SDH Ring Topologies

Ring topology and alternate routes

A point-to-point link between two sites has no alternate physical route when its fiber span fails. A ring topology connects each node to two neighbors, so traffic can use the other direction around the ring when one span is unavailable. This layout is used in metro and transport networks where continuity after a cable fault is part of the service design.

In a four-node ring, for example, Node A can reach Node C through B or through D. These routes are physically distinct only if the fibers and civil paths are separated. Protection logic selects or reroutes the service after a monitored fault. The network elements that add and remove tributaries while forwarding the remaining traffic are the Add-Drop Multiplexers (ADMs).

Two Ring Protection Models

SDH and SONET ring protection uses two main models. Path protection duplicates a selected service and chooses a good copy at the destination. Line protection reserves or shares capacity and switches a group of services around a failed span.

Path-switched rings (UPSR / SNCP)

In North American SONET terminology, this ring implementation is called a Unidirectional Path-Switched Ring (UPSR). In SDH, the general path-protection function is called Sub-Network Connection Protection (SNCP). Both use duplicated path traffic and destination selection, although SNCP can also be deployed outside a physical ring.

Architecture and Operation

A UPSR is typically a 2-fiber ring where the two fibers carry traffic in opposite directions. The protection mechanism follows the 1+1 "bridge and select" principle:

1. Permanent Bridge: When a service, such as a VC-12 carrying an E1 tributary in SDH or a VT1.5 carrying a T1 tributary in SONET, is sent from Node A to Node C, the ADM at Node A creates a permanent bridge. It transmits identical copies onto both directions simultaneously, one via B and the other via D.
2. Continuous Monitoring and Selection: The destination ADM at Node C receives both copies of the service, one from each direction. Its internal switching fabric continuously monitors the health of both signals based on information in the Path Overhead. It selects the signal from the pre-designated "working" path to send to its local tributary port.
3. Local Selection: If a fiber cut occurs between nodes B and C, the signal on the clockwise path is lost. The selector at Node C detects the defect, such as Loss of Signal or an excessive error rate, and selects the copy already arriving from the counter-clockwise path.

Characteristics of UPSR/SNCP

• Switching: The destination has a copy available before the fault, so no ring-wide APS coordination is needed for the path selector. The recovery target depends on defect detection and equipment processing, with 50 ms commonly used for specified protection designs.
• Simple selection: The source bridge and destination selector use a small set of local decisions based on path defects.
• Bandwidth cost: Every protected service is duplicated. Both directions carry the service, so the working service capacity is about half of the aggregate two-fiber resources. This can be unsuitable for capacity-intensive core transport but useful when per-service protection is required.

Line-switched rings (BLSR / MS-SPRing)

For core and long-haul networks, a shared line-protection ring can preserve more working capacity than duplicating every service. The SONET name is the Bidirectional Line-Switched Ring (BLSR), and the SDH name is the Multiplex Section-Shared Protection Ring (MS-SPRing). The common arrangement has two fibers.

Architecture and Bandwidth Allocation

A 2-fiber BLSR/MS-SPRing uses a pair of fibers between each node, one for clockwise transmission and one for counter-clockwise. Both fibers carry working and protection channels, and this arrangement supports ring switching. In the common shared-protection arrangement, the capacity on each fiber is logically divided into working and protection portions:

• Working Capacity (50%): Used to carry high-priority service traffic under normal conditions.
• Protection Capacity (50%): Reserved for recovery. It can carry extra traffic when configured, but that traffic is preempted during a protection switch.

Loopback switch after a cable cut

A complete cable cut removes both fibers of a span. The ring protection response is a loopback, or ring switch, that uses the reserved capacity on the remaining route.

1. Normal Operation: Traffic between, for example, Node A and Node C flows on the working channels over the shortest path (e.g., through Node B). Protection channels remain available for recovery and may carry preemptible extra traffic when configured.
2. Failure: A construction crew accidentally severs the entire cable bundle between Node B and Node C. Both fibers are cut.
3. Detection and Signaling: The ADMs at Node B and Node C immediately detect a Loss of Signal. They use the APS bytes (K1, K2) in the Multiplex Section Overhead to broadcast a failure alert to all other nodes on the ring.
4. Coordinated Switch: The nodes adjacent to the break, B and C, execute the ring switch:
   • At Node B: Working traffic arriving from Node A for Node C is bridged to the protection channels of the intact link back towards Node A.
   • At Node C: Working traffic arriving from Node D for Node B is looped back onto the protection channels of the intact link towards Node D.
5. Result: The network changes its bridge state and traffic takes the intact side of the ring through the reserved protection capacity. The failed span is bypassed, while the remaining services retain the line-protection state defined by the APS protocol.

4-Fiber Rings and a Comparison

Operators can deploy 4-Fiber BLSR/MS-SPRing architectures when the service plan requires separate working and protection fibers.

• Architecture: It uses four fibers between each node: two working fibers, one in each direction, and two dedicated protection fibers, also assigned by direction.
• Capacity: With the same line rate per fiber, it can provide twice the working fiber capacity of the common 2-fiber shared arrangement.
• Protection: A span switch can move traffic from a failed working fiber to its parallel protection fiber. A complete cable cut can require a ring switch, or loopback, over the other nodes.