Optimal Time-dependent Spectrum Sharing Between Neighboring Channels In Elastic Optical Networks

With the rapid development of broadband services, cloud computing and data center,optical network is under immense pressure posed by the explosive growth of data traffic.Flexibility and high capacity are the trends for the future optical transport network.Wavelength Division Multiplexing (WDM) technology can multiplex and transmitoptical signals at different wavelengths in the same fiber, and the characteristics of hightransmission capacity, good technical adaptability, and ease of implementation enable itto be one of major transmission technologies in today’s backbone networks. However,WDM technology also suffers from significant drawbacks due to its inflexible grid andcoarse bandwidth granularity. With the development of Orthogonal Frequency DivisionMultiplexing (OFDM) technology, elastic optical networks are receiving more researchinterest recently due to their better spectrum efficiency and flexibility. In the elasticoptical network, lightpaths with flexible bandwidth and adaptive frequency spacing canbe established. The elastic optical network based on the OFDM technology is consideredas a promising technique for the next generation optical network. The key researchcontributions in this thesis are summarized as follows:First, we propose the concept of time-dependent optical spectrum sharing foradjacent elastic optical channels, which adapts elastic optical channels to carrytime-dependent traffic, so as to improve spectrum resources utilization in the elasticoptical network. Second, we apply the concept of time-dependent optical spectrumsharing to the single-link case. We establish an ILP optimization model to optimallyorder a set of elastic optical channels on a single link so as to maximally share opticalspectra between neighboring channels that carry time-dependent traffic demands, therebyminimizing the total required spectrum. We also develop several algorithms, such asNearest Neighboring algorithm for the above optimization problem. Finally, we apply the concept of time-dependent spectrum sharing to a mesh network. We develop two ILPmodels to assign spectra for lightpaths to maximally share optical spectra betweenneighboring channels and therefore minimize the total required spectrum in the wholenetwork. Based on the results, we find that no matter for a single link or a mesh network,the strategy of time-dependent spectrum sharing can save more spectra to improve thespectrum utilization for the elastic optical network.