TY: THES T1 - Transceivers for TWDM-PON A1 - Bento, Susana Gonçalves N2 - In recent years, Internet has been assuming a fundamental role in everyday life. Traffic demands are increasing in such a way that the available technologies will presumably no longer satisfy the raised requirements. For the last years, operators have expressed a clear interest in the implementation and development of Passive Optical Network (PON) to provide several services and applications to a high flow rate per client. Comparing to other access technologies, PON is very attractive mainly due to reduction of maintenance and to the operational cost efficiency. As a consequence, PON systems were standardized and developed in the whole world, but the everincreasing bandwidth demand makes this type of network need to evolve. Therefore, the current standardized technologies Gigacapable PON and XG-PON need to be upgraded to Next-Generation PON2. In order to protect the initial investment and to reduce the operational costs, operators should keep the current optical distribution network, providing the technologies coexistence in the same fiber. The principle of NG-PON2 is to improve previous technologies, in terms of capacity, ODN compatibility, bandwidth and cost-efficiency. In April 2012, Full Service Access Network (FSAN) selected the time and wavelength multiplexing PON (TWDM-PON) technology as the solution of choice for NG-PON2. Almost one year later, ITU-T G.989.1 came out, providing some wavelength plans proposals. The ability to operate on existing fiber ODN, coexisting with legacy PON is the most important requirement. The current dissertation is based on the study of TWDM-PON upstream transmission. Both GPON and XG-PON work in burst mode for upstream direction, therefore in the current study also that type of data transmission is considered for upstream TWDM-PON. Once using this transmission mode, some parameters have to be taken into consideration, as the packets size and their separation length in order to understand which frame fits the best, considering the system performance. In the actual study, it was supposed to visualize transients in each packet, however it was experimentally proved that once the lifetime of the carriers is less than the burst time, it was not possible to identify any of them. It was also verified that increasing the guard time will decrease the performance of the system. UR - https://ria.ua.pt/handle/10773/14570 Y1 - 2014 PB - Universidade de Aveiro