Este trabalho apresenta uma análise sobre a Arquitetutra PCE (Path Computation Element), como alternativa para o roteamento inter-domínios em redes heterogêneas, ou seja, em ambientes onde ocorre a integração do domínio óptico, das redes de transporte ópticas, com o domínio eletrônico, das redes convencionais, tratando inclusive com questões de qualidade de transmissão do sinal óptico, ou QoT (Quality of Transmission). Os objetivos e o funcionamento do mecanismo do PCE em redes integradas são detalhados. São apresentados também os conceitos e definições das redes ópticas, bem como seus principais elementos, a proposta de interconexão das mesmas com as redes convencionais provendo uma rede única e integrada, o papel dos planos de controle, e os principais mecanismos de roteamento. A análise do PCE foi direcionada para o comportamento do principal protocolo utilizado na sua arquitetura, denominado PCE Protocol. De forma mais específica, foi realizada a análise de suas mensagens de requisição de rotas e das mensagens de respostas a estas requisições, e sua interferência no tráfego dos enlaces das redes onde são introduzidas. A partir daí é apresentada uma proposta de extensão de parâmetros para o protocolo PCE Protocol, com o objetivo de atender necessidades específicas de QoT, baseado em estudos das particularidades das características das transmissões ópticas. A partir destas definições, foram realizadas análises adicionais a fim de verificar o impacto com relação ao tráfego gerado em enlaces de diferentes larguras de banda, com a definição original do protocolo, e com a extensão proposta, permitindo uma comparação entre ambos. / This thesis presents an analysis of Path Computation Element Architecture (PCE) as an alternative to be implemented in heterogeneous network inter-domain routing, in other words, environments where optical domains are integrated with the electronic domains of conventional networks. Also, the ability of PCE to manage issues regarding the transmission quality of the optical signal, the QoT (Quality of Transmission), is presented. Objectives and functionality of the PCE mechanism in integrated networks is detailed. Concepts and definitions regarding optical networks and their main elements, the aim of interconnection with conventional networks to build a unified and integrated network, the role of control planes, and the main routing mechanisms are also presented. The analysis of the PCE was focused on the behavior of its main protocol, called PCE Protocol. Specifically, the analysis was performed on both its request and response messages, in addition, the interference in the traffic of the links where these messages had been introduced was also analyzed. Afterwards, an extension of the PCE Protocol parameters was proposed based on the study of optical transmission characteristics, with the aim of meeting specific QoT requirements. From these definitions, further analyses of the impact on the traffic in links with different bandwidths were performed, using the PCE protocol, with and without the proposed extension, allowing a comparison between them.

Uma das premissas fundamentais em redes fotônicas baseadas na tecnologia de multiplexação por comprimento de onda (WDM) é o controle dos caminhos ópticos. Um caminho óptico, ou caminho de luz, é uma conexão puramente óptica estabelecida entre dois nós da rede, que pode atravessar diversos nós intermediários. Para dar suporte eficiente à demanda exigida de uma rede WDM sobre a qual executam aplicações de alto desempenho, os caminhos ópticos devem ser estabelecidos e encerrados dinamicamente, de tal forma que as rotas e os comprimentos de onda escolhidos minimizem a probabilidade de ocorrência de bloqueios de conexão por falta de recursos disponíveis. O elemento central de uma rede WDM é o sistema que controla os comutadores ópticos, determina as rotas, aloca os comprimentos de onda, e estabelece, mantém e encerra as conexões ópticas entre os nós da rede. O objetivo deste trabalho é apresentar estratégias de roteamento e de atribuição de comprimentos de onda para redes fotônicas usando heurísticas baseadas em inteligência de enxame. Os resultados demonstram que as estratégias desenvolvidas apresentam baixa probabilidade de bloqueio de requisições e buscam, sempre que possível, os caminhos de menor custo. As estratégias desenvolvidas levam em conta o dinamismo da rede, sendo as decisões de caminho baseadas em dados atualizados em cada nó da rede em tempo real, sem a necessidade do conhecimento global do estado da rede em cada nó. / A fundamental aspect of photonic networks based on wavelength division multiplexing (WDM) technology is the control of optical paths. An optical path, or lightpath, is a purely optical connection established between two network nodes, which can pass through several intermediate nodes. To efficiently support the demands required from a WDM network that runs high performance applications, the optical paths should be established and finished dynamically, in such a way that the selected routes and wavelengths tend to minimize the blocking probability due to the lack of available resources. The central element of a WDM network is the system that controls the optical switches, determines routes, allocates the wavelengths, and establishes, maintains and finishes connections between optical network nodes. The purpose is to present strategies for routing and wavelength assignment over photonic networks using heuristics based on swarm intelligence. Results show that the developed strategies exhibit low blocking probability of requests and find the paths of minimum cost whenever possible. The strategies developed take into account the dynamics of the network, and decisions are based on updated data at each network node in real time without the need of a complete knowledge of network state at each node.

This thesis puts focus on the technological challenges for Wavelength Division Multiplexed Passive Optical Network (WDM-PON) implementation, and presents novel semiconductor optical devices for deployment at the optical network unit (ONU). The first-ever reported L-band Reflective semiconductor optical amplifier (RSOA) is presented based on InP-base material. A theoretical model is developed to estimate the optical gain and the saturation power of this device compared to a conventional SOA. Experiments on this device design show long-range telecom wavelength operation, with polarization-independent gain of greater than 20 dB, and low saturation output power of 0 dBm suitable for PON applications. Next, the effect of the amplified spontaneous emission noise of RSOA devices on WDM-PON system is investigated. It is shown through theoretical modeling and simulations that the RSOA noise combined with receiver noise statistics increase probability of error, and induce considerable power penalties to the WDM-PON system. By improving the coupling efficiencies, and by distributing more current flow to the input of these devices, steps can be taken to improve device noise characteristics. Further, in spectrally-spliced WDM-PONs deploying RSOAs, the effect of AWG filter shape on system performance is investigated. Simulation modeling and experiments show that deployment of Flat-band AWGs is critical for reducing the probability of error caused by AWG spectral shape filtering. Flat-band athermal AWGs in comparison to Gaussin-shape counterparts satisfy the maximum acceptable error probability requirements, and reduce the power penalty associated with filtering effect. In addition, detuning between two AWG center wavelengths impose further power penalties to the WDM-PON system. In the last section of this thesis, motivated by RSOA device system limitations, a novel injection-locked Fabry-Perot (IL-FP) device is presented which consists of a gain section monolithically integrated with a phase section. The gain section provides locking of one FP mode to a seed source wavelength, while the phase modulator allows for adjusting the wavelength of the internal modes by tuning bias current to maintain mode-locking. This device counters any mode drifts caused by temperature variations, and allows for cooler-less operation over a wide range of currents. The devices and the performance metrics subsequently allow for a hybrid integration platform on a silicon substrate and integrate many functionalities like reflective modulator with thin film dielectric filter and receiver on a single chip for deployment at the user-end of future-proof low cost WDM-PONs.

Fault management is a challenging task in all-optical wavelength division multiplexing (WDM) networks. However, fast fault localization for shared risk link groups (SRLGs) with multiple links is essential for building a fully survival and functional transparent all-optical mesh network. Monitoring trail (m-trail) technology is an effective approach to achieve the goal, whereby a set of m-trails are derived for unambiguous fault localization (UFL). However, an m-trail traverses through a link by utilizing a dedicated wavelength channel (WL), causing a significant amount of resource consumption. In addition, existing m-trail methods incur long and variable alarm dissemination delay. We introduce a novel framework of real-time fault localization in all-optical WDM mesh networks, called the monitoring-burst (m-burst), which aims at initiating a balanced trade-off between consumed monitoring resources and fault localization latency. The m-burst framework has a single monitoring node (MN) and requires one WL in each unidirectional link if the link is traversed by any m-trail. The MN launches short duration optical bursts periodically along each m-trail to probe the links of the m-trail. Bursts along different m-trails are kept non-overlapping through each unidirectional link by scheduling burst launching times from the MN and multiplexing multiple bursts, if any, traversing the link. Thus, the MN can unambiguously localize the failed links by identifying the lost bursts without incurring any alarm dissemination delay. We have proposed several novel m-trail allocation, burst launching time scheduling, and node switch fabric configuration schemes. Numerical results show that the schemes, when deployed in the m-burst framework, are able to localize single-link and multi-link SRLG faults unambiguously, with reasonable fault localization latency, by using at most one WL in each unidirectional link. To reduce the fault localization latency further, we also introduce a novel methodology called nested m-trails. At first, mesh networks are decomposed into cycles and trails. Each cycle (trail) is realized as an independent virtual ring (linear) network using a separate pair of WLs (one WL in each direction) in each undirected link traversed by the cycle (trail). Then, sets of m-trails, i.e., nested m-trails, derived in each virtual network are deployed independently in the m-burst framework for ring (linear) networks. As a result, the fault localization latency is reduced significantly. Moreover, the application of nested m-trails in adaptive probing also reduces the number of sequential probes significantly. Therefore, practical deployment of adaptive probing is now possible. However, the WL consumption of the nested m-trail technique is not limited by one WL per unidirectional link. Thus, further investigation is needed to reduce the WL consumption of the technique.

This thesis puts focus on the technological challenges for Wavelength Division Multiplexed Passive Optical Network (WDM-PON) implementation, and presents novel semiconductor optical devices for deployment at the optical network unit (ONU). The first-ever reported L-band Reflective semiconductor optical amplifier (RSOA) is presented based on InP-base material. A theoretical model is developed to estimate the optical gain and the saturation power of this device compared to a conventional SOA. Experiments on this device design show long-range telecom wavelength operation, with polarization-independent gain of greater than 20 dB, and low saturation output power of 0 dBm suitable for PON applications. Next, the effect of the amplified spontaneous emission noise of RSOA devices on WDM-PON system is investigated. It is shown through theoretical modeling and simulations that the RSOA noise combined with receiver noise statistics increase probability of error, and induce considerable power penalties to the WDM-PON system. By improving the coupling efficiencies, and by distributing more current flow to the input of these devices, steps can be taken to improve device noise characteristics. Further, in spectrally-spliced WDM-PONs deploying RSOAs, the effect of AWG filter shape on system performance is investigated. Simulation modeling and experiments show that deployment of Flat-band AWGs is critical for reducing the probability of error caused by AWG spectral shape filtering. Flat-band athermal AWGs in comparison to Gaussin-shape counterparts satisfy the maximum acceptable error probability requirements, and reduce the power penalty associated with filtering effect. In addition, detuning between two AWG center wavelengths impose further power penalties to the WDM-PON system. In the last section of this thesis, motivated by RSOA device system limitations, a novel injection-locked Fabry-Perot (IL-FP) device is presented which consists of a gain section monolithically integrated with a phase section. The gain section provides locking of one FP mode to a seed source wavelength, while the phase modulator allows for adjusting the wavelength of the internal modes by tuning bias current to maintain mode-locking. This device counters any mode drifts caused by temperature variations, and allows for cooler-less operation over a wide range of currents. The devices and the performance metrics subsequently allow for a hybrid integration platform on a silicon substrate and integrate many functionalities like reflective modulator with thin film dielectric filter and receiver on a single chip for deployment at the user-end of future-proof low cost WDM-PONs.

Σκοπός αυτής της διπλωματικής εργασίας είναι η ανάλυση και αξιολόγηση της απόδοσης των οπτικών δικτύων μεταγωγής ριπών με την μέθοδο της προσέγγισης μειωμένου φορτίου. Η μέθοδος αυτή είναι αρκετά ευέλικτη για να προσαρμοστεί σε ένα οπτικό δίκτυο μεταγωγής οπτικών ριπών το οποίο υιοθετεί 4 διαφορετικά είδη πολιτικών και οι οποίες είναι η JET, η κατάτμηση ριπής, όπως επίσης και δυο πολιτικές προτεραιοτήτων. Η ακρίβεια της μεθόδου επιβεβαιώνεται από την προσομοίωση και τέλος οι διάφορες πολιτικές συγκρίνονται μεταξύ τους. Το πρώτο κεφάλαιο αυτής της εργασίας έχει αφιερωθεί στο να εισάγει τον αναγνώστη σε βασικές έννοιες σχετικά με τις οπτικές ίνες και τα οπτικά δίκτυα από την πρώτη έως την τρίτη γενιά. Αναφέρεται στα τρία είδη οπτικής μεταγωγής που υπάρχουν και ακολούθως γίνεται μια σύγκριση αυτών των τεχνολογιών οπτικής μεταγωγής. Στο δεύτερο κεφάλαιο αυτής της εργασίας, γίνεται μια πιο ενδελεχής ανάλυση των δικτύων δρομολόγησης με οπτικές ριπές και παρουσιάζονται όλοι οι απαραίτητοι μηχανισμοί για την ύπαρξη και άρτια λειτουργία αυτών των δικτύων. Αναλυτικότερα, παρέχονται πληροφορίες για την επίλυση θεμάτων ανταγωνισμού και την σηματοδοσία την συναρμολόγηση των ριπών. Στη συνέχεια, στο τρίτο κεφάλαιο γίνεται η παρουσίαση της μεθόδου προσέγγισης μειωμένου φορτίου για κλασσικά τηλεφωνικά δίκτυα. Σκοπός είναι να αναλυθεί εκτενώς ο τρόπος με τον οποίο μπορεί να χρησιμοποιηθεί η μέθοδος αυτή σε οπτικά δίκτυα μεταγωγής οπτικών ριπών και τι απαραίτητες αλλαγές πρέπει να γίνουν για την κάθε πολιτική που εφαρμόζεται ξεχωριστά. Παράλληλα, για κάθε πολιτική παρέχονται πίνακες με αποτελέσματα όπως και γραφήματα για να γίνουν πιο κατανοητά τα αποτελέσματα της μεθόδου. Έπειτα, παρουσιάζονται τα συμπεράσματα της εργασίας όπως και διάφορες προτάσεις για το πώς μπορεί να αξιοποιηθεί η κάθε πολιτική με τον καλύτερο τρόπο. Τέλος στο τέταρτο κεφάλαιο παρατίθενται οι κώδικες των αναλυτικών μοντέλων για κάθε μια από τις πολιτικές που χρησιμοποιήθηκε σε γλώσσα προγραμματισμού C, έτσι ώστε αν θέλει ο αναγνώστης να έρθει σε επαφή και με το προγραμματιστικό περιβάλλον του δικτύου που υλοποιήθηκε. / The subject of this Diploma Thesis is the analysis and evaluation of the performance of Optical Burst Switching Networks by using the method of reduced load approximation. This method is flexible enough to be adjusted to the type of networks mentioned above which normally adopt four different kinds of policies, like the JET policy, the burst segmentation policy and also two policies of priorities. Thus, the main goal of this project is the confirmation of the accuracy of this method by performing the necessary simulations in order to compare these different policies. The first chapter of the project is dedicated to the task of making the reader familiar with the basic elements of the science of optical fibers and optical networks in general, all the way from the very first generation of them to the third one that is widely used at present. During this chapter, the three types of optical switching networks are described. These types are further examined next, during the second chapter, where a thorough analysis of network routing with optical bursts is performed. In addition, all the necessary mechanisms for the existence and well-functioning of these networks are presented. Specifically, a wide range of information related to the solving of competition issues and signaling for the assembly of the bursts, is given. Then, during the third chapter, the method of reduced load approximation is presented for the classic telephony networks. This method can be used in optical burst switching networks, and different variations of it are used, according to each separate policy that is followed. Result tables accompanied by the appropriate charts are supplied in order to clarify the results of this method. Next, the results of this Diploma Thesis are aggregated and accompanied by proposals about reclaiming each policy the best way. In the last chapter, the programming codes of each analytical model of these four optical networks policies, expressed in the programming language C, are supplied, giving the reader the chance to get in touch with the programming environment that was used in order to derive the optical network that was simulated.

Orientador: Prof. Dr. Gustavo Sousa Pavani / Dissertação (mestrado) - Universidade Federal do ABC, Programa de Pós-Graduação em Ciência da Computação, 2014. / Por conta da grande demanda por computação de alto desempenho, maiores taxas de envio de dados e disponibilidade de recursos de um ambiente de computação em grade, este trabalho apresenta um estudo realizado sobre o uso de redes ópticas que operam com Multiplexação por Divisão de Comprimento de Onda (Wavelength Division Multiplexing (WDM)) no suporte a grades computacionais, cujo ambiente é chamado de lambda grid. Assim, é proposto neste trabalho uma nova arquitetura de lambda grid capaz de ampliar a disponibilidade de seus recursos computacionais e de rede, por meio da inserc~ao de um suporte a reserva futura destes recursos, ampliando a disponibilidade da grade a tempos futuros e não apenas a tempos correntes de operação. Para o gerenciamento desse ambiente, foi proposto o uso de um algoritmo baseado em Otimização por Colônia de Formigas (Ant Colony Optimization (ACO)) integrado ao plano de controle Generalized Multi-Protocol Label Switching (GMPLS), o qual é capaz de gerenciar dinamicamente e de forma conjunta os recursos de processamento e de rede, alem de prover os mecanismos de agendamento e de reserva futura desses recursos. A arquitetura proposta, juntamente com os demais algoritmos desenvolvidos, foi testada e validada por meio de simulações. Os resultados das simulações permitiram a avaliação do desempenho da nova arquitetura e da necessidade de se incorporar o agendamento conjunto com reserva futuras na execução da ação da probabilidade de bloqueio. / Due to the increasing demand of high-performance computing, higher data transport rates and the availability of the resources on a grid computing environment, this work presents a study of the use of Wavelength Division Multiplexing (WDM) optical networks on the support of grid computing, whose environment is known as lambda grid. It is proposed in this work a new architecture of lambda grid which is capable of increasing the availability of its network and computing resources by the use of an advance reservation approach, which makes the grid able to other its resources in future times of operation and not only in the current time. For the management of this environment, it was proposed the use of an Ant Colony Optimization (ACO) based algorithm integrated to the Generalized Multi-Protocol Label Switching (GMPLS) control plane, which is capable to dynamically manage the server and network resources, in addition to provide the co-scheduling and advance reservation of those resources. The proposed architecture and algorithms were tested and evaluated by means of simulations. The obtained results allowed the evaluation of the performance of the new architecture and the need to incorporate the co-scheduling of server and optical network resources in their advance reservation in order to minimize the blocking probability.

abstract: Access Networks provide the backbone to the Internet connecting the end-users to the core network thus forming the most important segment for connectivity. Access Networks have multiple physical layer medium ranging from fiber cables, to DSL links and Wireless nodes, creating practically-used hybrid access networks. We explore the hybrid access network at the Medium ACcess (MAC) Layer which receives packets segregated as data and control packets, thus providing the needed decoupling of data and control plane. We utilize the Software Defined Networking (SDN) principle of centralized processing with segregated data and control plane to further extend the usability of our algorithms. This dissertation introduces novel techniques in Dynamic Bandwidth allocation, control message scheduling policy, flow control techniques and Grouping techniques to provide improved performance in Hybrid Passive Optical Networks (PON) such as PON-xDSL, FiWi etc. Finally, we study the different types of software defined algorithms in access networks and describe the various open challenges and research directions. / Dissertation/Thesis / Doctoral Dissertation Electrical Engineering 2015

Submitted by Fernando Souza (fernandoafsou@gmail.com) on 2017-08-21T11:30:58Z No. of bitstreams: 1 arquivototal.pdf: 1795088 bytes, checksum: fb7c00f2437bd31139648401e67e1ad9 (MD5) / Made available in DSpace on 2017-08-21T11:30:58Z (GMT). No. of bitstreams: 1 arquivototal.pdf: 1795088 bytes, checksum: fb7c00f2437bd31139648401e67e1ad9 (MD5) Previous issue date: 2014-07-31 / Coordenação de Aperfeiçoamento de Pessoal de Nível Superior - CAPES / In this work the problem of routing and route allocation (RSA) for elastic optical networks has been studied. This is similar to the problem of routing and wavelength allocation problem, the latter being characteristic routed optical networks in wavelengths. The problem RSA aims to allocate the least amount of resources to elastic optical network so it can cater to the maximum number of demands set out in customer traffic matrix. Besides the study of the problem, in this paper, two metaheuristics were implemented, which were the Iterated Local Search (ILS) and the Greedy Randomized Adaptive Search Procedure (GRASP), which use SSA and LP-CA mathematical models in the local search phase. To implement metaheuristics and mathematical models, we used IBM ILOG CPLEX IDE. The results are presented in graphs, which showed that metaheuristics able to be efficient and competitive, with good quality solutions in an acceptable and even lower than the results presented by accurate modeling computational time. Finally, this paper presents a discussion of the advantages and limitations of this study, and presents directions for future work in this field of study. / Neste trabalho foi estudado o problema de roteamento e alocação de rota (RSA) em redes ópticas elásticas. Este é um problema similar ao problema de roteamento e alocação de comprimentos de onda, este último sendo característico em redes ópticas roteadas em comprimentos de onda. O problema RSA tem como objetivo atribuir a menor quantidade de recursos de uma rede óptica elástica de tal forma que consiga atender ao máximo número de demandas definidas na matriz de tráfego cliente. Além do estudo sobre o problema, neste trabalho foram implementadas duas metaheurísticas, sendo elas o Iterated Local Search (ILS) e o Greedy Randomized Adaptative Search Procedure (GRASP), as quais utilizam os modelos matemáticos SSA e LP-CA na fase de busca local. Para implementar os modelos matemáticos e metaheurísticas, foi utilizada a IDE IBM iLOG CPLEX. Os resultados obtidos são apresentados através de gráficos, os quais demonstraram que as metaheurísticas conseguiram ser eficientes e competitivas, apresentando soluções de boa qualidade em um tempo computacional aceitável e até menor que os resultados apresentados pelos modelos exatos. Por fim, este texto apresenta uma discussão sobre os diferenciais e limitações deste trabalho, e apresenta direcionamentos para trabalhos futuros neste campo de estudo.

O objetivo deste trabalho é utilizar o método do tráfego aleatório equivalente para analisar e dimensionar redes ópticas sem comutação para arquitetura de nível único. Uma modelagem analítica, baseada nesse método, que possibilita otimizar os parâmetros da rede considerando padrões mínimos de desempenho é descrita. A rede em estudo utiliza apenas um nó, onde se realizam as funções de roteamento passivo. Em anexo ao nó existem seqüências de conversores de comprimentos de onda gerenciados por um controlador de rede. A estrutura da rede consiste de um certo número de redes ópticas passivas (PONs) conectadas às portas do nó. Essas PONs são distribuídas pela área geográfica onde estão os clientes a serem atendidos. O acesso ao meio constituído pelas PONs é disciplinado por meio da utilização de um esquema WDMA / TDMA administrado pelo controlador de rede. Os comprimentos de onda pertinentes a cada PON conectada a uma porta de entrada do nó são alocados de tal forma que cada um deles se direciona a uma diferente porta de saída e, por conseguinte, a uma diferente PON de destino. Uma conexão entre dois terminais é estabelecida pela seleção de um intervalo de tempo (slot de tempo) pertinente ao comprimento de onda que conecta suas respectivas PONs através do nó. Caso esse comprimento de onda já esteja com sua capacidade plenamente utilizada, conexões adicionais podem ser estabelecidas utilizando-se um comprimento de onda direcionado a uma seqüência de conversores de comprimentos de onda anexa ao nó. Após a conversão, é utilizado outro comprimento de onda que é direcionado pelo nó à PON de destino. A modelagem analítica investigada é utilizada para obter as relações entre os parâmetros da rede, possibilitando a escolha dos seus melhores valores. / The purpose of this work is to use the equivalent random traffic method for analysis and dimensioning of switchless optical networks for single-layer architecture. This work reports an analytical model that allows the network dimensioning according to some relevant design parameters and the required system performance. The analyzed network is based on a single node providing passive routing functions. Wavelength converter arrays, controlled by a network controller, are added in the node. The network structure consists of a set of passive optical networks (PONs) connected to the node ports. Those PONs are distributed along the geographical area where the customers to be assisted are. The access to the medium constituted by the PONs is disciplined through the use of a WDMA / TDMA scheme administered by the network controller. The pertinent wavelengths to each PON connected to an input port of the node are allocated in such a way that each one is addressed to a different output port and, consequently, to a different destination PON. A connection between two terminals is set up by the selection of a time slot pertinent to the wavelength that connects its respective PONs through the node. If the capacity of the wavelength is fully used, additional connections can be established by the use of wavelength directed to a wavelength converter array enclosed to the node. After the conversion, another wavelength addressed by the node to the destination PON is used. The connections achieved between the network parameters are detailed to allow the choice of the best network parameters.