10.4. Routing with Constraints
As discussed earlier in the chapter, routing in optical networks needs to be cognizant of different resource types and resource constrains. In particular, types and capabilities of optical links and switching nodes are of interest. Also important are some of the optical constraints that need to be met for error-free data transfer over the connections. In the following, we discuss some of the important resources, and the associated characteristics and constraints.
10.4.1. Link Characteristics
10.4.1.1. ENCODING, SWITCHING CAPABILITY, AND CAPACITY AVAILABILITY
A node computing an end-to-end path needs to know the characteristics of different types of links that are available. Some of the parameters that define the characteristics of a link are:
Link encoding: One of the important characteristics of a link is the type of payload it can carry. For example, a link can be completely transparent and can carry any traffic irrespective of its format or bit-rate. Transparent links are available in all-optical (OOO, see Chapter 1) networks only, which are still not widely deployed. In most optical networks, links are opaque and carry specific types of payload. For example, a link can carry SONET, SDH, Gigabit Ethernet, or Fiber Channel payloads. It is also possible for a link to carry more than one type of payload.
Interface switching capability: The payload type by itself does not completely characterize a link. It is also important to identify the switching capabilities of the nodes on both sides of the link. It is possible that the switching nodes on two sides of the link have different switching capabilities. Switching capabilities are associated with switch interfaces and are referred to as interface switching capabilities. The interface switching capabilities that are relevant to optical routing are:
Time-Division Multiplex (TDM) capable: A node receiving data over a TDM capable interface can multiplex or demultiplex channels within a SONET/SDH payload. For TDM capable interfaces, additional information such as switching granularity and support for different types of SONET/SDH concatenation must also be indicated.
Lambda-switch capable: A node receiving data over a lambda-switch capable interface can recognize and switch individual wavelengths (lambdas) within the interface. Additional information, such as data rate, must be indicated for these interfaces. For example, a lambda-switch capable interface could support the establishment of connections at both OC-48c and OC-192c data rates.
Fiber-switch capable: A node receiving data over a fiber-switch capable interface can switch the entire contents to another interface (without distinguishing lambdas, channels, or packets). That is, a fiber-switch capable interface switches at the granularity of an entire interface and cannot extract individual lambdas within the interface.
Available capacity: Available capacity is an important parameter that determines whether a specific connection can be carried on a link. Capacity of links bound by TDM-, lambda-, or fiber-switch capable interfaces is defined in discrete units (e.g., number of available time slots on a TDM link). Besides the total available capacity on a link, additional information such as minimum and maximum units of available capacity may be indicated. All of these are important for making routing decisions.
10.4.1.2. PROTECTION CAPABILITY
Links in an optical network may have different protection capabilities. During path selection, this information is used to ensure that the protection requirements of the connection will be satisfied by all the links en route. Typically, the minimum acceptable protection is specified at path instantiation, and the path selection algorithm finds a path that satisfies at least this level of protection. The following are possible protection capabilities that could be associ ated with a link:
Preemptible: This indicates that the link is a protection link, that is, it protects one or more other links (see Chapter 8). The traffic carried on a protection link will be preempted if any of the corresponding working links fail and traffic is switched to the protection link. This means that only extra traffic can be routed over this link.
Unprotected: This indicates that there is no other link protecting this link. The connections on a link of this type will not be protected if the link fails.
Protected: This indicates that there are one or more links that are protecting this link. There can be different flavors of protection. A link that is protected by another dedicated link is said to have dedicated protection. Multiple links, which are protected by one or more common protection links, are said to have shared protection.
10.4.2. Optical Constraints
Routing in all-optical networks is subject to additional constraints. Specifically, the routing protocols and algorithms need to ensure that signal impairments are kept within acceptable limits.
As discussed in Chapter 1, impairments can be classified into two categories, linear and nonlinear. Linear effects are independent of signal power and affect wavelengths individually. Amplifier spontaneous emission (ASE), polarization mode dispersion (PMD), and chromatic dispersion are examples of linear impairments [Stern+99]. Nonlinear impairments are significantly more complex and generate not only impairments on each channel, but also cross talk between channels. In the following, we briefly discuss different types linear and nonlinear impairments and how they can be abstracted for the purpose of routing.
Polarization Mode Dispersion (PMD): PMD degrades signal quality and limits the maximum length of a transparent segment. With older fibers, the maximum length of the transparent segment should not exceed 400 km and 25 km for bit rates of 10Gb/s and 40Gb/s, respectively. Due to recent advances in fiber technology, the PMD-limited distance has increased dramatically. With newer fibers, the maximum length of the transparent segment without PMD compensation is limited to 10,000 km and 625 km for bit rates of 10 Gb/s and 40 Gb/, respectively. Typically, the PMD requirement is not an issue for most types of fibers at 10 Gb/s or lower bit rate. It cannot be ignored, however, at bit rates of 40 Gb/s and higher. The only link dependent information needed by the routing algorithm is the square of the link PMD, denoted as link-PMD-square. It is the sum of the span-PMD-square of all spans on the link.
Amplifier Spontaneous Emission (ASE): ASE degrades the optical signal-to-noise ratio or OSNR. OSNR level depends on the bit rate, transmitter-receiver technology (e.g., availability of forward error correction [FEC]), and margins allocated for the impairments, among other things. OSNR requirement can translated into constraint on maximum length of the transparent segment and number of spans. For example, current transmission systems are often limited to 6 spans each 80 km long. The only link-dependent information needed for ASE by the routing algorithm is the link noise, which is the sum of the noise of all spans on the link. Hence, the ASE constraint that needs to be satisfied is that the aggregate noise of the transparent segment, which is the sum of the link noise of all links, cannot exceed the ratio of transmit power and OSNR.
There are many other types of impairments that affect signal quality. Most of the impairments generated by network elements such as OXCs or OADMs (see Chapter 1) can be approximated using a network-wide margin on the OSNR.
Besides linear impairments, there are many nonlinear impairments, which affect optical signal quality [Stern+99]. It is extremely difficult to deal with nonlinear impairments in a routing algorithm because they lead to complex dependencies, for example, on the order in which specific fiber types are traversed. A full treatment of the nonlinear constraints would likely require very detailed knowledge of the physical infrastructure. An alternative approach is to assume that nonlinear impairments are bounded and result in a margin in the required OSNR level for a given bit rate.