Evaluation of the impact of resource disaggregation in future optical transport networks

The communications industry is experiencing constant transformation. Telecom operators consider Network Disaggregation a potential approach to redesign their current network architecture in order to thrive. It focuses on decoupling the optical transport hardware into independent functional blocks, combined with open-source orchestration systems, to abstract the control layer from the physical layer, acquiring software-based control and automation features. The key perceived benefits include reducing vendor lock-in, drive innovation and evolution, and offer flexibility to deploy the best-in-class equipment that fits the network needs, regardless of the supplier. However, there is uncertainty regarding the performance of a disaggregated system, with the introduction of interoperable open nodes to the network, a demerit on the reach of the propagated optical signals is expected compared to a traditional single-vendor system. In this regard, this thesis evaluates the impact of disaggregation on the network performance, specifically, considering several levels of lightpaths' reach reduction, the average number of connections which are denied service is estimated. Since signal regeneration emerges as the straight answer to overcome the reach limitations, its influence on the behavior of the network is analyzed as well. Moreover, the traffic grooming concept is introduced as an alternative to improve the network performance. The idea is to take benefit of the additional optical to electrical conversions performed for signal regeneration and groom low rate traffic streams in the already established lightpaths, looking to increase the capacity of the network. To this end, a network simulator was developed to test a disaggregated system under several conditions as varying the number of regenerators available in the network, allowing or not traffic grooming, and applying different modulation formats or connections' traffic profiles, all of them are compared to the behavior of a traditional integrated system. Overall the simulation results demonstrate an important demerit of the performance of the network as a function of the transmission reach reduction, forcing the use of additional regenerator nodes. Nevertheless, when implementing traffic grooming with a limited number of regenerators, a significant decrease of the network blocking probability is achieved. Thus, network operators must analyze if this benefit plus the advantages of an open model prevails to the cost generated by the additional OEO nodes