Ionescu, MV; (2015) Digital signal processing for sensing in software defined optical networks. Doctoral thesis , UCL (University College London).

Preview

Text Maria Ionescu - PhD Thesis.pdf Download (4MB) | Preview

Abstract

Optical networks are moving from static point-to-point to dynamic configurations, where transmitter parameters are adaptively changing to meet traffic demands. Dynamic network reconfigurability is achievable through software-defined transceivers, capable of changing the data-rate, overhead, modulation format and reach. Additionally, flexibility in the spectral allocation of channels ensures that the available resources are efficiently distributed, as the increase in fibre capacity has reached a halt. The complexity of such highly reconfigurable systems and cost of their maintenance increase exponentially. Implemented as part of digital signal processing of coherent receivers, sensing is an enabling technology for future software defined optical networks, as it makes possible to both control and optimise transmission parameters, as well as to manage faulty links and mitigate channel impairments in a cost-effective manner. Symbol-rate is one of the parameters most likely to adaptively change according to existing fibre impairments, such as optical signal-to-noise ratio or chromatic dispersion. A single-channel symbol-rate estimation technique is demonstrated initially, yielding a sufficient accuracy to distinguish between different typical error-correction overheads, in the presence of dispersion and white Gaussian noise. Further increasing the capacity over fibre to 1 Tb/s and beyond means moving towards superchannel configurations that employ Nyquist pulse shaping to increase spectral efficiency. Novel sensing techniques applicable to such information dense configurations, that can jointly monitor the channel bandwidth, frequency offset, optical signal-to-noise ratio and chromatic dispersion are proposed and demonstrated herein. Based on time-domain and frequency-domain functions derived from the theory of cyclostationarity, the performance of this joint estimator is investigated with respect to a wide range of parameters. The required acquisition time of the receiver is approximately 6.55 μs, three orders of magnitude faster compared to the round-trip time in core networks. The pulse shaping at the transmitter limits the performance of this estimator, unless the excess bandwidth is 30% of the symbol-rate, or more.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as CSVXMLJSON

Downloads by country - last 12 months

Export as XMLCSVJSON

Archive Staff Only

View Item