This appendix provides an in-depth explanation of ONS 15454 SDH circuit routing and VC low-order path tunneling in mixed protection or meshed environments, such as the network shown in Figure A-1. For circuit creation and provisioning procedures, see "Circuits and Tunnels."
Figure A-1 Multiple protection domains
Automatic Circuit Routing
If you select automatic routing during circuit creation, Cisco Transport Controller (CTC) routes the circuit by dividing the entire circuit route into segments based on protection domains. For unprotected segments of protected circuits, CTC finds an alternate route to protect the segment in a virtual SNCP fashion. Each path segment is a separate protection domain, and each protection domain is protected in a specific fashion (virtual SNCP, MS-SPRing, or 1+1).
Circuit Routing Characteristics
The following list provides principles and characteristics of automatic circuit routing:
Circuit routing tries to use the shortest path within the user-specified or network-specified constraints. VC low-order path tunnels are preferable for VC high-order path circuits because VC low-order path tunnels are considered shortcuts when CTC calculates a circuit path in extended SNCP mesh networks.
If you do not choose Fully Path Protected during circuit creation, circuits may still contain protected segments. Because circuit routing always selects the shortest path, one or more links and/or segments may have some protection. CTC does not look at link protection while computing a path for unprotected circuits.
For 1+1 and MS-SPRing, if a link is down, a fully protected circuit will be provisioned. The working route uses the link that is up (short or long path), and the protect route uses the link that is down. SNCP circuit routing will not use links that are down. If you want all links to be considered for routing while creating an SNCP ring, do not create circuits when a link is down.
Circuit routing computes the shortest path when you add a new drop to an existing circuit. It tries to find a shortest path from the new drop to any nodes on the existing circuit.
Bandwidth Allocation and Routing
Within a given network, CTC will route circuits on the shortest possible path between source and destination based on the circuit attributes, such as protection and type. CTC will consider using a link for the circuit only if the link meets the following requirements:
The link has sufficient bandwidth to support the circuit
The link does not change the protection characteristics of the path
The link has the required time slots to enforce the same time slot restrictions for MS-SPRing
If CTC cannot find a link that meets these requirements, it displays an error.
The same logic applies to VC high-order path circuits on VC low-order path tunnels. Circuit routing typically favors VC low-order path tunnels because, based on topology maintained by circuit routing, VC low-order path tunnels are shortcuts between a given source and destination. If the VC low-order path tunnel in the route is full (no more bandwidth), CTC asks whether you want to create an additional VC low-order path tunnel.
Secondary Sources and Drops
CTC supports secondary sources and drops. Secondary sources and drops typically interconnect two "foreign" networks, as shown in Figure A-2. Traffic is protected while it goes through a network of ONS 15454 SDHs.
Figure A-2 Secondary sources and drops
Several rules apply to secondary sources and drops:
CTC does not allow a secondary destination for unidirectional circuits because you can always specify additional destinations (drops) after you create the circuit
Primary and secondary sources should be on the same node
Primary and secondary destinations should be on the same node
Secondary sources and destinations are permitted only for regular VC connections (not for VC low-order path tunnels and multicard EtherSwitch circuits)
For point-to-point (straight) Ethernet circuits, only SDH VC4 endpoints can be specified as multiple sources or drops
For bidirectional circuits, CTC creates an SNCP connection at the source node that allows traffic to be selected from one of the two sources on the ONS 15454 SDH network. If you check the Fully Path Protected option during circuit creation, traffic is protected within the ONS 15454 SDH network. At the destination, another SNCP connection is created to bridge traffic from the ONS 15454 SDH network to the two destinations. A similar but opposite path exists for the reverse traffic flowing from the destinations to the sources.
For unidirectional circuits, an SNCP drop-and-continue connection is created at the source node.
Manual Circuit Routing
Routing circuits manually allows you to:
Choose a specific path, not just the shortest path chosen by automatic routing
Choose a specific VC on each link along the route
Create a shared packet ring for Multicard EtherSwitch circuits
Choose a protected path for Multicard EtherSwitch circuits, allowing virtual SNCP segments
CTC imposes the following rules on manual routes:
You cannot create manual low-order path circuits (DS3i or E3 cards).
All circuits, except Multicard EtherSwitch circuits in a shared packet ring, should have links with a direction that flows from source to destination. This is true for Multicard EtherSwitch circuits that are not in a shared packet ring (see Figure A-1).
If you enabled Fully Path Protected, choose a diverse protect (alternate) path for every unprotected segment (see Figure A-3). The white arrow shows the primary path and the black arrows show the alternate path.
Figure A-3 Alternate paths for virtual SNCP segments
For Multicard EtherSwitch circuits, the Fully Path Protected option is ignored. Each high-order path circuit must be manually selected to complete a packet ring.
For a node that has an SNCP selector based on the links chosen, the input links to the SNCP selectors cannot be 1+1 or MS-SPRing protected (see Figure A-4). The same rule applies at the SNCP bridge.
Note The white arrows show the primary path and the black arrows show the alternate path.
Figure A-4 Mixing 1+1 or MS-SPRing protected links with an SNCP
Choose the links of Multicard EtherSwitch circuits in a shared packet ring to route from source to destination and back to source (see Figure A-5). Otherwise, a route (set of links) chosen with loops is invalid.
Figure A-5 Ethernet shared packet ring routing
Multicard EtherSwitch circuits can have virtual SNCP segments if the source or destination is not in the SNCP domain. This restriction also applies after circuit creation; therefore if you create a circuit with SNCP segments, Ethernet node drops cannot exist anywhere on the SNCP segment (see Figure A-6).
Note The white arrows show the primary path and the black arrows show the alternate path.
Figure A-6 Ethernet and SNCP
VC low-order path tunnels cannot be an endpoint of an SNCP segment. An SNCP segment endpoint is where the SNCP selector resides.
If Fully Path Protected is chosen, CTC verifies that the route selection is protected at all segments. A route can have multiple protection domains with each domain protected by a different mechanism.
The following tables summarize the available node connections. Any other combination is invalid and will generate an error.
Table A-3 Multicard Group Ethernet Shared Packet Ring Circuit
# of Inbound Links
# of Outbound Links
# of Sources
# of Drops
Connection Type
At intermediate nodes only
2
1
—
—
SNCP
1
2
—
—
SNCP
2
2
—
—
Double SNCP
1
1
—
—
Two way
At source or destination nodes only
1
1
—
—
Ethernet
Table A-4 Bidirectional VC Low-Order Path Tunnels
# of Inbound Links
# of Outbound Links
# of Sources
# of Drops
Connection Type
At intermediate nodes only
2
1
—
—
SNCP
1
2
—
—
SNCP
2
2
—
—
Double SNCP
1
1
—
—
Two way
At source nodes only
—
1
—
—
VC low-order path tunnel end point
At destination nodes only
1
—
—
—
VC low-order path tunnel end point
Although virtual SNCP segments are possible in VC low-order path tunnels, VC low-order path tunnels are still considered unprotected. If you need to protect VC high-order path circuits either use two independent VC low-order path tunnels that are diversely routed or use a VC low-order path tunnel that is routed over only 1+1 or MS-SPRing links (or a mix of both link types).
Constraint-Based Circuit Routing
When you create circuits, you can choose Fully Protected Path to protect the circuit from source to destination. The protection mechanism used depends on the path that CTC calculates for the circuit. If the network is comprised entirely of MS-SPRing and/or 1+1 links, or the path between source and destination can be entirely protected using 1+1 and/or MS-SPRing links, no extended SNCP mesh network (virtual SNCP) protection is used.
If virtual SNCP (extended SNCP mesh network) protection is needed to protect the path, set the level of node diversity for the extended SNCP mesh network portions of the complete path on the Circuit Creation dialog box:
Nodal Diversity Required—Ensures that the primary and alternate paths of each extended SNCP mesh network domain in the complete path have a diverse set of nodes.
Nodal Diversity Desired—CTC looks for a node diverse path; if a node diverse path is not available, CTC finds a link diverse path for each extended SNCP mesh network domain in the complete path.
Link Diversity Only—Creates only a link diverse path for each extended SNCP mesh network domain
When you choose automatic circuit routing during circuit creation, you have the option to require and/or exclude nodes and links in the calculated route. You can use this option to:
Simplify manual routing, especially if the network is large and selecting every span is tedious. You can select a general route from source to destination and allow CTC to fill in the route details.
Balance network traffic; by default CTC chooses the shortest path, which can load traffic on certain links while other links are either free or used less. By selecting a required node and/or a link, you force CTC to use (or not use) an element, resulting in more efficient use of network resources.
CTC considers required nodes and links to be an ordered set of elements. CTC treats the source nodes of every required link as required nodes. When CTC calculates the path, it makes sure the computed path traverses the required set of nodes and links and does not traverse excluded nodes and links.
The required nodes and links constraint is only used during the primary path computation and only for extended SNCP mesh network domains/segments. The alternate path is computed normally; CTC uses excluded nodes/links when finding all primary and alternate paths on extended SNCP mesh networks.
