Strategies for protected Optical SDH Network
Swapnil P N Srivastava, Wipro Limited
Abstract-High availability is a key requirement for any network. The optical layer of a network is the basic carrier layer and it is responsible for protecting all kind of services being carried on network. There are number of protection schemes suggested by different standardization bodies. This paper talks about the protection schemes at the SONET/SDH layer. The protection schemes for SONET or SDH networks are almost same. In some cases terminologies and granularities are different.
Keywords: APS (Automatic Protection Switching), Unidirectional Path Switched Rings (UPSR), Protection & Restoration, SNC Protection, Bidirectional Line Switched Ring (BLSR), Multiplex Section- Shared Protection Ring (MS-SPRing)
Introduction
The present optical networks are supposed to carry multiple flavors at a very high line rate to which these networks were never subjected in past. The earlier networks had different set of requirements as they were carrying only voice traffic than the today’s network, which carries multiple flavors of traffic at a very high data rate. The present traffic is more sensitive to hits as compared to voice traffic.
Due to the change in traffic flavor new protection schemes like ASON are also getting popular. But the earlier protection schemes are very popular and even today many parts of a network operate with the same protection schemes. In this paper the basic protection techniques used in SDH networks is discussed in liner and ring topology.
Protection & Restoration
The telecom network has an inherent requirement of being the carrier grade network. The carrier grade network in general has a maximum stipulated down time less than 0.001% of the total time of use. This means that the carrier grade network has to be up for at least 99.999% time on average. The optical network carries most of the services in a network at physical layer. This makes an inevitable choice of making the optical network 99.999% available or high available (commonly used term for this kind of network), which means the optical network has a downtime permission of maximum five minutes in a year.
In an optical network there are two kinds of possible faults; first is the fault in equipment and other is the fault in fiber i.e. fiber cut. If we analyze the equipment related faults, these may cost user few minutes to several hours to repair depending upon the nature of fault and availability of spares. In a SDH or WDM system the easiest and most effective way of protection against the possible module failures is provisioning of redundant hot standby modules to each of the working module.
The second case i.e. fiber cut is the main cause of failure of optical networks in India. Due to massive highway expansions and construction activities, fiber cut is the major cause of network failure. The fiber repairing time is also high because it requires digging, locations are remote (most of the time) and at many of the location it is not feasible to dig during daytime in urban areas. The protection for fiber network is given by transmitting the signal on multiple fibers and there are many schemes for protection based on the topologies, number of fiber available and speed of transmission.
Unlike earlier optical network where voice or TDM traffic was the only type of traffic carried on the network, present optical network typically carries more than 60% data traffic in form of Ethernet, MPLS, ATM and HDLC etc. Most of this kind of traffic is comparatively less sensitive to latency and has a capability to reroute and run protection algorithm on the higher layer of OSI model. In that way current traffic is protected generally on two layers. In this paper, focus is only on the optical layer of SDH network. The SDH network can have protection at channel layer or multiplex section layer. In this paper focus is on MS Section protection as the MS Section protection works as link level protection.
Protection and Restoration are two closely related terms. Protection is a paradigm in which dedicated scheme for the protection, based on the backup resource is deployed. The Restoration is a paradigm in which there is no dedicated backup resource but a pool of resource is there to be utilized as backup.
Protection in a Linear Point to Point SDH Network: APS (Automatic Protection Switching) is the most common form of protection in a linear topology. It has two ways of protection viz. 1:1 and 1+1 protection. 1:1 protection mechanism has a variant of 1:N protection as well. In 1:1 Protection, there are two fibers for each direction. One fiber is
working and another one is protection fiber. The switch matrix changes the transmit and receive fiber if any fault occurs in working fiber. This switching action is supposed to happen in less than 50ms as per ITU-T recommendation.
In 1: N topology there is only one fiber working as protection fiber for N number of fibers.
1+1 protection mechanism is implemented by transmitting the signal on both the fibers; working and protected but the receiver decides the fiber from which it has to receive the data based on the health of signal.
SNC protection
A subnetwork is a single layer of network. This includes a set of network elements or a different subnetwork itself. A subnetwork may be constituted from one network element to multiple sub-networks itself. Subnetwork connection protection is a dedicated protection mechanism that can be used on any physical topology. It may be applied at any path layer in a layered network. It can be used to protect a portion of a path (e.g. in case of more than one path available), or the full end-to-end path. It switches on server failures (using inherent monitoring) or it switches using client layer information (using non-intrusive monitoring).
In a subnetwork the protection uses the philosophy of unidirectional 1+1 protection. When it is performed by end devices, it is not a bandwidth efficient mechanism of protection but for a finer granularity it is a good idea to use this scheme in 1:N method.
The following figure is an example of 1: N subnetwork protection. In this figure the two protection connection share the capacity on the links between NE A and NE C. Here the subnetwork provides two diverse paths for protection and restoration.
Ring architecture and self- healing mechanism
The most popular form of carrier topology is the ring networks. Even a mesh type network can be treated as number of rings operating simultaneously. These ring-based architectures are capable of protecting the traffic automatically by detecting and rerouting the traffic so that the traffic remains unaffected. Based on the number of fiber and the philosophy of protection, there are two type of protection mechanism available for ring: UPSR and BLSR.
a. Unidirectional path switched rings (UPSR)
This is a simple protection mechanism. As shown in figure it has two fibers, one fiber being working and another being protection. The working fiber is carries traffic in clockwise direction and protection fiber carries traffic in opposite direction. To understand the traffic flow, let NE A to NE B be the example. The traffic flowing from the working fiber flows directly from A to B but the traffic from B to A on working fiber flows from B to A via NE C and D. If the working fiber between A and B gets disconnected, then the traffic from A to B flows via NE D and C till B on the protection fiber.
b. BLSR or MS-SPRING
Multiplex Section shared protection ring (MS-SPRING) is also known as Bidirectional Line Switched Rings (BLSR) in SONET. It has an advantage of bandwidth efficiency over UPSR. This protection scheme works on multiplex section layer and line layer. In this scheme two fibers are used as working fiber, and two fibers are used as protection fiber. The direction of traffic flow in working and protection fiber is opposite to each other i.e. if working fiber is carries traffic clockwise, the protected fiber carries the traffic anti-clockwise.
This scheme supports a maximum of 16 nodes in the ring numbered from 0000 to 1111. The order of node is not important in numbering. The nodes may appear in any order irrespective of their id/number. In case of a working traffic’s fiber cut in any span, the traffic is retrieved from the protection fiber. With a distance less than 1200Kms the switching time is targeted to be less than 50ms.
In case of BLSR or MS-SPRing using 2- fibers, the half bandwidth of each fiber carrying the working traffic is considered to be protection fiber.
Conclusion
There are number of protection mechanism in SONET/SDH layer of network. Along with the other protections at client layer, these network elements form a highly survivable network. The protection switching time is targeted to be less than 50ms. Theoretically, Four- fiber MS-SPRing is the most reliable protection philosophy.
