4.6. Multilayer Issues

As noted above, the networks composing the Internet are multilayer in nature and survivability and restoration can be addressed at multiple layers with different techniques at each layer. However, the majority of the literature focuses on providing survivability for various network technology (WDM, SONET, MPLS, ATM, IP, Overlay/Application) layers independently and multilayer issues are an area of ongoing research [49] that present several challenges.

Consider, for example, a two-layer backbone network such as an MPLS/WDM network. Survivability techniques in two-layer networks can be classified as survivability at the bottom layer, survivability at the top layer, and survivability at both layers, depending on the layer in which the survivability technique is deployed [13]. In the bottom-layer approach, recovery from a failure is performed only at the bottom layer (e.g., recovering failed light paths in the WDM network). This scheme has the benefits that it is simple and provides fast recovery of aggregate traffic. However, the major drawback of this scheme is that it cannot recover from failures that occur in the top layer, such as the failure of a top-layer router or its interfaces. With survivability at the top-layer, failure recovery is performed only at the top layer e.g., by recovering failed label-switched paths (LSPs) in an MPLS network using fast reroute. The advantage of this scheme is that it can recover from failures that occur in both layers. It also allows a service differentiation among top-layer flows by recovering each individual flow in the top layer, which is difficult in the bottom-layer survivability scheme where an aggregate of top-layer flows is recovered. Among the drawbacks of the top-layer approach are its complexity and slower speed of fault recovery.

One of the major problems in the survivability of multilayer networks is fault propagation, which occurs when the failure of a bottom-layer link or node results in the simultaneous failure of multiple higher-layer links. For example, consider Figure 4.13, which shows an MPLS over WDM network. At the MPLS layer, the three nodes are connected into a ring topology that can be provisioned with enough capacity for any single MPLS layer link (i.e., an LSP) to be fault tolerant. However, suppose in the lower WDM layer, the MPLS link 1-3 is routed on the WDM path 1-4-3, the MPLS link 1-5 is routed on the WDM path 1-4-5, and the MPLS link 5-3 is routed on the WDM path 5-3. We can immediately see that the MPLS links 1-3 and 1-5 are not diverse at the lower WDM level. Thus, if the WDM link 1-4 fails at the lower layer, both the upper-layer MPLS links 1-3 and 1-5 will be affected. Note, that fault propagation is essentially a correlated failure at the higher layers. An additional cause of fault propagation is shared-risk link groups (SRLGs), which are defined by a set of links that fail together due to physical placement of cabling in conduits or a common infrastructure [50] (e.g., separate cables crossing a bridge).

If failure propagation is not considered appropriately in multilayer networks, the survivability at the top-layer technique may fail to recover the communication services after a failure. This is an especially important point that has been ignored in much of the higher-layer survivability schemes in the literature. In part due to failure propagation, each layer of a network will typically employ self-healing capabilities to address faults occurring in their own layer. In such a multilayer scheme, coordination between layers is required to provide an efficient recovery process upon a failure. This coordination is called an escalation strategy, which determines which layer will perform a recovery first in response to a particular failure, and when and how responsibility will be transferred to another layer if the current layer fails to recover from the failure [13].

The design of survivable multilayer networks has been considered in the literature [13, 49]. The common theme in the current literature involves an exchange of interlayer topology information. For example, in Liu et al. [51], we consider the problem of provisioning spare capacity in two-layer backbone networks using a shared backup path protection approach in order to meet survivability requirements. First, two SCA optimization problems are formulated as ILP models for protection in the top layer against failures in the bottom layer. Both problems use overlay mapping information between the two layers to determine the location and the amount of spare capacity as well as the backup routes to protect a set of given working routes. The first model captures failure propagation across network layers, so the backup paths meet diversity requirements correctly. The second model improves bandwidth efficiency by moving the spare-capacity sharing from the top layer to the bottom layer. It requires the top-layer backup route be able to reserve bottom-layer spare capacity. This exposes a trade-off between bandwidth efficiency and extra cross-layer operation. In addition, the SCA model for shared path protection at both layers is developed. This approach is also called common pool protection. It allows spare capacity to be shared across layers to further reduce the network redundancy with significant savings possible. However, the implementation of such approaches requires cross-layer signaling and fault management coordination.

While the current literature provides a foundation for general survivable multilayer networks, there are still major open research issues, such as determining what type of survivability mechanism should be implemented at each layer to meet speed and availability requirements, coordination of alarms and resilience mechanisms at the layers, prioritization of traffic for fault recovery within and among the layers, techniques for collection and dissemination of cross-layer topology information, and understanding the overall system behavior to avoid instability when multilayer restoration schemes operate concurrently.