PROTECTION SCHEMES 211
Working fiber
Protection fiber
A
B
Figure 9.5 An optical unidirectional path sharing ring OUPSR.
static rings to advanced dynamic rings. Below, we examine the protection scheme of three WDM rings: the optical unidirectional path sharing ring OUPSR, the two-fiber optical
bidirectional link sharing ring 2F-OBLSR , and the four-fiber optical bidirectional link
sharing ring 4F-OBLSR .
OUPSR is unidirectional. It consists of a working and a protection ring transmitting in opposite directions see Figure 9.5. 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 is transmitted over the working and protection rings see Figure 9.5 from A to B. The
destination selects the best signal. When a fiber link is broken, the receiver continues to receive the signal along the other path. The OUPSR provides a simple and robust
architecture without complex protection signaling protocols. This type of ring is typically used as a metro edge ring, and it connects a small number of nodes e.g. access networks
and customer sites 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.
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, and protection schemes are implemented using a real- time distributed protection signaling protocol known as the optical automatic protection
switching optical APS .
The two-fiber optical bidirectional link shared ring 2F-0BLSR uses two rings, trans- mitting in opposite directions as in OUPSR. Each fiber is partitioned into two sets of
wavelengths; one set of working wavelengths and one set of protection wavelengths. If a fiber fails, the traffic will be rerouted onto the protection wavelengths of the other fiber.
The four-fiber optical bidirectional link shared ring 4F-OBLSR uses two working fibers and two protection fibers see Figure 9.6. 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.
Let us consider a lightpath from user A to user B see the solid line in Figure 9.6. This lightpath is routed through nodes 1, 2, and 3. Assume that the lightpath fails on
the link between nodes 2 and 3. In this case, the protection mechanism will switch the lightpath over to the protection fiber from nodes 2 to 3 see the dotted line labeled “span
switching” in Figure 9.6. If the working fiber from nodes 2 to 3 fails as well, then all of the lightpaths will be switched onto its protection fiber from nodes 2 to 3, as in the
212 WAVELENGTH ROUTING OPTICAL NETWORKS
Working fibers
Protection fibers
Node 1 Node 2
Node 3 Node 4
A
B Span
switching Ring
switching
Figure 9.6 A four-fiber optical bidirectional link sharing ring 4F-OBLSR.
case of the lightpath above. This is known as span switching. When all four fibers are cut between nodes 2 and 3, then the traffic will be diverted to the working fibers in the
opposite direction. This is known as ring switching. In this case, the lightpath from A to B will be diverted; that is, it will be routed back to node 1, and then to nodes 4 and 3.
See the dotted line labeled “ring switching” in Figure 9.6.
9.2.3 Mesh Optical Networks
A mesh network can employ both path and link protection. Link protection can be implemented using the point-to-point 1 + 1, 1:1, and 1:N schemes see Section 9.2.1.
Path protection uses dedicated or shared back-up paths. Alternatively, an arbitrary mesh topology can be organized into a set of WDM optical rings, which permits ring-based
protection schemes.
The 1 + 1 path protection scheme is the simplest form of protection. It is also the most expensive and bandwidth-inefficient. The user signal is split into two copies, and each
copy is transmitted simultaneously over two separate lightpaths. The lightpaths might be diversely routed i.e. they follow different geographical paths or they might go through
the same OXCs but use different fibers. 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.
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 work-
ing lightpath fails, the source and destination switches to the protection lightpath. Since the bandwidth allocated to the protection lightpath is not utilized during normal operation,
it can be shared by multiple working lightpaths. This is the 1:N path protection scheme.
An important concept in these protection schemes is the concept of the shared risk link group SRLG
. An SRLG is a group of links that share the same physical resources, such as a cable, a conduit, and an OXC. Failure of these physical resources will cause failure
of all of the links. Each common physical resource is associated with an identifier called the SRLG. When setting up a working and a protection lightpath, care is taken so that
the two lightpaths are not routed through the same SRLG. For example, let us consider the optical network shown in Figure 9.7. The working lightpath from OXC 1 to OXC 2
THE ITU-T G.709 STANDARD – THE DIGITAL WRAPPER 213
1 2
3 6
7 8
11 12
13 4
5 9
10 OXC 1
OXC 2
Figure 9.7 Path protection.
that uses links {1, 6, 11} and its protection lightpath that uses links {3, 8, 13} do not use the same SRLG. That is, they are SRLG-disjoint.
The concept of SRLG can also be used in the 1:N shared protection scheme. For instance, in Figure 9.7, 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}. This is because a single failure of a physical
resource along the path of either working lightpaths excluding the originating and termi- nating OXCs will not cause both working lightpaths to fail at the same time. That is, in
this case, the protection lightpath will only be used by one of the two working lightpaths.
In protection schemes, the backup protections routes are pre-planned and the neces- sary resources e.g. wavelengths, fibers, and bandwidth within an OXC are allocated
in advance. During the normal operation of the network, these resources are either kept idle, or they are used to transmit low priority traffic which can be preempted any time a
failure occurs. This guarantees a fast recovery from a failure at the expense of inefficient resource utilization. An alternative strategy, known as dynamic restoration, is to calculate
a protection path and allocate resources for recovery at the moment when a network fail- ure occurs. This approach has a more efficient resource utilization but the recovery time
is longer than in the case of a protection scheme. Dynamic restoration is a promising new approach that is being further studied.
9.3 THE ITU-T G.709 STANDARD – THE DIGITAL WRAPPER
Information on a lightpath is typically transmitted using SONETSDH framing. Also, Ethernet frames can be transmitted over an optical network. In the future, it is expected
that information will be transmitted over the optical network using the new ITU-T G.709 standard, otherwise known as the digital wrapper. This standard defines the network node
interfaces between two optical network operators, or between subnetworks of vendors within the same network of an operator. The following are some of the features of the
G.709 standard:
• Types of traffic
: The standard permits the transmission of different types of traffic, such as IP packets and gigabit Ethernet frames using the generic framing procedure GFP,
ATM cells, and SONETSDH synchronous data. •
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 SONETSDH, but is appropriate for terabit networks, since it avoids the large number of low bit-rate paths
that would have to be used with SONETSDH.
