Optimized Routing and Buffer Design for Optical Transport Networks Based on Virtual Concatenation

The emergence of new telematic applications and services is creating significant growth in the traffic carried over the optical transport network (OTN). One potential cost-effective approach for satisfying these demands consists in applying inverse-multiplexing techniques, such as virtual concatenation (VCAT). In this context, efficient load balancing can be achieved by exploiting multipath routing at the cost of introducing differential delay in the concatenated circuits. This effect can be compensated through appropriate delay compensation techniques (e.g., electrical buffering), which can act either in a centralized way with the buffering at the end nodes, or by distributing the differential delay throughout the intermediate nodes. In order to properly solve the routing and differential delay compensation distribution problem in the OTN, we propose two novel methods: a dual-step integer linear programming (ILP) model and a tabu search multi-stage heuristic. These strategies are compared with our two former proposals consisting of a single-step ILP model and an iterative search heuristic. The four solutions are further expanded to consider the possibility of using homogeneous or heterogeneous VCAT. The performance of these optimization methods is examined in two network topologies assuming that 100 Gb/s Ethernet streams are carried over 40 Gb/s channels (homogeneous case) or over a mixed combination of 40 Gb/s and 10 Gb/s channels (heterogeneous case) in the OTN. The results obtained show that, as expected, the smallest buffering requirements are obtained by both ILP models, with the dual-step version exhibiting also a significant reduction in the number of optical-to-electrical and electrical-to-optical operations used for intermediate compensation. In addition, the novel tabu search framework is able to surpass our former heuristic by simultaneously leveraging minimum link capacities and reducing the buffer sizes.