-1-
Appendix I – OTN network topology utilizing Transitional Links
This appendix provides a specific networking example illustrating the use of Transitional Links.
The intent is to demonstrate the usage of the Transitional Link construct to represent various
network architectures from a routing topology perspective. The modeling shown below is provided
to indicate the general direction of Q12’s work, and represents a work in progress.
Introduction
In G.800, the Transitional Link allows for the representation of inter-layer relationships that exist in
a multi-layer network topology for the purposes of performing path computation. This construct is
also extremely useful in networks that support multiple different mappings of a signal (i.e., resulting
in multiple instances of a single ODU layer at different bit rates). The multiple mapping cases can
be caused by a number of conditions:
-
Heterogeneous support of adaptations in the network
-
Heterogeneous server layer technologies in the network.
Considering an end-to-end connection spanning multiple domains, a client signal could be mapped
into a low-rate ODU at the OTN boundary, and then have that ODU respectively mapped into
progressively higher rate ODUs as it progresses across domain boundaries, before it is
demultiplexed back down as it approaches its termination point. This appendix explores the usage
of Transitional Links in representing such topologies, from a path computation perspective, in
ASON networks.
Background
As discussed in Amendment 2 of G.872, the heterogeneous OTN multiplexing hierarchy supports
various network architectures, including those optimized to minimize stranded capacity, minimize
managed entities, support carrier’s carrier scenarios, and/or enable ODU0/ODUflex traffic to transit
a region of the network that does not support these capabilities.
This document elaborates on a few cases, and is not intended to provide an exhaustive compilation
of all possible cases.
There are heterogeneous network environments for which it is necessary to enable carriage of the
newly defined G.709v3 (12/2009) signal types (e.g. ODU0 and ODUflex using 1.25G Tributary
Slots) across an existing network based on G.709 (3/2003). Thus, these existing networks do not
support ODU0 nor do they support use of 1.25G Tributary Slots in their ODU2 and ODU3 links.
One way of accommodating this is by taking the new signals and multiplexing them in a G.709
(3/2003) signal for transmission though the G.709 (3/2003) network. In doing so, there is clearly a
possibility for more than one stage of multiplexing between the client ODU0/ODUflex signals and
the higher-rate ODUk signals (being carried directly over their respective OTUk signals).
Thus, it is clear that routing solutions must be capable of handling all feasible network architecture
and equipment scenarios, including those involving more than one stage of multiplexing.
The architecture of ASON (G.8080) supports the notion of “potential” connectivity. A potential link
connection is a member of a link, and enables the link to be selected during path computation. Any
actions needed to configure the link connection are delayed until signalling time (when resources
are finalized along the computed path). This allows paths to be computed before resources are fully
configured and enables different layer networks to share a common server layer before (flexible)
adaptation has been configured.
-2The transitional link similarly presents potential resources to path computation before the
configuration of those resources has been finalized. It is not necessary that the adaptation functions
that will eventually support the transition have been fully configured before path computation can
use the transitional link.
In this document, all discussion focuses on the topology before a new path has been setup. Just as in
single layer operations, the topology available to the next request can change as a result of the set
up of this request. If, for example, a particular link becomes exhausted as a result of this request,
that link may be advertised with zero capacity.
The behavior for transitional links is exactly analogous to that for links, with the additional
possibility that setting up a server connection between a pair of transitions can add a new client link
into the topology. Because this document is focused on topology before signalling, no further
discussion is provided on operations occurring during and after signalling.
Example G.805 model for Single-stage and Multi-stage multiplexing
This section provides Box/Line figures that provide alternate representations of the G.805
examples. These figures depend on the following two box-level detailed descriptions, which are
color coded in Figure 1.
ODU2
ODU3
ODU2
ODU2
ODU1
ODU2
ODU1
OTU2
termination
OTU3
ODU1
ODU2
termination and
ODU (de)muxing
ODUflex
OTU2
ODU3 termination and ODU
(de)muxing
ODUflex
ODU1
OTU3 termination
ODUflex
ODU1
ODU0
OTU2 termination
ODU3 termination and ODU (de)muxing
OTU3 termination
OTU3
ODU1
ODU0
ODU2 termination and ODU (de)muxing
ODU0
OTU2
ODU2
ODU3
ODU2
G.709 (3/2003)
ODU3
ODU3
G.709 (12/2009)
Figure 1. Box Description contents
The rest of this section will discuss the example below. Both box and functional diagrams are
shown. Note that these figures are provided to only discuss topology, and they do not make use of
the new models provided in G.709 (2010). In addition, the new ODUflex capability is not discussed
and is not shown. While G.709 (2010) models a single, rate independent matrix, topology is
sensitive to the rate dependent fragments, which are depicted here.
-3G.709
(12/2009)
OTU3
ODU0,1,2,3
Switching
G.709
(3/2003)
OTU2
ODU1,2,3
Switching
G.709
(3/2003)
OTU3
G.709
(12/2009)
OTU2
ODU1,2,3
Switching
ODU0
OTU3
ODU0,1,2,3
Switching
ODU0
ODU1
ODU1
ODU1
ODU1
ODU2
ODU2
ODU2
ODU2
ODU3
ODU3
ODU3
ODU3
OTU
2
OTU
2
OTU
3
OTU
3
OTU
3
Figure 2. Box and corresponding G.805 diagram
For this network fragment, one possible corresponding routing topology with Transitional link is as
follows:
A0
D0
TL2/0
TL2/0
A1
TL2/1
TL2/1 B
1
TL3/1
A2
B2 TL3/2
A3
B3
TL3/1
C1
TL2/1
TL2/1 D
1
TL3/2 C2
D2
C3
D3
Figure 3. Corresponding Transitional Links
Transitional links are identified with an icon on the link, and in Figure 3 are labelled with the type
of adaptation they represent.
While transitional links can make any and all equipment capability available to path computation, in
general operator policy is likely to be applied to limit which variability is actually presented. This is
no different to operator policy being added to topology in single layer networks. For example, a
Shared Risk Group is no more than operator policy added to a link. Thus, the use of none, some, or
many, transitional links in a routing topology can be done by applying operator policy. Figure 4
shows an example of the effect of policy on the topology in Figure 3 where there are fewer
transitional links present.
A0
D0
TL2/0
TL2/0
B1
C1
TL2/1 D
1
A2
B2
C2
D2
A3
B3
C3
D3
A1
TL2/1
Figure 4. Pruned Transitional Links
-4Note that the representation of transitional links in multilayer topology as shown in Figure 3 does
not proscribe or define how they are encoded in a routing protocol.
Example Single-layer Routing Topology
The resulting advertisement for the network in Figure 2 above is shown in Figure 5 below:
w ODU0 w ODU0
x ODU1 x ODU1
y ODU2 y ODU2
z ODU3
x ODU1 x ODU1
y ODU2 y ODU2
z ODU3
A
B
x ODU1
y ODU2
z ODU3
x ODU1
y ODU2
w ODU0 w ODU0
x ODU1 x ODU1
y ODU2 y ODU2
z ODU3
C
D
Figure 5. Efficient representation
In past versions of G.8080, there is a requirement for routing to be performed for a specific layer
network. In order to facilitate the G.8080 requirement, G.7715.1 requires the information for a
specific layer be discerned from the efficient representation. This can be performed through use of
a per-layer filtering function. The resulting topology for each of the layers shown in Figure 5 is
shown in Figure 6.
w ODU0
w ODU0
w ODU0
A
x ODU1
B
x ODU1
z ODU3
z ODU3
A
D
y ODU2
y ODU2
B
x ODU1
x ODU1
C
y ODU2
y ODU2
A
w ODU0
D
x ODU1
x ODU1
B
y ODU2
y ODU2
x ODU1
x ODU1
A
C
y ODU2
y ODU2
C
D
C
D
z ODU3
B
z ODU3
w ODU0 w ODU0 Routing views derived from an efficient encoding
w ODU0
w ODU0
Figure 6. Layer-specific
(per G.7715/G.7715.1)
x ODU1 x ODU1
x ODU1 x ODU1
x ODU1
x ODU1
x ODU1
x ODU1
y ODU2
z ODU3
y ODU2
y ODU2
y ODU2
z ODU3
y ODU2
z ODU3
y ODU2
y ODU2
y ODU2
z ODU3
Note in Figure 6, the ODU0 link between A and B is shown because the efficient encoding of
Figure 5 only shows ODU0 capability on node A. The resulting per-layer filter ends up showing
the link, but only places capacity information on the node A link end. This capacity gives rise to the
transitional link TL 2/0.
A
B
C
D
Example Routing Topology for Multistage Multiplexing
In Figure 6, two stage multiplexing capability may be possible. This is illustrated in Figure 7.
-5-
ODU0
1-2
ODU1
1-4
2
1
1-32
1-16
ODU1
1
1-8
1-4
1-8
1-2
1-2
1-8
2
2
ODU2
1-4
ODU2
3
3
ODU3
ODU3
OTU3
OTU2
OTU3
Figure 7. Two-stage ODU muxing in left node in both OTU3 and OTU2 interface ports
Figure 8 shows the complete capability of the equipment in network defined in Figure 6 with two
stage mutliplexing shown in Figure 7.
A0
TL1/0
TL1/ 0
TL2/0
TL2/1+1/ 0
A1
TL2/1
D0
TL2/ 0
TL2/ 1 B1
TL3/ 1
A2
B2 TL3/2
A3
B3
TL3/1
C1
TL2/1
TL2/ 1+1/0
TL2/1
D1
TL3/2 C
2
D2
C3
D3
Figure 8. Topology using link-attribute form
While Figure 8 shows the complete equipment capability, many of the Transitional Links shown are
not useful and can be pruned from the routing topology based on Network Operator Policy. For the
remaining links, Network Operator Policy will also dictate the costs advertised, making some paths
through the network desirable over other paths. An example resulting topology is shown in Figure
9. Note that no connections have yet been configured and allocated in the topology. In Figure 9 TL
2/1+1/0 may not be preferred by operator policy over TL 2/0, e.g., by setting cost.
-6-
A0
A1
TL1/0
TL1/ 0
TL2/0
TL2/ 0
TL2/1+1/0
TL2/1
TL2/1+1/0
TL2/1
D0
B1
C1
D1
A2
B2
C2
D2
A3
B3
C3
D3
Key:
Non-preferred TL
Figure 9. Topology after applying Network Operator Policy
Specific Network Example
Two network scenarios are illustrated in Figure 10 that results in more than one stage of
multiplexing end-to-end across the network:
1) Tunneling of G.709 (12/2009) signals (e.g., ODU0 and ODUflex) through a G.709 (3/2003)
network (e.g., an ODU3 network which couldn’t support 1.25G TS)
2) Carrier’s Carrier, utilizing a higher order ODU connection (e.g., an ODU2) that is carried
over a yet higher speed connection (e.g., an ODU3 mapped into an OTU3) to provide a
variety of lower order connections (e.g., ODU0, ODU1, ODUflex, and ODU2).
It should be emphasized that these are simply illustrative scenarios. They should not be interpreted
as the only scenarios that can result in more than one stage of multiplexing.
These two scenarios are described by the Red and Blue connections shown in Figure 10. The color
legend in Figure 10 is as follows:
ODU0/flex
XC
A node supporting ODU0 and ODUflex switching
capability based on 1.25G TS
ODU0/flex
XC
A node supporting ODU0 and ODUflex switching
capability based on 1.25G TS. It also support the
multi stages multiplexing capability.
ODU1/2 XC
A node only supporting ODU1 and ODU 2 switching
capability based on 2.5G TS.
-7GigE 2
ODU0/1/flex
XC
12
11
ODU0/1/flex
XC
ODU 2
Network 4
（1.25G TS）
13
STM-16 1
ODU0/1/flex
XC
G4
ODU012
XC
2
OTU3
GigE 2
ODU0/1/flex
XC
ODU12
XC
1
STM-16 1
GigE 1
ODU0/1/flex
XC
ODU 2
Network 1
（1.25G TS）
ODU0-ODU1-ODU3
ODU0-ODU2-ODU3
ODU1-ODU2-ODU3
ODUflex-ODU2-ODU3
G1
4
ODU12
XC
8
ODU0-ODU2-ODU3
ODUflex-ODU2-ODU3
G3
7
ODU 3
Network 2
（2.5G TS）
ODUflex 1
ODU0/flex
XC
5
ODU012
XC
OTU3
ODU0-ODU1-ODU3
ODU0-ODU2-ODU3
ODU12
XC
OTU3
ODU02
XC
10
ODU 2
Network 3
（1.25G TS）
ODU0/flex
XC
GigE 1
ODUflex 1
3
ODU0/1/flex
XC
6
9
ODU12
XC
ODU0/flex
XC
Figure 10. Example Network Deployment Utilizing Multi-stage Multiplexing
To increase understanding of the deployment scenario, some further information is provided
regarding implementation options, and illustrative end-to-end service examples. The intent of this
discussion is not to restrict the equipment implementation, but rather to provide some examples
about how a node providing multi-stage multiplexing capability can be implemented. Modeling
methodology (e.g., usage of transitional links) is independent of equipment design.
G1/G3 Implementation Option Examples
Three options for the implementation of G1/G3 nodes in the following figures are presented for
illustration, and do not represent all possible implementations:
o Option 1: Integrated Line Card and Equipment
In this implementation option, illustrated in Figure 11, a unified cross-connect supports ODU0,
ODU1, ODUflex, and ODU2 cross connection. Line card 1 solely provides single-stage
multiplexing capability. Line card 2 provides the multi-stage multiplexing capability within one
single line card in order to interworking with the 2.5G TS network. Both multi-stage and singlestage multiplexing are integrated in a single line card.
If the line card only supports fixed multiplexing (e.g., only supports ODU0-ODU1-ODU3), the
hierarchy of multi-stage multiplexing cannot be configured.
Note: If the 2.5G TS network just switches ODU2 directly in order to minimize managed
entities in the core network, ODU1 can be mapped to ODU2, with ODU2 then mapped into
ODU3. So here the line card also provides the ODU1-ODU2-ODU3 multiplexing capability.
Dependent upon the implementation, it may be the case that not every ODUk is equally
available for cross-connection. In Figure 11, for example, the ODU3 must be always terminated
on the line boards.
-8A generic routing solution should be able to clearly differentiate between bandwidth that can
only be terminated and bandwidth that can be cross-connected.
Line Card 1
ODU1
ODU1
OD
0-OD
ODU U13
0
U
2U
D
O
0- U 3
D
O
-O
DU
1
U1
OD OD U
U 3 2-
D
OTU3
U
O
2
DU
OD U
O
OTU
2
U
U
ODU0
D
D
D
ODU0
O
O
O
2-
Line Card 2
Unified Cross
O
1D
3
2-OD
U
OD U
Uflex
ODUflex
ODUflex
OD
le
Uf
x- O
O
2-O
DU
D
U
O
2-
D
3
DU
U
3
ODU2
ODU2
G1
Figure 11
o Option 2: Cascade Card
In this implementation option, illustrated in Figure 12, the line card does not provide multi-stage
multiplexing within a single card or only uses single stage multiplexing. In this case the signal
must be back-hauled to another line card that can support the re-muxing. (This involves back-toback multiplexing within the NE).
There is a selector for the different hierarchies of multi-stage multiplexing (e.g., the selection
for ODU0-ODU1-ODU3 or ODU0-ODU2-ODU3). If the line card just supports a fixed
interface (e.g., only support ODU0-ODU1-ODU3), the hierarchy of multi stage multiplexing
can’t be configured. There is no selector within the node.
In the same manner as in Option 1, dependent upon the Option 2 implementation it may be the
case that not every ODUk is equally available for cross-connection.
A generic routing solution should be able to clearly differentiate between bandwidth that can
only be terminated and bandwidth that can be cross-connected.
Line Card 1
Unified Cross
ODU0
Line Card 4 Line Card 3
ODU0
ODU0-ODU1 ODU1
0
ODU1-ODU3
O
D
U
2-
O
D
U
Line Card 5
OTU
2
U
OD
OD
D
2-O
U2
Uf
lex
ODU0
ODUflex
ODUflex
-O
D
OD
U0
-O
ODU1
U1
ODU1
ODU1
DU
2
OTU3
ODUflex-ODU2 ODU2
OD
-O
U1
D
U2
ODU2-ODU3
ODU1-ODU3
ODU1
ODU2-ODU3
ODU2
ODU2
G1
-9Figure 12.
o Option 3: Cascaded NEs
In this option, illustrated in Figure 13, multi-stage multiplexing is obtained by cascading
network elements (back-to-back multiplexer). The signal has to be back-hauled to equipment
that can support the re-muxing.
Note: The cross-connect in back-haul equipment may not be necessary for ODU1/ODU2 being
mapped into ODU3.
Line Card 1
Unified Cross
ODU0
OTU1
ODU1
ODU1
Ufle
ODU1
ODUflex
U3
OTU3
O
-O
D
ODU0
U1
D
ODU1
U
ODU1
O
0D
-O
2
U1
U
OD
ODUflex
U2
-O
D
x
OD
U2
1
3
DU
OD
D
-O
DU
-O
U2
U0
-O
Line Card 6
U1
OTU
2
OD
U
U
OD
0
Unified Cross
OD
O
D
O
2-
D
Line Card 5
Line Card 4
ODU0
DU
2
OTU2
ODU2
ODU2
ODUflex-ODU2
ODU2
ODU2
Line Card 3
Line Card 6
Backhaul Equipment
G1
Figure 13.
The unified cross connection matrix in back-haul equipment (very often a DWDM system) may just
consist of a few internal physical links.
A generic routing solution must be capable of representing constrained connectivity of electrical
matrixes.
End-to-End Service Use-Case Examples
An example about the end-to-end service, which is configured by connected equipment, is shown in
Figure 14. This example does not present all possible implementation arrangements. Refer again to
Figure 10, and using the same color conventions.
Multi stage multiplexing capability can be integrated in a single line card or the composition of line
card (e.g., G-3). Also using back-haul equipment and/or line cards can provide the multi-stage
multiplexing capability (e.g., G-1 and NE-4 of Figure 14.
- 10 Optical colored network with
3R regenerators or DXC with
colored interfaces used as
regenerators.
Multi-stage multiplexing
obtained cascading network
elements. Overall topology is
asymmetric because of the
different approach on G-1
and G-3
ODU 2
Network-1
TS 1.25
NE-1
NE G709 V3
G-1
NE G709 V2
NE-4
(OTMx.2)
(OTU2)
ODU 3
ODU0
(OTU2)
ODU2
ODU2
(OTMx.2)
ODU0
GE
ODU2
ODU 0
ODU 0
DWDM
ODU 2
ODU 2
ODU 3
(OTMx.3)
(OTU3)
Internal fixed
(cabled) cross
connection
ODU 2
Network-3
TS 1.25
NE-10
NE G709 V3
G-3
NE G709 V3 with muxponder
ODU0
ODU2
ODU2
ODU 0
ODU 0
DWDM
ODU 3
ODU 3
ODU 2
(OTU2)
ODU 0
ODU 3
GE
(OTMx.2)
NE-7
ODU 2
ODU 0
ODU 0
Multi-stage multiplexing. It
may be either a single line
card or a composition of line
cards. The hierarchy is fixed.
ODU-2 terminated. ODU0
switched. Single multiplexing
stage
Figure 14 Network Example.
The functional model representation is provided in Figure 15 below:
G-1
NE-1
NE G709 V3
GE
NE G709 V2
ODU 2
Network-1
TS 1.25
ODU0
ODU2
G-1
ODU3
ODU2
NE-4
DWDM
(OTMx.2)
(OTU2)
(OTU2)
(OTU3)
NE G709 V3 with muxponder
NE G709 V3
ODU0
ODU-3
Network 2
TS 2.5
G-3
NE-10
GE
(OTMx.3)
ODU2
(OTMx.2)
ODU0
ODU 2
Network-3
TS 1.25
ODU2
ODU2
ODU2
NE-7
ODU3
DWDM
(OTMx.2)
ODU3
(OTU2)
Figure 15 Functional Representation.
Application of Transitional Link Construct
Focusing upon the portion of the example network deployment, considering G1 and NE4 from
Figure 10, the routing topology representation using Links and Transitional Links is illustrated in
Figure 16. Note that, independent of the equipment implementation, the model of transitional link
- 11 construct is the same for three implementation options illustrated in Figure 11, Figure 12 and Figure
13.
Note: the modeling methodology must be capable of supporting any equipment design. Figure 16
gives an example of how to model the G1 and NE-4 of Figure 10 into the topology view.
NE-4 cannot support the ODU0 and ODUflex switching capability.
G1 supports the following multi stage multiplexing capability when signal is mapped into ODU3.
Path computation must know multi stage multiplexing capability constraint information.
 ODU0-ODU1-ODU3
 ODU0-ODU2-ODU3
 ODU1-ODU2-ODU3
 ODUflex-ODU2-ODU3
Because the ODU0-ODU1-ODU2-ODU3 multi stage multiplexing hierarchy is not supported
within the G1 NE of Figure 10, the constraint of prohibiting ODU0 in ODU1 in ODU2 must be
expressed in the topology view. However, there are some limitations from G.805 conventions. So
there should be some constraint information attached with Transitional Link between ODU0 and
ODU1 subnetwork (i.e., TL1/0 of Figure 16) to prohibit ODU0 being mapped into ODU1-ODU2ODU3. It depends on the routing encoding to express this constraint information. The path
computation entity has to consider this constraint information. Costs can be associated with TLs for
directing path computation.
NE2Flex2
G1Flex2
NE20
G10
NE21
G11
NE41
NE51
NE22
G12
NE42
NE52
NE23
G13
NE43
NE53
= TL1/0
= TL2/1 & 3/2
= TL2/0 & 3/2
= TL2/0
= TL3/1
= TL 1/0 & 3/1
= TL2/Flex2
= TL3/2
Figure 16 Transitional Link representation