Rapid changes in population distribution across Canada and the introduction of new telecommunication services to the consumer market have resulted in a number of significant challenges for existing network infrastructure. Fast growing populations in metropolitan regions require high density access networks to meet the growing need for bandwidth that results. Furthermore, new services such as high definition TV, online gaming and real-time video teleconferencing are becoming increasingly popular among consumers. These services require higher bandwidth to be available to end users. Changes in the Canadian economy will soon lead to a transition in Canadian industry from manufacturing to services and exploration of natural resources. This will create opportunities for new industrial development and growth in northern regions. Expanding industrialization towards northern Canada will require deployment of reliable telecommunication infrastructure. The combination of open source software, Linux operating system and Personal Computer (PC) based hardware platform is proposed to become the foundation for low cost and flexible technology that will provide transition towards all-optical infrastructures. An innovative prototype of a low-cost optical gigabit Ethernet switch is presented and its benchmark results are discussed. Scalability of the switch and its future applications in optical networks are studied. A prototype of a software based data encapsulation system was designed and implemented in a PC based platform, and its performance was evaluated using real data that was captured in commercial LAN. Semiconductor optical amplifiers (SOA) are studied as a building block in next generation switching devices for all-optical access networks. A prototype of an SOA-based low-cost optical switching device with implemented FPGA based controlling mechanism is presented and its characteristics are discussed. SOA is also studied as an energy efficient optical amplifier that can be deployed in end user facilities. The presented results provide proof of concept of a low cost flexible platform that can be used to design and build network devices to facilitate the transition of existing telecommunication networks towards next generation optical access infrastructure.

In recent years, the ever-increasing environmental friendliness concern has made energy efficiency in telecom networks as an important theme in their operations. Meanwhile, mission-critical (MC) services and systems (such as healthcare, police, and firefighting) have been acquiring special attention from telecom designers and operators. The currently deployed MC network technologies are indigent in terms of bandwidth capacity, and thus they are not able to support the emerging MC multimedia applications. Therefore in this thesis, we first explore the possibility of provisioning the MC services over the integration of fiber-wireless (FiWi) technologies, which has been considered as a promising candidate for the deployment of high-speed and mobile broadband access networks. We then investigate the energy efficiency problem in the FiWi integration, which consists of WiMAX in the wireless plane, and of Ethernet Passive Optical Network (EPON) - the most popular variant of the next-generation PON (NG-PON) technology, in the optical plane. In WiMAX, the energy saving protocol has been extensively investigated and standardized. Conversely, it has been recently studied in NG-PON, which currently consumes the least power among all the high-speed access networks. However, NG-PON has notably matured in the past few years and is envisioned to massively evolve in the near future. This trend will increase the power requirements of NG-PON and make it no longer coveted. Therefore we address the energy efficiency problem in NG-PON. For each of our contributions, we conduct extensive simulations to demonstrate the effectiveness and advantages of the proposed solutions.

Rapid changes in population distribution across Canada and the introduction of new telecommunication services to the consumer market have resulted in a number of significant challenges for existing network infrastructure. Fast growing populations in metropolitan regions require high density access networks to meet the growing need for bandwidth that results. Furthermore, new services such as high definition TV, online gaming and real-time video teleconferencing are becoming increasingly popular among consumers. These services require higher bandwidth to be available to end users. Changes in the Canadian economy will soon lead to a transition in Canadian industry from manufacturing to services and exploration of natural resources. This will create opportunities for new industrial development and growth in northern regions. Expanding industrialization towards northern Canada will require deployment of reliable telecommunication infrastructure. The combination of open source software, Linux operating system and Personal Computer (PC) based hardware platform is proposed to become the foundation for low cost and flexible technology that will provide transition towards all-optical infrastructures. An innovative prototype of a low-cost optical gigabit Ethernet switch is presented and its benchmark results are discussed. Scalability of the switch and its future applications in optical networks are studied. A prototype of a software based data encapsulation system was designed and implemented in a PC based platform, and its performance was evaluated using real data that was captured in commercial LAN. Semiconductor optical amplifiers (SOA) are studied as a building block in next generation switching devices for all-optical access networks. A prototype of an SOA-based low-cost optical switching device with implemented FPGA based controlling mechanism is presented and its characteristics are discussed. SOA is also studied as an energy efficient optical amplifier that can be deployed in end user facilities. The presented results provide proof of concept of a low cost flexible platform that can be used to design and build network devices to facilitate the transition of existing telecommunication networks towards next generation optical access infrastructure.

This thesis focuses on system for persistence of GPON communication. Theoretical part deals with problems of GPON and NG-PON optical networks, NoSQL database systems and MongoDB database. Practical part contains design of a database schema for MongoDB database and source code in programming languages Python and C# for working with this database. The thesis is finalized by performance testing, proving that the database design and source code implementation is capable of handling real world traffic.

This masters thesis is focusing on the issue of passive optic access networks. It is dealing with the limiting factors for broadband connections, especially attenuation of fibre optics and dispersion. It is also dealing with most commonly used standards of passive optic networks. It clarifies the issue of wavelength-division multiplexing, its advantages and disadvantages, used standards, various types of components for the WDM-networks, their characteristics, application etc. In the practical part of the work you can find the measuring and diagnostics of the passive optic network EPON. It contains measuring by the direct method OLTS, measuring by the Pon Power Meter and optic reflectometer OTDR. At the end I am dealing with the measuring of splitter in the rate of 1:16.

Cílem této práce je analyzovat současný stav přístupových optických sítí,a hlouběji prozkoumat PON metodiku a návrh sítě. Navrhnul jsem síť založenou na konceptu FTTH pro 90 rodinných domů ovládané z centrální stanice vzdálené 20 km. Zabýval jsem se problémy při návrhu FTTH, konkrétně pasivní optickou sítí. Dále jsem se zabýval ATM-PON a Ethernet PON (EPON), které jsou založeny na společné síťové architektuře, PON, ale používají jiné přenosové technologie na podporu integrovaných služeb a jiné protokoly. Po zadání (optimálního) síťového řešení, které se hodí pro nabídku Triple-Play služeb. Dále budu popisovat, jak fungují síťová propojení a také budou popsány metody měření a testování navržené sítě "po aktivaci a nastavení sítě" a "před aktivací síťových služeb ". Při měření jsem také používal OTDR a prakticky ověřil charakteristiku linky. V práci je také naznačeno monitorování optické sítě.

This thesis deals with the most recent passive optical network standard NG-PON2, describes the sublayer model which includes error correction coding during propagation in optical fibres. Assesses the impact of split ratios using the simulation environment created from defined model and compares various scenarios. Discusses possible sources of errors of simulation model in compare to real deployment.

Wavelength division multiplexing passive optical network (WDM-PON) can meet growing bandwidth demand in access network by providing high bandwidth to the end users. Failure in the access network is becoming critical as a large volume of traffic might be affected. Therefore, an effective supervision mechanism to detect and localize the fault is required to shorten the service interruption time. Meanwhile, open access provides a certain freedom for end users to choose the service and hence boosts competition among service/network providers. On the other hand, to offer open access in WDM-PON could result in a substantial change on architectural design, e.g., multiple feeder fibers (FFs) instead of a single one may be required to connect different service/network providers. Consequently, the traditional supervision mechanisms don’t work properly in open WDM-PON. To fill in this gap, several fault supervision mechanisms to support open access in WDMPON are proposed in this thesis. They can be applied to both disjoint and co-located FF layout where the choice of providers is done through wavelength selection. The feasibility of such solutions has been validated by evaluating transmission performance. We have carried out simulations in VPItransmissionMaker for different deployment scenarios. The results have confirmed that no significant degradation of the transmission performance is introduced by the proposed monitoring schemes compared to the benchmark, where no any fault supervision method is implemented.

La croissance rapide des applications interactives et ses demandes en termes de bande passante et de (QoS), a motivé le déploiement de la technologie optique aux réseaux métropolitains. Beaucoup de dispositifs optiques intelligents sont créés pour gérer efficacement la capacité énorme du réseau. L'explosion de la bande passante est principalement attribuable à la croissance du trafic de données dont la nature devient de plus en plus complexe. Actuellement, la plupart des trafics de données sont transportés dans les réseaux qui sont dominées par la technologie de commutation de circuit optique. Néanmoins, cette technique a quelques inconvénients tels que les propriétés inflexibles. Technologie de commutation de paquets optiques (OPS), qui offre des gains significatifs en termes de capacité de passage à l’échelle et d'efficacité de gestion des ressources du réseau, peut surmonter ces limitations. Cela a motivé l'orientation de la commutation de circuit optique à la commutation de paquet optique dans l’infrastructure future du réseau. En réalité, un réseau en anneau à commutation de paquet optique ( OPSR), qui combinent la flexibilité et la mise à l’échelle de la technologie de commutation de paquet avec les avantages de la topologie en anneau tels que la restitution rapide du service en cas de panne et un bon gain de multiplexage statistique du trafic, promettraient une bonne solution pour les réseaux MAN du future. Une nouvelle architecture du réseau OPSR qui se base sur l’infrastructure tout optique a proposé par des équipes de recherche chez Alcatel-Lucent. La nouvelle architecture offre des fonctions intelligentes, avec un coût moins élevé en optimisant le temps de traitement. Elle est envisagée de remplacer les architectures optoélectroniques existantes tels que Resilient Packet Ring. L'élément fondamental du réseau est Packet Optical Add/Drop Multiplexer (POADM) qui est implémenté à l'intérieur des nœuds d’accès, permettant aux nœuds d'exploiter la transparence optique. Cette thèse se concentre donc sur l'évaluation des performances de la nouvelle génération des réseaux optiques métropolitains. Dans cette thèse, nous avons analysé des performances (en termes de délais d’accès et de la distribution du taux de l’occupation au tampon local des nœuds d'accès) d'un bus optique synchrone qui soutient des paquets à taille fixe. Nous avons modélisé chaque nœud d'accès par une chaîne de Markov à temps discret (EDTMC). La solution de l'EDTMC nous permet de calculer la probabilité approximative où les nœuds d'accès dans le bus peuvent "voir" slots libres dans la ligne de transit. En utilisant une technique d'analyse récursive, nous déduisons une formule mesurant le temps moyen d'attente des paquets du client venant de la couche supérieure, ainsi que la distribution du taux de l’occupation au tampon local des nœuds d'accès pour les deux cas: avec et sans garantie QoS). Pour caractériser des performances d'un réseau OPSR asynchrone, nous avons évalué l'impact de la création de containeur optique à taille fixe sur les performances du réseau. L’analyse des performances de ce système nous a permis d'identifier une combinaison raisonnable de certains paramètres (la durée de timeslot, les valeurs d’expiration des temporisateurs, le profil du trafic client, la charge du réseau) qui peuvent améliorer l'utilisation de la bande passante du réseau. Le point le plus important que nous avons mentionné dans la thèse est la comparaison entre deux approches architecturales: le modèle de Variable Length - Optical Packet Format (VL-OPF) versus le modèle de Fixed Length - Optical Packet Format (FL-OPF). En plus de Cos-Upgrade Mécanisme (CUM) qui est proposé d'améliorer le taux de remplissage du conteneur optique, nous avons proposé un nouveau mécanisme appelé Dynamic Upgrade Cos-Mécanisme (DCUM), où les valeurs des temporisateurs sont modifiés dynamiquement en fonction de l'état des tampons locaux aux nœuds d’accès et le trafic circulant dans le réseau / The rapid growth in client application demands, in terms of bandwidth and (QoS), has motivated the deployment of the optical technology at Metro Access and Metro Core Networks. More diverse and more intelligent optic devices are required for efficiently management of huge capacity in the network. The explosion in demand for network bandwidth is mainly due to the growth in data traffic whose nature is also becoming more and more complex. In general, most of service-based traffics are transported in networks which are now being dominated by the optical switching technology. Nevertheless, such technology has some drawbacks such as inflexible and non-scalable properties. (OPS), which offers significant benefits in terms of both network efficiency and control scalability, may overcome these limitations. This has motivated the orientation from optical circuit switching to optical packet switching in the future network infrastructure. In reality, optical packet switching ring (OPSR) networks, which combine the packet switching technology with the well-known advantages of ring topology such as fast service restoration in cases of failure and high gain of statistical traffic multiplexing over the ring, appear to be the technology of choice for the next generation of Metro Area Networks. A new OPSR architecture which is based on the all-optical infrastructure that offers intelligent features with lower cost while maximizing processing time and is now considering being replaced existing opto-electronic architectures such as Resilient Packet Ring, have been presented in the scope of the dissertation. The key element of such network is Packet Optical Add/Drop Multiplexer (POADM) which is implemented inside a ring node, allowing the node to exploit the optical transparency. This dissertation hence focuses on the performance evaluation of the new generation of OPSR network. In this dissertation, we have investigated the performance analysis (in terms of packet delay and queue-length distribution of access nodes) of an optical synchronous bus-based metropolitan network supporting fixed-size packets. We have modeled each access node by an embedded discrete time Markov chain (EDTMC). The solution of the EDTMC allows us to compute the approximate probability that access nodes on the bus “see” free slots in the transit line. Using a recursive analysis technique, we approximately outline the mean waiting time of client packets coming from the upper layer as well as an approximate queue-length distribution of local buffers at access nodes in two cases: with and without QoS guarantees. To characterize the performance of an asynchronous OPSR network, we have evaluated the impact of Optical fixed-Size Packet Creation on the network performance without and with different quality of services (QoS) requirements. Performance analysis of such systems has allowed us to identify a reasonable combination of some parameters (timeslot duration, timer expiration values, profile of client traffic and network load) which may be able to improve the bandwidth utilization of the network for a given traffic matrix. The most important point that we have mentioned in the dissertation is the comparison of two architectural approaches: Variable Length – Optical Packet Format (VL-OPF) model supporting empty packets versus Fixed Length – Optical Packet Format (FL-OPF) model. In addition to CoS-Upgrade Mechanism (CUM) which is proposed to improve the filling ratio of the optical container, we have proposed a novel mechanism named Dynamic CoS-Upgrade Mechanism (DCUM) where timers are dynamically changed according to the state of the local buffer of network nodes and the traffic circulating in the network, in order to create containers with a high filling ratio while limiting the time needed for their creation.

Un défi majeur dans les réseaux d’aujourd’hui est de combler l’écart entre la haute vitesse de la transmission optique et la vitesse plus limitée du traitement électronique des données. Une option est de commuter les données directement dansle domaine optique. Dans cette thèse, nous proposons plusieurs solutions permettant la commutation dans le domaine optique à une granularité plus fine que la longueur d’onde, technique que nous appelons commutation sous-longueurd’onde. Pour montrer la pertinence des solutions proposées, nous analysons leur performance en termes de capacité de trafic, de débit et de délai. La performance est évaluée à la fois par des simulations et en utilisant des modèles de filesd’attente appropriés. Nous considérons d’abord le cas des réseaux métropolitains (Metropolitan Area Networks, MAN) et nous étudions la performance d’un anneau optique avec multiplexage en longueur d’onde (Wavelength Division Multiplexing, WDM) dans lequel la communication entre les noeuds du réseau se fait par insertion/extraction de données dans des créneaux temporels. Nous présentons un protocole entièrement distribué conçu pour assurer l’équité dansce réseau. Nous proposons également un mécanisme d’assemblage de paquets capable d’assurer des délais faibles ainsi que des taux de remplissage élevés. Nous proposons ensuite des solutions de commutation sous-longueur d’onde qui peuvent être appliquées dans le cas plus général des réseaux asynchrones. D’abord, nous proposons de résoudre le problème des collisions de la commutation optique par rafale (Optical Burst Switching, OBS) par la mise en oeuvre d’un mécanisme de réservation. Afin de maximiser l’utilisation des ressources, nous proposons d’adapter la taille de la rafale optique à la charge du réseau.Ensuite, nous proposons une solution alternative pour construire un réseau coeur tout-optique. A cette architecture, nous associons un protocole d’accès ainsi qu’un algorithme d’allocation dynamique de bande passante et nous analysons les performances de la solution proposée. Par le biais d’une étude de cas, nousmontrons que notre solution est capable de réduire considérablement la consommation énergétique par rapport aux architectures actuelles basées sur des routeurs IP. Enfin, nous proposons un nouveau dispositif optique capable derésoudre la contention directement dans le domaine optique. Nous montrons que ce dispositif simple peut être utilisé pour construire des réseaux optiques dynamiques à courte portée tels que les réseaux d’accès ou les centres de traitement de données. / A key challenge in today’s networks is to bridge the gap between high-speed optical transmission and limited electronic processing. This can be achieved by enabling payload to be switched directly in the optical domain.A simple solutionto provide optical switching is by allocating one wavelength channel to each source-destination pair, a technique called Optical Circuit Switching (OCS). Due to lack of sharing, OCS suffers from limited scalability. To overcome this issue,the capacity of each wavelength channel must be dynamically shared among different source-destination pairs. This requires data to be switched at subwavelength granularity by means of subwavelength switching. In this thesis, wepropose several solutions which enable subwavelength switching in optical networks. To show the relevance of the proposed solutions, we analyse their performance in terms of traffic capacity, flow throughput and packet delay. Performance is evaluated both through simulations and by means of appropriate queueing models. We first consider the case of Metropolitan Area Networks (MAN) and we study the performance of synchronous time-slotted Wavelength DivisionMultiplexing (WDM) ring in which network nodes communicate by inserting and extracting data from time-slots. We present a fully distributedMedia Access Control (MAC) protocol designed to ensure fairness. We also propose a burst assembly mechanism able to ensure low assembly delays and high fill rates of the optical time-slots. We then propose subwavelength switching solutions which can be applied in the more general case of asynchronous wide area networks. We first propose to solve the contention problems of conventional Optical Burst Switching (OBS) and the low utilization issue of wavelength-routed OBS byimplementing a two-way reservation OBS scheme in which the size of the opticalburst increases proportionally with the network load so as to maximize resourceutilization. Next, we propose a solution for building an all-optical wide area network based on multipoint-to-multipoint lightpath sharing. We also design an associated MAC protocol and a dynamic bandwidth allocation algorithm and analyse the performance of the proposed solution. By means of a case study, we show that the proposed solution has the potential to considerably reduce power consumption with respect to current router-based architectures. Finally, we propose a novel optical device able to solve contention directly in the optical domain withoutrequiring any optical buffering, electronic signalling or header processing. We show that thissimple device can be used as a building block for dynamic and power efficient short-range optical networks such as access networks or data centers.