularities: 2.488 Gbps, 9.95 Gbps, and
39.81 Gbps.
This granularity is coarser than that of
SONET/SDH, but is appropriate for terabit
networks, since it avoids the large number of
sub-rate units.
Connection-Oriented Networks - Harry Perros 44
Connection monitoring:
Monitoring capabilities permit to monitor a
connection on an end-to-end basis over several
carriers.
Forward error correction (FEC):
It is used to detect and correct bit errors caused by
physical impairments in the transmission links.
(Useful in under-water transoceanic cables, and
long-haul links across the continent.)
23
Connection-Oriented Networks - Harry Perros 45
The optical transport network
In ITU-T, an optical network is referred to
as the optical transport network (OTN). It
consists of three layers:
Optical channel (Och),
Optical multiplex section (OMS),
Optical transmission section (OTS).
Connection-Oriented Networks - Harry Perros 46
Och
OMS
OTS OTS OTS
The OTN layer structure
24
Connection-Oriented Networks - Harry Perros 47
Optical channel (Och): An optical connection
between two users that uses an entire lightpath.
Optical multiplex section (OMS): Optical channels
are multiplexed and transmitted as a single signal
over a fiber. The OMS is the section between a
multiplexer and a demultiplexer that carries the
combined signal.
Optical transmission section (OTS): This the
transport between two access points over which
the multiplexed signal is transmitted.
Connection-Oriented Networks - Harry Perros 48
The optical channel (Och) frame
The user data is transmitted in frames which
contain several different types of overhead,
the user payload, and the forward error
correction (FEC).
Och overhead Och payload FEC
25
Connection-Oriented Networks - Harry Perros 49
Client payload
Client payload OPU OH
OPU ODU OH
OTU OH ODU FEC
OPU
ODU
OTU
The optical channel overheads
OPU OH (Och payload unit ): It includes information related to the client
signal (i.e. the type of traffic submitted by the user).
ODU OH (Och data unit): It provides tandem connection monitoring, and
end-to-end path supervision
OTU OH (Och transport unit): It includes information for monitoring the
signal on a section
Connection-Oriented Networks - Harry Perros 50
The format of the OTU frame
The OUT frame is arranged in a matrix consisting
size of 4 rows of 4080 bytes.
Data is transmitted serially beginning at the top
left, first row, followed by the second row, etc
1 . . . 7 8 . . . 14 15 16 17 . . . 3824 3825 . . . 4080
row
Column (byte)
1
2
3
4
FAS OTUOH
ODUOH
Client payload
O
P
U
O
H
26
Connection-Oriented Networks - Harry Perros 51
Transmission rates
The following three rates have been defined
(including overheads):
2,666,057 Gbps,
10,709 Gbps,
43,018,413 Gbps
In SONET/SDH, the frame repeats every 125
sec. Higher rates are achieved by
transmitting bigger frames every 125 sec.
Connection-Oriented Networks - Harry Perros 52
In G.709 the frame remains the same, but it is
transmitted at different rates. The three rates
are:
Every 48.971 sec for 2,666,057 Gbps
Every 12.191 sec for 10,709 Gbps
Every 3.035 sec for 43,018,413 Gbps
27
Connection-Oriented Networks - Harry Perros 53
The OPU overhead
Located at rows 1 to 4, columns 15 and 16.
Provides information related to the client
signal. It terminates where the client signal
is originated and terminated.
Connection-Oriented Networks - Harry Perros 54
All bytes are reserved except the payload structure
identifier (PSI). This field transports a 256-byte
message aligned with the ODU multi frame.
The first byte contains the payload type (PT)
which is used to identify the type of payload
carried in the OPU.
1 . . . 7 8 . . . 14 15 16 17 . . . 3824 3825 . . . 4080
row
Column (byte)
FAS OTUOH
ODUOH
Client payload
O
P
U
O
H
1
2
3
4
PSI
FEC
28
Connection-Oriented Networks - Harry Perros 55
The ODU overhead
Located at rows 2 to 4, columns 1 to 14.
The ODU OH provides two important overheads:
the path monitoring overhead, and the tandem
connection monitoring (TCM).
The ODU path monitoring OH enables the
monitoring of particular sections within the
network as well as fault location in the network.
The tandem connection monitoring enables signal
management across multiple networks.
Connection-Oriented Networks - Harry Perros 56
RES: Reserved
TCM/ACT: Activation/deactivation of the TCM fields
TCMi: Tandem Connection Monitoring of ith connection
FTFL: Fault Type & Fault Location reporting channel
PM: Path Monitoring
EXP: Reserved for experimental purposes
GCC: General Communication Channel
APS/PCC: Automatic Protection Switching and protection
communication channel
RES
TCM
/ACT
TCM6 TCM5 TCM4
TCM3 TCM2 TCM1 PM EXP
GCC1 APS/PCC RES GCC2
FTFL
1 . . . 7 8 . . . 14 15 16 17 . . . 3824 3825 . . . 4080
row
Column (byte)
1
2
3
4
FAS OTUOH
ODUOH
Client payload
O
P
U
O
H
1 2 3 4 5 6 7 8 9 10 11 12 13 14
2
3
4
FEC
29
Connection-Oriented Networks - Harry Perros 57
The path monitoring (PM) OH
PM OH occupies columns 10, 11, 12 of row 3.
The following are some of the defined fields:
Trail trace identifier (byte 10): It is used to identify the
signal from the source to the destination. Similar to the
J0 byte in SONET/SDH.
BIP-8 (byte 11): BIP-8 is computed over the whole
OPU and it is inserted two frames later.
Connection-Oriented Networks - Harry Perros 58
The tandem connection monitoring
(TCM) overhead
The TCM OH are defined in row 2, columns 5 to
13, and row 3, columns 1 to 9.
The TCM functionality implemented in the OTN
enables a network operator to monitor the error
performance of a connection that originates and
terminates within its own network, while it
traverses different operators.
30
Connection-Oriented Networks - Harry Perros 59
Carrier
A
Carrier
B
Carrier
A
User B User A
An example of network monitoring:
End-to-end, over several carriers, and per carrier basis monitoring
can be deployed at the same time
Over several carriers
Per carrier basis
End-to-end
Connection-Oriented Networks - Harry Perros 60
FAS and OUT OH
FAS MFAS SM GCC RES
1 . . . 7 8 . . . 14 15 16 17 . . . 3824 3825 . . . 4080
row
Column (byte)
1
2
3
4
FAS OTUOH
ODUOH
Client payload
O
P
U
O
H
1 2 3 4 5 6 7 8 9 10 11 12 13 14
1
FEC
31
Connection-Oriented Networks - Harry Perros 61
Frame alignment
The frame alignment signal (FAS) is carried in the
6-byte frame alignment field
FAS is used by the receiving equipment to identify
the beginning of the ODU frame. The value of
FAS is the same as in SONET/SDH:
F6F6F6282828, and it is transmitted unscrambled.
Connection-Oriented Networks - Harry Perros 62
Multi-frame alignment signal
Some of the OTU and ODU overheads span
several OTU frames.
Because of this, a multi-frame alignment signal
(MFAS) byte is defined in row 1 column 6.
The value of MFAS byte is incremented each
frame, thereby providing a multiframe consisting
of 256 frame.
It is is transmitted scrambled along with the
remaining of the OUT frame
32
Connection-Oriented Networks - Harry Perros 63
OTU overhead
Located at row 1, columns 8 to 14.
Provides supervisory functions for section
monitoring and conditions the signal for transport
between 3R (re-timing, reshaping, and
regeneration) points in the OTN.
Fields:
SM - section monitoring
GCC0 - general communication channel
Reserved for future
Connection-Oriented Networks - Harry Perros 64
Forward error correction (FEC)
The FEC implementation in G.709 utilizes the Reed-
Solomon code RS(255/239).
1 . . . 7 8 . . . 14 15 16 17 . . . 3824 3825 . . . 4080
row
Column (byte)
1
1
1
1
FAS OTUOH
ODUOH
Client payload
O
P
U
O
H
FEC
33
Connection-Oriented Networks - Harry Perros 65
Client signals
The following types of traffic can be
mapped onto the OPU payload:
SONET/SDH
IP and Ethernet over GFP
ATM traffic
Test signals
Connection-Oriented Networks - Harry Perros 66
Mapping SONET/SDH into OPU
STS-48, STS-192, and STS-768 data
streams are mapped onto an OPU payload
using a locally generated clock or a clock
derived from the SONET/SDH signal.
34
Connection-Oriented Networks - Harry Perros 67
Mapping IP and GbE into OPU
This is done using the Generic Framing
Procedure (GFP), as described in the GFP
presentation.
Connection-Oriented Networks - Harry Perros 68
Mapping ATM cells into OPU
A constant bit rate ATM cell stream with a
capacity identical to the OPU payload is mapped
by aligning the ATM cell bytes to the OPU bytes.
Rate coupling maybe necessary.
A cell may straddle over two successive OPU
payloads.
Cell delineation is derived using the HEC field.
35
Connection-Oriented Networks - Harry Perros 69
Control plane architectures
The control plane consists of protocols that
are used to support the data plane, which is
concerned with the transmission of data.
The control plane protocols are concerned
with signaling, routing, and networking
management.
Connection-Oriented Networks - Harry Perros 70
There are two different control plane
architectures:
In the first control plane architecture:
The user is isolated from the network via a
user-network interface (UNI)
The user is not aware of the networks
topology, its control plane, and its data plane.
The nodes inside the network interact with each
other via a network-node interface (NNI).
ATM is a good example of this architecture
36
Connection-Oriented Networks - Harry Perros 71
In the second control plane architecture:
The user is not isolated from the network
through a UNI
The nodes inside the network do not interact
with each other via a separate NNI.
All users and nodes run the same set of
protocols
The IP network is a good example of this
architecture
Connection-Oriented Networks - Harry Perros 72
An optical network provides inter-connectivity to client
networks, such as IP, Frame Relay, ATM and SONET/SDH
Optical Network
IP network
ATM network
IP network
37
Connection-Oriented Networks - Harry Perros 73
A large optical network will typically
consist of interconnected smaller optical
sub-networks, each representing a separate
control domain.
Each of these smaller networks could be a
different administrative system.
Also, the equipment within a smaller
network could all be of the same vendor,
with their own administrative and control
procedures.
Connection-Oriented Networks - Harry Perros 74
UNI E-NNI
Optical subnetwork
I-NNI UNI I-NNI
Optical subnetwork
Client network Client network
Optical Network
Interfaces defined in the first control plane
architecture:
User-network interface (UNI)
Internal network-node interface (I-NNI)
External network-node interface (E-NNI)
38
Connection-Oriented Networks - Harry Perros 75
The OIF UNI
OIF has specified a UNI which provides
signaling for clients to automatically create
a connection.
The UNI is based on LDP and RSVP-TE
protocols
Connection-Oriented Networks - Harry Perros 76
IETF control plane architectures
IETF has defined the following three
control plane architectures:
Peer model
Overlay model
Augmented model
39
Connection-Oriented Networks - Harry Perros 77
The peer model
The peer model utilizes the second control plane
architecture.
That is, the client networks and the optical
networks are treated as a single network from the
point of view of the control plane.
The generalized MPLS (GMPLS), an extension of
MPLS, is used in the control plane.
Connection-Oriented Networks - Harry Perros 78
Routing in the peer model
The IP and optical networks run the same IP routing
protocol, i.e OSPF with suitable optical extensions.
The topology and link state information maintained
by all nodes (OXCs and routers) is identical.
A router can compute an LSP end-to-end.
An LSP can be established using CR-LDP extended
for GMPLS or RSVP-TE extended for GMPLS.
40
Connection-Oriented Networks - Harry Perros 79
The overlay model
This model utilizes the first control plane
architecture.
An IP client network is connected to the optical
network via an edge IP router which has an optical
interface to its ingress optical node, i.e. the optical
node to which it is directly attached.
An edge IP router has to request the establishment
of a connection from its ingress optical node,
before it can transmit over the optical network.
This is done using a signaling protocol.
Connection-Oriented Networks - Harry Perros 80
A connection over the optical network may be a
permanent or a switched lightpath or sub-channel.
The edge router is not aware of the topology of the
optical network nor is it aware of its control and
data planes.
The control plane of the optical network may be
based on GMPLS. However, a strict separation of
the client networks and the optical network is
maintained through the UNI.
41
Connection-Oriented Networks - Harry Perros 81
The augmented model
The IP and optical networks use separate
control planes.
However, information from one routing
protocol is passed to the other.
For instance IP addresses from one IP
network can be carried by the optical
network to another IP network to allow
reachability.
Connection-Oriented Networks - Harry Perros 82
Generalized Multi-Protocol Label
Switching (GMPLS)
GMPLS is an extension of MPLS.
MPLS was designed originally to introduce label-
switched paths into the packet-switched network
GMPLS was designed with a view to applying
label-switching techniques to time-division
multiplexing (TDM) networks and wavelength
routing networks in addition to packet-switching
networks.
42
Connection-Oriented Networks - Harry Perros 83
GMPLS, like MPLS, can be used to setup
an LSP through an IP network and other
packet-switched networks.
It can also be used to:
setup a circuit-switched connection in a
SONET/SDH network.
setup a lightpath in a wavelength routing
optical network.
Connection-Oriented Networks - Harry Perros 84
In GMPLS
IP routers, ATM switches, Frame Relay
switches, Ethernet switches, DCSs and OXCs
are all treated as a single IP network from
the control point of view.
There are no UNIs and NNIs, since GMPLS
is a peer-to-peer protocol.
43
Connection-Oriented Networks - Harry Perros 85
A GMPLS-capable LSR may support the following
interfaces:
Packet-switch capable (PSC) interfaces
Time-division multiplex capable (TDM) interfaces
Lamda switch capable (LSC) interfaces
Fiber-switch capable (FSC) interfaces
GMPLS interfaces
Connection-Oriented Networks - Harry Perros 86
Packet-switch capable (PSC) interfaces:
These are the different interfaces used to receive
and transmit packets, such as IP packets, ATM
cells, Frame Relay frames, and Ethernet frames.
Forwarding of these packets is based on an
encapsulated label, VPI/VCI field, DLCI field.
Time-division multiplex capable (TDM)
interfaces:
They forward data based on the datas slot(s)
within a frame. This interface is used in a
SONET/SDH DCS.
44
Connection-Oriented Networks - Harry Perros 87
Lamda Switch Capable (LSC) interfaces
They forward data from an incoming
wavelength to an outgoing wavelength. This
interface is used in OXCs.
Fiber-switch capable (FSC) interfaces
They forward data from one (or more)
incoming fibers to one (or more) outgoing
fibers. They are used in an OXC that can
operate at the level of one (or more) fibers.
Connection-Oriented Networks - Harry Perros 88
A hierarchy of interfaces
These four interfaces form a hierarchy used
to support hierarchical LSPs
PSC
TDM
LSC
FSC
Highest level
Lowest level
45
Connection-Oriented Networks - Harry Perros 89
Hierarchical LSPs
An LSP may start and end at a packet-
switched interface (PSC) .
It can be then nested together with other
LSPs within an LSP that starts and ends on
a TDM interface, which in turn is nested
(together with other) within an LSP that
starts and ends on a lamda switched
interface (LSC).
Connection-Oriented Networks - Harry Perros 90
An example of a hierarchical LSP
IP router
A
IP router
B
TDM
Switch A
TDM
Switch B
IP router
C
OXC A OXC B
1 GE OC-48
SONET
OC-192
32 OC-192
SONET
OC-192
1 GE
Packet LSP1
TDM LSP2
lamda LSP3
46
Connection-Oriented Networks - Harry Perros 91
The generalized label request
The generalized label request is used to request
the establishment of an LSP.
The following information is carried:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP Enc. Type Switching Type G-PID
Connection-Oriented Networks - Harry Perros 92
LSP Encoding Type (8 bits): It indicates how the data to be
transmitted over the LSP will be encoded. The following values have
been defined.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP Enc. Type Switching Type G-PID
Value Type
1 Packet
2 Ethernet V2/DIX
3 ANSI PDH
4 ETSI PDH
5 SDH ITU-T G.707
6 SONET ANSI T1.105
7 Digital wrapper
8 lamda (photonic)
9 Fiber
10 Ethernet 802.3
11 Fiber Channel
47
Connection-Oriented Networks - Harry Perros 93
Switching Type (8 bits):
It indicates the type of switching that should be performed on a
particular link.
This field is needed on links that advertise more than one type of
switching capability
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP Enc. Type Switching Type G-PID
Connection-Oriented Networks - Harry Perros 94
Generalized payload identifier (G-PID) - 16 bits: It used to identify the
payload carried by an LSP. It is used by the endpoints of the LSP.
Some of the values are:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP Enc. Type Switching Type G-PID
Value Type Technology
0 Unknown All
14 Byte synchronous mapping of E1 SDH
17 Bit synchronous mapping of DS1/T1 SDH
28 POS- No scrambling, 16 bit CRC SONET
32 ATM mapping SONET, SDH
33 Ethernet Lamda, Fiber
34 SDH Lamda, Fiber
35 SONET Lamda, Fiber
36 Digital wrapper Lamda, Fiber
37 Lamda Fiber
48
Connection-Oriented Networks - Harry Perros 95
The generalized label
Several new forms of labels are required to deal
with the widened scope of MPLS into the optical
and time-division multiplexing domains.
The generalized label not only allows for the
familiar MPLS-type label that travels in-band with
the associated packet, but also it allows for labels
which identify time-slots, wavelengths, or fibers.
Connection-Oriented Networks - Harry Perros 96
The generalized label may carry a label that
represents:
Generic MPLS label Frame Relay label, ATM
label
A set of time-slots within a wavelength, or fiber
A single wavelength within a waveband, or fiber
A single waveband within a fiber
A single fiber in a bundle
These new forms of labels are collectively
referred to as the generalized label.
49
Connection-Oriented Networks - Harry Perros 97
Since the node using GMPLS knows the type of
link used, the generalized label does not contain a
type field.
The generalized label is not hierarchical. When
multiple level of labels are required, each LSP
must be established separately.
Format:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Label
Connection-Oriented Networks - Harry Perros 98
The suggested label
This is used to provide a downstream node with
the upstream nodes label preference.
This permits the upstream node to start
configuring its hardware with the proposed label
before the label is communicated by the
downstream node. (Useful, if time to configure a
label is non-trivial).
It can be over-ridden by the downstream node.
Suggested label format: Same as generalized label
50
Connection-Oriented Networks - Harry Perros 99
The label set
The label set is used to limit the label choice
of a downstream node to a set of acceptable
labels.
The receiver must restrict its choice of
labels to one which is in the label set.
Connection-Oriented Networks - Harry Perros 100
There are four cases where a label set is useful in
the optical domain:
Case 1: The end equipment is only capable of
transmitting/receiving on a small specific set of
wavelengths
Case 2: There is a sequence of interfaces which cannot
support wavelength conversion, and require the same
wavelength to be used over a sequence of hops, or even
the entire path.
Case 3: Limit the number of wavelength conversion
along the path.
Case 4: Two ends of a link support different sets of
wavelengths
51
Connection-Oriented Networks - Harry Perros 101
A label set is composed of one or more elements.
Each element is referred to as a subchannel identifier
and it has the same format as a generalized label.
The information carried in a label set is:
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Label type
Subchannel 1
Reserved Action
Subchannel N
.
.
.
Connection-Oriented Networks - Harry Perros 102
Action: This 8-bit field indicates how the label set
is to be interpreted. The following values have
been defined:
Inclusive list (value set to 0)
Exclusive list (value set to 1)
Inclusive range (value set to 2
Exclusive range (value set to 3)
Label type: A 14-bit field used to indicate the type
and format of the labels carried in the object/TLV.
52
Connection-Oriented Networks - Harry Perros 103
Bi-directional LSPs
In MPLS two unidirectional LSPs have to
be established in order to provide bi-
directional connectivity.
Double latency
Twice control overhead
Route selection may be complicated
In GMPLS, bi-directional optical LSPs can
be set-up
Connection-Oriented Networks - Harry Perros 104
Protection information
It is used to indicate the required protection
desired for the LSP., i.e., dedicated 1+1, dedicated
1:1, shared 1:N, unprotected.
Protection information also indicates if the LSP is
a primary or a secondary LSP.
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Reserved S Link flags
53
Connection-Oriented Networks - Harry Perros 105
Secondary (S): A 1-bit field used to indicate that
the requested LSP is a secondary LSP
Link flags: It indicates the desired protection type:
Enhanced: A protection scheme which is more reliable
than dedicated 1+1 should be used, i.e., 4 fiber BLSR.
Dedicated 1+1
Dedicated 1:1.
Shared (1:N)
Unprotected
Extra traffic: It indicates that the requested LSP should
use links that are protecting other primary LSPs.
Connection-Oriented Networks - Harry Perros 106
Protocols for GMPLS
GMPLS is an architecture, and as in MPLS, it requires
a signaling protocol for the reliable distribution of label
bindings.
Both CR-LDP and RSVP-TE have been extended to
support GMPLS. The extensions are presented below.
IS-IS and OSPF have also been extended to support
GMPLS.
54
Connection-Oriented Networks - Harry Perros 107
CR-LDP extensions for GMPLS
New TLVs have been introduced in CR-
LDP to support GMPLS. Specifically,
The generalized label request TLV
The generalized label TLV
The suggested label TLV
The label set TLV
Connection-Oriented Networks - Harry Perros 108
Type
length F
Label
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
U
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP enc. type G-PID Switching type
Type length U F
The CR-LDP generalized label request TLV
The CR-LDP generalized label TLV
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Action Label type Reserved
Sub-channel 1
Sub-channel N

Type length U F
The CR-LDP label set TLV
55
Connection-Oriented Networks - Harry Perros 109
The establishment of a CR-LDP
Label request message
Time
B A E C D
Label mapping
message
Connection-Oriented Networks - Harry Perros 110
Bidirectional LSPs
They are set-up using the same process of establishing a
unidirectional LSP with an upstream label added to the
label request message.
A receiving node provides a new upstream label and then
forwards the request message to the next downstream
node.
As the request message propagates towards the destination
LSR E, labels for the path from LSR E to LSR A are
being setup.
The labels for the path from LSR A to LSR E are setup as
the mapping message propagates towards LSR A.
56
Connection-Oriented Networks - Harry Perros 111
RSVP-TE extension for GMPLS
As in the case of CR-LDP, new objects
have been introduced in RSVP-TE to
support the generalized label operation.
The generalized label request object
the generalized label object
the suggested label object
the label set object
Connection-Oriented Networks - Harry Perros 112
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
LSP enc. type G-PID Switching type
Length Class-Num C-type
Label
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Length Class-Num C-type
0 1 2 3
0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 0 1
Action Label type Reserved
Sub-channel 1
Sub-channel N

Length Class-Num C-type

The RSVP-TE generalized label request object
The RSVP-TE generalized label object
The RSVP-TE label set object
57
Connection-Oriented Networks - Harry Perros 113
Bidirectional LSPs
They are setup using the same process of
establishing a unidirectional LSP with some
additions.
An upstream label is added to the Path message,
which permits the allocation of labels along the
path from the destination LSR to the source LSR.
Labels along the path from the destination LSR to
the source LSR are allocated as in the
unidirectional LSP using the Resv message.
Connection-Oriented Networks - Harry Perros 114
The OIF UNI 1.0
OIF-UNI specifies signaling procedures for clients
to automatically create, delete, and query the
status of a connection over an optical network.
The UNI is based on the GMPLS signaling
protocols: LDP and RSVP-TE.
Also it uses extensions of Link Management
Protocol (LMP)
58
Connection-Oriented Networks - Harry Perros 115
Client:
This is network equipment connected to the
transport network, such as IP routers, ATM
switches, SONET/SDH cross connect
UNI-C and UNI-N
UNI-C is the UNI interface at the clients side
UNI-N is the UNI interface at the networks
side
Connection-Oriented Networks - Harry Perros 116
The UNI data plane
The link between the client and its ingress
node, known as terminal network element
(TNE), is SONET/SDH.
The transmission rate can be up to STS-76.
59
Connection-Oriented Networks - Harry Perros 117
IPCC
UNI signaling messages are transported in
IP packets between the UNI-C and UNI-N.
The UNI defines an IP control channel
(IPCC) for this purpose.
This channel may be
in-fiber or
out-of-fiber.
Connection-Oriented Networks - Harry Perros 118
In-fiber IPCC
The signaling messages are carried over a
common channel embedded in the data-
carrying optical link between the client and
the TNE.
For this, the user can use the line and
section overhead D bytes in the SONET
frame.
60
Connection-Oriented Networks - Harry Perros 119
Out-of-fiber IPCC
The signaling messages are carried over a
dedicated communication link, which is separate
from the link that is used to transmit the data.
Such a link can be over a Ethernet or an IP
network.
UNI-C
Client
UNI-N
TNE
Data link
IP network
Connection-Oriented Networks - Harry Perros 120
Addressing - TNA
The topology, resources, and addressing of the
optical network is not revealed to the clients.
A transport network administrative (TNA)
address is used by the UNI to identify a client.
The TNA address is a globally uniquely defined
address and distinct from the native address spaces
of both the clients and the network.
61
Connection-Oriented Networks - Harry Perros 121
To maintain compatibility with network
devices that use different addressing types,
the TNA may be in the form of IPv4, IPv6,
and NSAP.
IPv4 uses a 32 bit address
IPv6 uses a 128 bit address
NSAP uses 160 bits (see ATM book)
The UNI allows a connection between two
different TNA type address
Connection-Oriented Networks - Harry Perros 122
Services offered over the UNI
The primary services offered to a client by
the transport network over the UNI is the
ability to create and delete connections on
demand.
Also, the following optional procedures are
specified:
Neighbor discovery
Service discovery
62
Connection-Oriented Networks - Harry Perros 123
The UNI abstract messages
A number of abstract signaling messages have
been defined. The actual implementation
depends on whether CR-LDP or RSVP-TE is
used.
These messages are used to create, delete, and
query the status of connections established
over the UNI.
Connection-Oriented Networks - Harry Perros 124
From the UNI point of view, a connection is
a circuit with fixed-size bandwidth between
an ingress and an egress optical node with a
specified frame.
At this moment only SONET/SDH framing
is used.
A connection could be unidirectional or bi-
directional.
63
Connection-Oriented Networks - Harry Perros 125
Abstract messages
Connection create request
Connection create response
Connection create confirmation
Connection delete request
Connection delete response
Connection status enquiry
Connection status response
Notification
Connection-Oriented Networks - Harry Perros 126
Connection create request
This message is sent by the UNI-C to UNI-
N to request the creation of a connection.
The egress UNI-N sends the message to the
destination UNI-C to indicate an incoming
connection request.
64
Connection-Oriented Networks - Harry Perros 127
Connection create response
This message is used to acknowledge
establishment of the connection to the UNI-C that
initiated the connection request.
It is sent from the destination UNI-C to
destination UNI-N, and from the ingress UNI-N to
the UNI-C that initiated the connection request.
The UNI-C client can start transmission upon
receipt of this message.
Connection-Oriented Networks - Harry Perros 128
Connection create confirmation
This message is used from the initiating UNI-C to
its ingress UNI-N to acknowledge completion of
the connection establishment.
The destination UNI-N to destination UNI-C to
indicate that the connection has been successfully
established.
65
Connection-Oriented Networks - Harry Perros 129
Connection establishment
Source
UNI-C
Source
UNI-N
Destination
UNI-N
Destination
UNI-C
Connection
create request
Connection
create response
Connection
create request
Connection
create response
Connection
create confirm
Connection
create confirm
Connection-Oriented Networks - Harry Perros 130
Connection delete request/response
The connection delete request message is used to
initiate the deletion of a connection, and it can be
sent by either UNI-C.
It can also be sent by the network in case of
internal network failure.
The connection delete response message is used to
signal the completion of the deletion of a
connection procedure.
66
Connection-Oriented Networks - Harry Perros 131
Connection status enquiry/response
The connection status enquiry message is
used to query the status and attributes of a
given connection.
The connection status response is used to
return the status of the specified connection
and its attributes.
Connection-Oriented Networks - Harry Perros 132
Notification
It is used by a UNI-N to either UNI-C to indicate a
change in the status of the connection.
67
Connection-Oriented Networks - Harry Perros 133
Attributes
Each abstract message has a number of mandatory
and optional attributes. These are organized
logically into the following groups:
Identification-related attributes
Service-related attributes
Routing-related attributes
Policy-related attributes
Miscellaneous attributes
Connection-Oriented Networks - Harry Perros 134
Identification-related attributes
Some of these attributes are:
TNA address
Source and destination TNA addresses
Logical port identifier
It indicates the client or TNE port number used for the
connection
Local connection ID
An identifier with local significance, assigned by the
UNI-C that initiates the connection. It is used in all
messages to identify which connections the messages
apply to.
68
Connection-Oriented Networks - Harry Perros 135
Service-related attributes
Some of the attributes are:
Encoding type:SONET or SDH.
SONET/SDH traffic parameters
Signal type and concatenation.
Directionality
It indicates unidirectional or bi-directional connection
Generalized payload identifier
It indicates the payload carried within the established
connection.
Connection-Oriented Networks - Harry Perros 136
Service level
It indicates a class of service. Since the optical
network is circuit-switched, the class of service
is not the typical one we encounter in packet-
switching, i.e. packet loss and end-to-end delay.
The class of service is related to issues such as
the restoration scheme, i.e. no restoration, 1+1
protection, etc), and the connection set-up and
hold priorities
69
Connection-Oriented Networks - Harry Perros 137
Routing-related attributes
The only attributes defined is:
Diversity
This attribute specifies at what level a new
connection must be diversely routed from
existing connections which start at the same
TNE. A new connection maybe disjoint from
an existing connections. It can also be routed
identically to an existing connection.
Connection-Oriented Networks - Harry Perros 138
Some of the attributes for the
connection create request
Source TNA (M)
Source logical port identifier (M)
Destination TNA address (M)
Destination logical port identifier (O)
Local connection ID (M)
Encoding type (M)
SONET/SDH traffic parameters (M)
Directionality (O)
Generalized payload identifier (O)
70
Connection-Oriented Networks - Harry Perros 139
LDP extensions for UNI signaling
Two main guiding principles were used
when extending LDP for UNI signaling :
Limit the introduction of new LDP messages.
Only two new messages were introduced to support
the status enquiry messages.
LDP extensions should be easily implemented
as simple addition to the existing LDP
implementation, without violating the LDP
semantics.
Connection-Oriented Networks - Harry Perros 140
LDP session initialization
A single LDP session between the UNI-C
and the UNI-N is established, regardless of
the number of data links between the client
and the TNE.
LDP hello and extended hello are used for
neighbor discovery
71
Connection-Oriented Networks - Harry Perros 141
Connection create using LDP
The connection create request is implemented
in LDP signaling using the label request msg.
This is sent from the source UNI-C to the
source UNI-N, and from the destination UNI-
N to destination UNI-C.
The label request message was extended to
support the signaling of the UNI attributes
Connection-Oriented Networks - Harry Perros 142
Connection establishment using LDP
Source
UNI-C
Source
UNI-N
Destination
UNI-N
Destination
UNI-C
Label request
Label mapping
Label request
Label mapping
Reservation confirm
Reservation confirm
72
Connection-Oriented Networks - Harry Perros 143
Connection deletion
LDP employs two mechanisms for an LSR
to inform its peer to stop using a particular
label:
label withdraw message and
label release message.
Connection-Oriented Networks - Harry Perros 144
Failure detection and recovery
The LDP keepAlive message is used to
detect signaling communication failures
between a UNI-C and a UNI-N.
73
Connection-Oriented Networks - Harry Perros 145
RSVP extensions for UNI signaling
Most of the UNI abstract messages are
directly supported by reusing existing
procedures, messages, and objects defined
in RSVP-TE and GMPLS extensions for
RSVP-TE.
There is no implied requirement that RSVP-
based signaling be supported within the
network.
Connection-Oriented Networks - Harry Perros 146
Connection establishment
UNI-C sends a Path message to its UNI-N.
The Path message includes a
GENERALIZED_LABEL_REQUEST object,
which indicates that a label binding is requested.
The traffic parameters of the connection are
encoded in a SONET/SDH SENDER_TSPEC
object in the Path message and a SONET/SDH
FLOWSPEC object in the corresponding Resv
message.
74
Connection-Oriented Networks - Harry Perros 147
Connection establishment in RSVP
Source
UNI-C
Source
UNI-N
Destination
UNI-N
Destination
UNI-C
Path
Resv
Path
Resv
ReserConf
ReserConf
Connection-Oriented Networks - Harry Perros 148
Connection deletion
A connection in RSVP can be deleted either using
a single PathTear messageor a ResvTear message
and PathTear message combination.
The deletion of a connection may cause the optical
network to think that the connection has failed,
which may lead to management alarms. For this
reason, a graceful connection deletion procedure is
followed.