Impact of physical layer impairments on large ROADM architectures

• Autor(a) principal: Morão, Diogo Miguel Cigarro
• Data de Publicação: 2020
• Tipo de documento: Dissertação
• Idioma: eng
• Título da fonte: Repositório Científico de Acesso Aberto de Portugal (Repositórios Cientìficos)
• Texto Completo: http://hdl.handle.net/10071/20973
• topic: Amplified spontaneous emission noise; Bank based add/drop; In-band crosstalk; Largescale ROADMs; Optical filtering; Crosstalk homódino; Estrutura de inserção/extracção bank based; Filtragem ótica; ROADMs de grandes dimensões; Ruído ASE

Most of today’s optical networks, use reconfigurable optical add/drop multiplexers (ROADMs) as nodes. To become more dynamic and flexible, these nodes architectures evolved over the years. The colorless, directionless and contentionless functionalities are now standard, however, current architectures have poor scalability due to limitations on wavelength selective switches dimensions. Hence, due to constant increase in data traffic, current architectures might become a bottleneck to manufacture future large-scale ROADMs. In this work, the hardware cost and in-band crosstalk generation inside different large-scale ROADM architectures, is compared with conventional architectures. Moreover, an analysis of optical filtering, amplified spontaneous emission (ASE) noise and in-band crosstalk impact in the performance of an optical network, with nodes based on the most promising large-scale architecture, the interconnected A architecture, is performed. This performance is assessed through Monte-Carlo simulation with 16 point quadrature amplitude modulation with polarization-division multiplexing (PDM-16QAM) and PDM- 32QAM signals with 200 Gb/s and 250 Gb/s, respectively. Two architectures are considered for the interconnected A express structure, Broadcast and Select (B&S) and Route and Select (R&S). For the add/drop structure, a bank-based structure is considered. The maximum number of cascaded ROADMs, considering all the studied impairments, is 5 and 7 nodes for a 32 GBaud 16QAM signal, respectively, for B&S and R&S architectures. A 32QAM signal reaches 3 and 4 nodes, respectively, for B&S and R&S architectures. The main penalty in transmission is the ASE noise generated by optical amplifiers throughout the network, having the in-band crosstalk and optical filtering penalties a lower contribution.