Background

Efficient resource allocation has been at the center of much of the recent work in the area of restoration. As challenging as this problem may be, it is not the focus of our work. This paper addresses the problem of measuring and comparing the delay inherent to the signaling schemes currently used for lightpath setup. We assume that a path with sufficient resources has been found and the only task remaining is to reserve it. We proceed by making the following assumptions about the network. A cut-through, all-optical (i.e., no intermediate o-e-o conversion) wavelength-routed-network (WRN) is primarily³ composed of fiber-links connecting nodes which contain optical switching elements such as Optical Add/Drop Multiplexers (OADMs) and Optical Cross Connects (OXCs) [2]. The all-optical nature of the signal path allows the network to be transparent and makes it ideal for use as a raw, high-bandwidth Optical Transport Network (OTN). A lower bandwidth IP network, known as the Data Communication Network (DCN) serves as a control plane to the OTN and is primarily used for exchange of signaling messages [3,4]. A DCN node controls an OXC in the OTN and can instruct the OXC to execute a particular cross-connect operation. When a lightpath request arrives at the ingress, it computes a primary path to the egress node and using a signaling scheme, sets up the lightpath by exchanging messages with other controlling DCN nodes. Once the path has been set up completely, the ingress node forwards incoming traffic onto it. When a fault occurs, the flow of traffic to the egress stops and the traffic stream is subjected to heavy packet-loss. In this situation, traffic delivery to the egress node needs to be resumed as quickly as possible [5] in order to minimize packet-loss. We shall focus on single or multiple link or node failures in the OTN. We assume the OTN to be 2-node connected⁴  and the DCN to be fault-free and reliable⁵.