|Resumo: ||In the past few years, Elastic Optical Networking (EON) emerged as the next generation core network technology, intended to surpass Wavelength-Division Multiplexing (WDM) weaknesses and limitations. WDM is the most successful and widely used technology in the backbone of the optical networks. However, in recent years Internet traffic in the core network has been doubling almost every two years, and predictions indicate that it will
continue to exhibit exponential growth due to emerging applications such as high-definition and real-time video communications.
To keep pace with the always greater demand for bandwidth, EON relies on Optical Orthogonal Frequency Division Multiplexing (OOFDM) and advanced modulation technologies which enhance spectral efficiency and flexibility. OOFDM allows the aggregation of multiple sub-carriers to form super-channels, thus changing the paradigm of the network from fixed-size WDM channels to variable-sized EON channels that can reduce spectrum
waste up to 60%. EON presents several other benefits such as high spectral and energy efficiency and flexible bandwidth adaptation over time. Despite all benefits, no technology is perfect, and the added EON efficiency and flexibility comes at the price of increased complexity and new problems, such as spectrum fragmentation and service unfairness.
A considerable amount of work has been done on both fragmentation and unfairness problems, introducing a broad range of solutions, which raises the following question: "how to compare existent solutions and how to identify which one is better suited for the required scenario?" In this context, it is presented the first contribution of this Doctoral Thesis, ElasticO++, an Elastic Optical Network Simulation Framework for OMNeT++. ElasticO++ is a framework created to enable testing a whole range of routing, modulation, spectrum assignment, defragmentation algorithms, parameters, and topologies. At present, the proposed framework is the first software available capable of working with fragmentation and defragmentation in dynamic network scenarios. The flexibility offered on the proposed
tool allows both academia and industry to develop new algorithms and techniques for Elastic Optical Networks.
The second contribution of this Doctoral Thesis is the Zone-Based Spectrum Assignment Algorithm. The proposed algorithm is an attempt to solve the unfairness and fragmentation problems, taking advantage of the spectrum management concept. In this Doctoral Thesis, it is presented two versions of the technique: the static version and the dynamic version.
The static version is intended to be used in cases where the information regarding the nature of the network traffic is known beforehand, whereas the dynamic version was developed as a solution in cases in which absolutely no information is known.|