In recent years, multi-threaded processors have become more and more popular in industry in order to increase the system aggregated performance and per-application performance, overcoming the limitations imposed by the limited instruction-level parallelism, and by power and thermal constraints. Multi-threaded processors are widely used in servers, desktop computers, lap-tops, and mobile devices. However, multi-threaded processors introduce complexities when accounting CPU (computation) capacity (CPU accounting), since the CPU capacity accounted to an application not only depends upon the time that the application is scheduled onto a CPU, but also on the amount of hardware resources it receives during that period. And given that in a multi-threaded processor hardware resources are dynamically shared between applications, the CPU capacity accounted to an application in a multi-threaded processor depends upon the workload in which it executes. This is inconvenient because the CPU accounting of the same application with the same input data set may be accounted significantly different depending upon the workload in which it executes. Deploying systems with accurate CPU accounting mechanisms is necessary to increase fairness among running applications. Moreover, it will allow users to be fairly charged on a shared data center, facilitating server consolidation in future systems. This Thesis analyses the concepts of CPU capacity and CPU accounting for multi-threaded processors. In this study, we demonstrate that current CPU accounting mechanisms are not as accurate as they should be in multi-threaded processors. For this reason, we present two novel CPU accounting mechanisms that improve the accuracy in measuring the CPU capacity for multi-threaded processors with low hardware overhead. We focus our attention on several current multi-threaded processors, including chip multiprocessors and simultaneous multithreading processors. Finally, we analyse the impact of shared resources in multi-threaded processors in operating system CPU scheduler and we propose several schedulers that improve the knowledge of shared hardware resources at the software level.

No description available.

With the recent switch in the design of general purpose processors from frequency scaling of a single processor core towards increasing the number of processor cores, parallel programming became important not only for scientific programming but also for general purpose programming. This also stressed the importance of programmability of existing parallel programming models which were primarily designed for performance. It was soon recognized that new programming models are needed that will make parallel programming possible not only to experts, but to a general programming community. Transactional Memory (TM) is an example which follows this premise. It improves dramatically over any previous synchronization mechanism in terms of programmability and composability, at the price of possibly reduced performance. The main source of performance degradation in Transactional Memory is the overhead of transactional execution. Our work on parallelizing Quake game engine is a clear example of this problem. We show that Software Transactional Memory is superior in terms of programmability compared to lock based programming, but that performance is hindered due to extreme amount of overhead introduced by transactional execution. In the meantime, a significant research effort has been invested in overcoming this problem. Our approach is aimed towards improving the performance of transactional code by reducing transactional data conflicts. The idea is based on the organization of the code in which highly conflicting data is promoted to dataflow tokens that coordinate the execution of transactions. The main contribution of this thesis is Atomic Dataflow model (ADF), a new task-based parallel programming model for C/C++ that integrates dataflow abstractions into the shared memory programming model. The ADF model provides language constructs that allow a programmer to delineate a program into a set of tasks and to explicitly define data dependencies for each task. The task dependency information is conveyed to the ADF runtime system that constructs a dataflow task graph that governs the execution of a program. Additionally, the ADF model allows tasks to share data. The key idea is that computation is triggered by dataflow between tasks but that, within a task, execution occurs by making atomic updates to common mutable state. To that end, the ADF model employs transactional memory, which guarantees atomicity of shared memory updates. The second contribution of this thesis is DaSH - the first comprehensive benchmark suite for hybrid dataflow and shared memory programming models. DaSH features 11 benchmarks, each representing one of the Berkeley dwarfs that capture patterns of communication and computation common to a wide range of emerging applications. DaSH includes sequential and shared-memory implementations based on OpenMP and TBB to facilitate easy comparison between hybrid dataflow implementations and traditional shared memory implementations. We use DaSH not only to evaluate the ADF model, but to also compare it with other two hybrid dataflow models in order to identify the advantages and shortcomings of such models, and motivate further research on their characteristics. Finally, we study applicability of hybrid dataflow models for parallelization of the game engine. We show that hybrid dataflow models decrease the complexity of the parallel game engine implementation by eliminating or restructuring the explicit synchronization that is necessary in shared memory implementations. The corresponding implementations also exhibit good scalability and better speedup than the shared memory parallel implementations, especially in the case of a highly congested game world that contains a large number of game objects. Ultimately, on an eight core machine we were able to obtain 4.72x speedup compared to the sequential baseline, and to improve 49% over the lock-based parallel implementation based on work-sharing. / Con el reciente cambio en el diseño de los procesadores de propósito general pasando del aumento de frecuencia al incremento del número de núcleos, la programación paralela se ha convertido en importante no solo para la comunidad científica sino también para la programación en general. Este hecho ha enfatizado la importancia de la programabilidad de los modelos actuales de programación paralela, cuyo objetivo era el rendimiento. Pronto se observó la necesidad de nuevos modelos de programación, para hacer factible la programación paralela a toda la comunidad. Transactional Memory (TM) es un ejemplo de dicho objetivo. Supone una gran mejora sobre cualquier método anterior de sincronización en términos de programabilidad, con una posible reducción del rendimiento como coste. La razón principal de dicha degradación es el sobrecoste de la ejecución transaccional. Nuestro trabajo en la paralelización del motor del juego Quake es un claro ejemplo de este problema. Demostramos que Software Transactional Memory es superior en términos de programabilidad a los modelos de programación basados en locks, pero que el rendimiento es entorpecido por el sobrecoste introducido por TM. Mientras tanto, se ha invertido un importante esfuerzo de investigación para superar dicho problema. Nuestra solución se dirige hacia la mejora del rendimiento del código transaccional reduciendo los conflictos con la información contenida en las transacciones. La idea se basa en la organización del código en el cual la información conflictiva es promocionada a señales del flujo de datos que coordinan la ejecución de las transacciones. La contribución principal de esta tesis es Atomic Dataflow Model (ADF), un nuevo modelo de programación para C/C++ basado en tareas que integra abstracciones de flujo de datos en el modelo de programación de la memoria compartida. El modelo ADF provee construcciones del lenguaje que permiten al programador la definición del programa como un conjunto de tareas, además de la definición explícita de las dependencias de datos para cada tarea. La información de dependencia de la tarea se transmite al runtime de ADF, que construye un grafo de tareas que es el que controla la ejecución de un programa. Adicionalmente, el modelo ADF permite que las tareas compartan información. La idea principal es que la computación es activada por el flujo de datos entre tareas, pero que dentro de una tarea la ejecución ocurre haciendo actualizaciones atómicas a un estado común mutable. Para conseguir este fin, el modelo ADF utiliza TM, que garantiza la atomicidad en las modificaciones de la memoria compartida. La segunda contribución es DaSH, el primer conjunto de benchmarks para los modelos de programación de flujo de datos híbridos y los de memoria compartida. DaSH contiene 11 benchmarks, cada uno representativo de uno de los Berkeley dwarfs que captura patrones de comunicaciones y procesamiento comunes en un amplio rango de aplicaciones emergentes. DaSH incluye implementaciones secuenciales y de memoria compartida basadas en OpenMP y TBB que facilitan la comparación entre los modelos híbridos de flujo de datos e implementaciones de memoria compartida. Nosotros usamos DaSH no solo para evaluar ADF, sino también para compararlo con otros dos modelos híbridos para identificar sus ventajas. Finalmente, estudiamos la aplicabilidad de dichos modelos híbridos para la paralelización del motor del juego. Mostramos que disminuyen la complejidad de la implementación paralela, eliminando o reestructurando la sincronización explícita que es necesaria en las implementaciones de memoria compartida. También se observa una buena escalabilidad y una aceleración mejor, especialmente en el caso de un ambiente de juego muy cargado. En última instancia, sobre una máquina con ocho núcleos se ha obtenido una aceleración del 4.72x comparado con el código secuencial, y una mejora del 49% sobre la implementación paralela basada en locks.

In recent years, the advent of new Future Internet (FI) applications is creating ever-demanding requirements. These requirements are pushing network carriers for high transport capacity, energy efficiency, as well as high-availability services with low latency. A widespread practice to provide FI services is the adoption of a multi-layer network model consisting in the use of IP/MPLS and optical technologies such as Wavelength Division Multiplexing (WDM). Indeed, optical transport technologies are the foundation supporting the current telecommunication network backbones, because of the high transmission bandwidth achieved in fiber optical networks. Traditional optical networks consist of a fixed 50 GHz grid, resulting in a low Optical Spectrum (OS) utilization, specifically with transmission rates above 100 Gbps. Recently, optical networks have been undergoing significant changes with the purpose of providing a flexible grid that can fully exploit the potential of optical networks. This has led to a new network paradigm termed as Elastic Optical Network (EON). In recent years, the advent of new Future Internet (FI) applications is creating ever-demanding requirements. A widespread practice to provide FI services is the adoption of a multi-layer network model consisting in the use of IP/MPLS and optical technologies such as Wavelength Division Multiplexing (WDM). Traditional optical networks consist of a fixed 50 GHz grid, resulting in a low Optical Spectrum (OS) utilization. Recently, optical networks have been undergoing significant changes with the purpose of providing a flexible grid that can fully exploit the potential of optical networks. This has led to a new network paradigm termed as Elastic Optical Network (EON). Recently, a new protection scheme referred to as Network Coding Protection (NCP) has emerged as an innovative solution to proactively enable protection in an agile and efficient manner by means of throughput improvement techniques such as Network Coding. It is an intuitive reasoning that the throughput advantages of NCP might be magnified by means of the flexible-grid provided by EONs. The goal of this thesis is three-fold. The first, is to study the advantages of NCP schemes in planning scenarios. For this purpose, this thesis focuses on the performance of NCP assuming both a fixed as well as a flexible spectrum grid. However, conversely to planning scenarios, in dynamic scenarios the accuracy of Network State Information (NSI) is crucial since inaccurate NSI might substantially affect the performance of an NCP scheme. The second contribution of this thesis is to study the performance of protection schemes in dynamic scenarios considering inaccurate NSI. For this purpose, this thesis explores prediction techniques in order to mitigate the negative effects of inaccurate NSI. On the other hand, Internet users are continuously demanding new requirements that cannot be supported by the current host-oriented communication model.This communication model is not suitable for future Internet architectures such as the so-called Internet of Things (IoT). Fortunately, there is a new trend in network research referred to as ID/Locator Split Architectures (ILSAs) which is a non-disruptive technique to mitigate the issues related to host-oriented communications. Moreover, a new routing architecture referred to as Path Computation Element (PCE) has emerged with the aim of overcoming the well-known issues of the current routing schemes. Undoubtedly, routing and protection schemes need to be enhanced to fully exploit the advantages provided by new network architectures.In light of this, the third goal of this thesis introduces a novel PCE-like architecture termed as Context-Aware PCE. In a context-aware PCE scenario, the driver of a path computation is not a host/location, as in conventional PCE architectures, rather it is an interest for a service defined within a context. / En los últimos años la llegada de nuevas aplicaciones del llamado Internet del Futuro (FI) está creando requerimientos sumamente exigentes. Estos requerimientos están empujando a los proveedores de redes a incrementar sus capacidades de transporte, eficiencia energética, y sus prestaciones de servicios de alta disponibilidad con baja latencia. Es una práctica sumamente extendida para proveer servicios (FI) la adopción de un modelo multi-capa el cual consiste en el uso de tecnologías IP/MPLS así como también ópticas como por ejemplo Wavelength Division Multiplexing (WDM). De hecho, las tecnologías de transporte son el sustento del backbone de las redes de telecomunicaciones actuales debido al gran ancho de banda que proveen las redes de fibra óptica. Las redes ópticas tradicionales consisten en el uso de un espectro fijo de 50 GHz. Esto resulta en una baja utilización del espectro Óptico, específicamente con tasas de transmisiones superiores a 100 Gbps. Recientemente, las redes ópticas están experimentado cambios significativos con el propósito de proveer un espectro flexible que pueda explotar el potencial de las redes ópticas. Esto ha llevado a un nuevo paradigma denominado Redes Ópticas Elásticas (EON). Por otro lado, un nuevo esquema de protección llamado Network Coding Protection (NCP) ha emergido como una solución innovadora para habilitar de manera proactiva protección eficiente y ágil usando técnicas de mejora de throughput como es Network Coding (NC). Es un razonamiento lógico pensar que las ventajas relacionadas con throughput de NCP pueden ser magnificadas mediante el espectro flexible proveído por las redes EONs. El objetivo de esta tesis es triple. El primero es estudiar las ventajas de esquemas NCP en un escenario de planificación. Para este propósito, esta tesis se enfoca en el rendimiento de NCP asumiendo un espectro fijo y un espectro flexible. Sin embargo, contrario a escenarios de planificación, en escenarios dinámicos la precisión relacionada de la Información de Estado de Red (NSI) es crucial, ya que la imprecisión de NSI puede afectar sustancialmente el rendimiento de un esquema NCP. La segunda contribución de esta tesis es el estudio del rendimiento de esquemas de protección en escenarios dinámicos considerando NSI no precisa. Para este propósito, esta tesis explora técnicas predictivas con el propósito de mitigar los efectos negativos de NSI impreciso. Por otro lado, los usuarios de Internet están demandando continuamente nuevos requerimientos los cuales no pueden ser soportados por el modelo de comunicación orientado a hosts. Este modelo de comunicaciones no es factible para arquitecturas FI como es el Internet de las cosas (IoT). Afortunadamente, existe un nueva línea investigativa llamada ID/Locator Split Architectures (ILSAs) la cual es una técnica no disruptiva para mitigar los problemas relacionadas con el modelo de comunicación orientado a hosts. Además, un nuevo esquema de enrutamiento llamado as Path Computation Element (PCE) ha emergido con el propósito de superar los problemas bien conocidos de los esquemas de enrutamiento tradicionales. Indudablemente, los esquemas de enrutamiento y protección deben ser mejorados para que estos puedan explotar las ventajas introducidas por las nuevas arquitecturas de redes. A luz de esto, el tercer objetivo de esta tesis es introducir una nueva arquitectura PCE denominada Context-Aware PCE. En un escenario context-aware PCE, el objetivo de una acción de computación de camino no es un host o localidad, como es el caso en lo esquemas PCE tradicionales. Más bien, es un interés por un servicio definido dentro de una información de contexto.

Hoy en día, los problemas de equilibrado de líneas de montaje se encuentran comúnmente en la mayoría de sistemas industriales y de manufactura. Básicamente, estos problemas consisten en asignar un conjunto de tareas a una secuencia ordenada de estaciones de trabajo, de manera que se respeten las restricciones de precedencia y se optimice una medida de eficiencia dada (como, por ejemplo, el número de estaciones de trabajo o el tiempo ciclo). Dada la complejidad de los problemas de equilibrado de líneas, en los trabajos de investigación tradicionalmente se consideraban numerosas simplificaciones en las que, por ejemplo, una sola línea serial procesaba un único modelo de un solo producto. Además, los problemas estaban principalmente restringidos por las relaciones de precedencia y el tiempo ciclo. Sin embargo, la disponibilidad de recursos computacionales de hoy en día, así como la necesidad de las empresas a adaptarse a los rápidos cambios en los procesos de producción, han motivado tanto a investigadores como a gerentes a tratar problemas más realistas. Algunos ejemplos incluyen problemas que procesan modelos mixtos, estaciones de trabajo y líneas en paralelo, consideran múltiples objetivos y restricciones adicionales, como la capacidad de proceso de las estaciones de trabajo y la ubicación de los recursos en la línea de montaje.Esta tesis doctoral trata un nuevo problema de equilibrado de líneas, que ha sido titulado ASALBP: the Alternative Subgraphs Assembly Line Balancing Problem, en el que se consideran variantes alternativas para diferentes partes de un proceso de montaje o de manufactura. Cada alternativa puede ser representada por un subgrafo de precedencias, que determina las tareas requeridas para procesar un producto particular, las restricciones de precedencia y los tiempos de proceso. Para resolver eficientemente el ASALBP, se deben resolver dos problemas simultáneamente: (1) el problema de decisión para seleccionar un subgrafo de montaje para cada parte que admite alternativas y (2) el problema de equilibrado para asignar las tareas a las estaciones de trabajo. El análisis del estado del arte revela que este problema no ha sido estudiado previamente en la literatura, lo que ha conducido a la caracterización y a la definición de un nuevo problema. Por otra parte, dado que no es posible representar las variantes de montaje en un diagrama de precedencias estándar, se propone el S-grafo como una herramienta de diagramación, para representar en un único grafo todas las alternativas de montaje.Habitualmente, los problemas de equilibrado de líneas que consideran alternativas de montaje se resuelven en dos etapas. En la etapa inicial, el diseñador de sistema selecciona una de las variantes posibles utilizando cierto criterio de decisión como por ejemplo tiempo total de proceso. Una vez que se han seleccionado las alternativas de montaje, y se dispone de un diagrama de precedencias (es decir, el problema de planificación ha sido resuelto), la línea de montaje es equilibrada en una segunda etapa. Sin embargo, utilizando dicho procedimiento de dos etapas no se puede garantizar que una solución óptima del problema global se pueda obtener, porque las decisiones tomadas por el diseñador de sistema restringen el problema y causan perdida de información; es decir, cuando se selecciona una alternativa priori los efectos de las posibilidades restantes quedan sin explorar. Por ejemplo, si el diseñador de sistema utiliza tiempo total de proceso como criterio de decisión, la alternativa con el tiempo total de proceso más grande será descartada a pesar de que pueda ser la que proporcione la mejor solución del problema (es decir, requiere el mínimo número de estaciones de trabajo o el mínimo tiempo ciclo). Por lo tanto, pareciera razonable considerar que para solucionar eficientemente un ALBP que implica alternativas de proceso, todas las alternativas de montaje deben ser tomadas en cuenta en el proceso de equilibrado. Para este propósito, en esta tesis el problema de selección de una variante de montaje y el problema de equilibrado de la línea se consideran conjuntamente en lugar de independientemente.Para resolver el Problema de Equilibrado de Líneas con Alternativas de Montaje (ASALBP) se usan varios enfoques. El problema se formaliza y se resuelve de manera óptima a través de dos modelos de programación matemática. Un enfoque aproximativo es usado para resolver problemas de tamaño industrial. Además, se proponen procedimientos de optimización local con el objetivo de mejorar la calidad de las soluciones obtenidas por los métodos heurísticos desarrollados en este trabajo. / Nowadays assembly line balancing problems are commonly found in most industrial and manufacturing systems. Basically, these problems seek to assign a set of assembly tasks to an ordered sequence of workstations in such a way that precedence constraints are maintained and a given efficiency measure (e.g. the number of workstations or the cycle time) is optimized.Because of the computational complexity of balancing problems, research works traditionally considered numerous simplifying assumptions in which, for example, a single model of a unique product were processed in a single line; moreover, problems were mainly restricted by precedence and cycle time constrains. Nevertheless, the current availability of computing resources and the enterprises need to adapt to rapid changes in production and manufacturing processes have encouraged researchers and decision-makers to address more realistic problems. Some examples include problems that involve mixed models, parallel workstations and parallel lines, multiple objectives and also further restrictions such as workstation processing capacity and resource allocation constraints. This doctoral thesis addresses a novel assembly line balancing problem, entitled here ASALBP: the Alternative Subgraphs Assembly Line Balancing Problem, which considers alternative variants for different parts of an assembly or manufacturing process. Each variant can be represented by a precedence subgraph that establishes the tasks required to process a particular product, their precedence requirements and their processing times. Therefore, to efficiently solve the Alternative Subgraphs Assembly Line Balancing Problem two subproblems need to be solved simultaneously: (1) the decision problem that selects one assembly variant for each part that admit alternatives and (2) the balancing problem that assigns the tasks to the workstations. The analysis of the state-of-the-art carried out revealed that the Alternative Subgraphs Assembly Line Balancing Problem has not been addressed before in literature studies, which leaded to the characterization and definition of this new problem. Moreover, due to the impossibility of representing assembly variants in a standard precedence graph, the S-Graph is proposed here as a diagramming tool to represent all available assembly alternatives in a unique diagram. Habitually, problems involving assembly alternatives are solved by using a two-stage based approach. In the initial stage, the system designer selects one of the possible variants according to criteria such as total processing time. Once the assembly alternatives have been selected, and a precedence graph is available (i.e. the assembly planning problem has been already solved), the line is then balanced in the second stage. However, by following this two-stage procedure it cannot be guaranteed that an optimal solution of the global problem can be obtained, because the decisions taken by the system designer restrict the problem and cause information loss; i.e., a priori selection of an alternative leaves the effects of the other possibilities unexplored. For instance, if the system designer uses total processing time as decision criterion, the alternative with largest total processing time will be discarded notwithstanding it may provide the best solution of the problem (i.e., it requires the minimum number of workstations or minimum cycle time). Therefore, it seems reasonable to consider that to solve efficiently an ALBP that involves processing alternatives all possibilities must be considered within the balancing process. For this purpose, in this thesis both the variant selection problem and the balancing problem are jointly considered instead of independently.Different approaches are used here to address the Alternative Subgraphs Assembly Line Balancing Problem (ASALBP). The problem is formalize and optimally solved by means of two mathematical programming models. An approximate approach is used to address industrial-scale problems. Furthermore, local optimization procedures are proposed aiming at improving the quality of the solutions provided by all heuristic methods developed here.

The Internet infrastructure, comprising both network and cloud facilities, has reached huge capacities but its development has not been compensated at the same rate as for its energy consumption. The energy consumption and the concomitant green house gases (GHG) emissions of the Internet are becoming major issues in the information and communication society. In such a context, there is a lack of a comprehensive energy-oriented paradigm for the Internet infrastructure that takes into account the absorbed energy, the emitted GHGs and the availability of renewable energy sources. This Thesis is focused on these very issues and tries to address the lack of such a paradigm in the Internet infrastructure by proposing energy models for energy-efficient architectures, energy-aware algorithms and protocols conceived to optimize the use of energy and minimize GHGs emissions, while preserving the traditional criteria such as network and datacenters load balancing to serve as many demands as possible and maximizing the system availability. In order to achieve the energy-oriented paradigm for the Internet infrastructure, specific problems were addressed step-wise, and then tied together in a comprehensive energy-oriented framework. Towards this goal, the power consumption of current and future energy-aware architectures was modeled through energy models that characterize the energy consumption of network equipment under different traffic loads, and power management strategies were assessed to allow network infrastructures to achieve advanced functionalities with limited energy budget. Integrated routing and wavelength assignment (RWA) schemes have been proposed (ILP formulations, heuristics and meta-heuristics, game theory, minimum affinity, minimum cut) in order to take advantage of different scenarios (complete or partial knowledge of network status, global control or individual selfishness of network elements, different requisites of computational and space complexity). Energy-aware RWA algorithms require an underlying routing protocol distributing up-to-date information about the energy consumption and GHG emissions of the network elements. Link state advertisement (LSA) messages of the OSPF-TE protocol have been extended to carry energy-related information. New TLVs have been added directly to the TE extensions of OSPF and flooded over the network. The connections re-optimization problem has been formulated as an iterative refinement process of multiple local search steps structured as a GRASP meta-heuristic, which re-reroutes connections to maintain the network traffic load balanced and free resources to serve incoming connections. To support the research tasks, a WDM-routed networks simulator, SimulNet, has been developed for the design and the evaluation of RWA and optimization algorithms. Energy-Farm, an energy manager for the modern and future grid/cloud data center infrastructures, was developed to reduce datacenters ecological footprint. Through the service-demand matching algorithm and the job aggregation capabilities, it allows turning off idle servers, while respecting both the demand requirements and the logical and physical dependencies. The risks related to energy-oriented attacks were pointed out for the first time and the potential impacts of network-based DoS attacks under the energy consumption perspective were evaluated. Finally, a holistic vision on the energy-oriented Internet is provided in which energy-efficient architectures are powered by a smart grid power distribution system employing renewable energy sources and are controlled by an intelligent energy-aware control plane, able to operate the Internet to minimize its ecological footprint. The research works leading to this Thesis delineate an energy-oriented paradigm for a sustainable high-performance Internet infrastructure that optimize the Internet ecological footprint while not disrupting the performance, towards sustainable society growth and prosperity. / La infraestructura de Internet, tanto de red como de centros de proceso de datos, ya alcanza un enorme volumen, pero este incremento no ha sido compensado con la misma rapidez en aspectos relacionados con el gasto energético. El consumo de energía y las emisiones de gases efecto invernadero (GEI) de Internet han pasado a ser un problema relevante en la sociedad de la información y las comunicaciones. En este entorno, falta un paradigma de largo alcance orientado a la energía, que considere el consumo de energía, las emisiones de GEI y la disponibilidad de recursos renovables. Esta Tesis está enfocada hacia estos problemas e intenta compensar la falta de ese paradigma en la infraestructura de Internet, proponiendo modelos energéticos para nuevas arquitecturas, así como algoritmos y protocolos conscientes de la energía para optimizar su uso y minimizar las emisiones de GEI, preservando los objetivos de calidad tradicionales de redes y centros de procesamiento de datos, así como asegurar la posibilidad de servir el mayor número de demandas posible y maximizar la disponibilidad del sistema. Para alcanzar una infraestructura de Internet orientada a la energía, se han solucionado problemas específicos y ligados a una estructura común de largo alcance. Hacia este objetivo, se ha modelado a través de modelos energéticos el consumo de los dispositivos bajo diferentes cargas, y se han valorado diversas estrategias de gestión de la energía para que las infraestructuras de red alcancen funcionalidades avanzadas con un presupuesto de energía limitado. Se han propuesto esquemas integrados de encaminamiento y asignación de longitud de onda (RWA) (formulaciones ILP, heurísticas y meta-heurísticas, teoría de los juegos, mínima afinidad, mínimo corte) para diferentes escenarios (conocimiento completo o parcial del estado de la red, control global o individual de los elementos de red, diferentes requisitos de computación y de espacio). Los algoritmos de RWA conscientes de la energía requieren un protocolo de encaminamiento que distribuya informaciones actualizadas sobre el consumo energético y las emisiones de GEI de los elementos de red. Se han desarrollado extensiones de los mensajes de aviso sobre el estado de la red (LSA) del protocolo OSPF-TE para transportar informaciones sobre la energía, añadiendo nuevos TLVs directamente a las extensiones TE de OSPF. El problema de la optimización de las conexiones se ha formulado como un proceso de refinado iterativo de pasos múltiples estructurado como una meta-heurística GRASP, que permite encaminar las conexiones para mantener el tráfico de la red balanceado y liberar recursos para servir posteriores conexiones. Para respaldar las tareas de investigación, se ha desarrollado SimulNet, un simulador de redes de encaminamiento de longitudes de ondas (WDM), para el diseño, optimización y evaluación de algoritmos de RWA. Se ha desarrollado EnergyFarm, un gestor de energía para los modernos centros de procesamiento de datos que, a través de un algoritmo de armonización entre demanda y servicio ofrecido y funcionalidades de agregación de las tareas, permite apagar los servidores no usados respetando los requisitos de las peticiones y las dependencias físicas y lógicas de los dispositivos. Se han evidenciado por primera vez los riesgos relacionados con los ataques orientados a la energía y se ha valorado su potencial impacto. Finalmente, se ha proporcionada una visión holística de Internet orientada a la energía, en la que arquitecturas eficientes energéticamente están alimentadas por una smart grid con fuentes renovables y controlada por un plano de control inteligente y consciente de la energía, capaz de operar en Internet para minimizar su huella ecológica. Los trabajos de investigación de esta Tesis conducen hacia un paradigma orientado a la energía para una infraestructura sostenible de Internet de alto rendimiento que optimice su huella ecológica sin afectar el rendimiento.

The microprocessor technology has experienced an enormous growth during the last decades. Rapid downscale of the CMOS technology has led to higher operating frequencies and performance densities, facing the fundamental issue of power dissipation. Chip Multiprocessors (CMPs) have become the latest paradigm to improve the power-performance efficiency of computing systems by exploiting the parallelism inherent in applications. Industrial and prototype implementations have already demonstrated the benefits achieved by CMPs with hundreds of cores.CMP architects are challenged to take many complex design decisions. Only a few of them are:- What should be the ratio between the core and cache areas on a chip?- Which core architectures to select?- How many cache levels should the memory subsystem have?- Which interconnect topologies provide efficient on-chip communication?These and many other aspects create a complex multidimensional space for architectural exploration. Design Automation tools become essential to make the architectural exploration feasible under the hard time-to-market constraints. The exploration methods have to be efficient and scalable to handle future generation on-chip architectures with hundreds or thousands of cores.Furthermore, once a CMP has been fabricated, the need for efficient deployment of the many-core processor arises. Intelligent techniques for task mapping and scheduling onto CMPs are necessary to guarantee the full usage of the benefits brought by the many-core technology. These techniques have to consider the peculiarities of the modern architectures, such as availability of enhanced power saving techniques and presence of complex memory hierarchies.This thesis has several objectives. The first objective is to elaborate the methods for efficient analytical modeling and architectural design space exploration of CMPs. The efficiency is achieved by using analytical models instead of simulation, and replacing the exhaustive exploration with an intelligent search strategy. Additionally, these methods incorporate high-level models for physical planning. The related contributions are described in Chapters 3, 4 and 5 of the document.The second objective of this work is to propose a scalable task mapping algorithm onto general-purpose CMPs with power management techniques, for efficient deployment of many-core systems. This contribution is explained in Chapter 6 of this document.Finally, the third objective of this thesis is to address the issues of the on-chip interconnect design and exploration, by developing a model for simultaneous topology customization and deadlock-free routing in Networks-on-Chip. The developed methodology can be applied to various classes of the on-chip systems, ranging from general-purpose chip multiprocessors to application-specific solutions. Chapter 7 describes the proposed model.The presented methods have been thoroughly tested experimentally and the results are described in this dissertation. At the end of the document several possible directions for the future research are proposed.

The hiring problem is a simple model for on-line decision-making under uncertainty, recently introduced in the literature. Despite some related work dates back to 2000, the name and the first extensive studies were written in 2007 and 2008. The problem has been introduced explicitly first by Broder et al. in 2008 as a natural extension to the well-known secretary problem. Soon afterwards, Archibald and Martínez in 2009 introduced a discrete (combinatorial) model of the hiring problem, where the candidates seen so far could be ranked from best to worst without the need to know their absolute quality scores. This thesis introduces an extensive study for the hiring problem under the formulation given by Archibald and Martínez, explores the connections with other on-line selection processes in the literature, and develops one interesting application of our results to the field of data streaming algorithms. In the hiring problem we are interested in the design and analysis of hiring strategies. We study in detail two hiring strategies, namely hiring above the median and hiring above the m-th best. Hiring above the median hires the first interviewed candidate then any coming candidate is hired if and only if his relative rank is better than the median rank of the already hired staff, and others are discarded. Hiring above the m-th best hires the first m candidates in the sequence, then any coming candidate is hired if and only if his relative rank is larger than the m-th best among all hired candidates, and others are discarded. For both strategies, we were able to obtain exact and asymptotic distributional results for various quantities of interest (which we call hiring parameters). Our fundamental parameter is the number of hired candidates, together with other parameters like waiting time, index of last hired candidate and distance between the last two hirings give us a clear picture of the hiring rate or the dynamics of the hiring process for the particular strategy under study. There is another group of parameters like score of last hired candidate, score of best discarded candidate and number of replacements that give us an indicator of the quality of the hired staff. For the strategy hiring above the median, we study more quantities like number of hired candidates conditioned on the first one and probability that the candidate with score q is getting hired. We study the selection rule 1/2-percentile rule introduced by Krieger et al., in 2007, and the seating plan (1/2,1) of the Chinese restaurant process (CRP) introduced by Pitman, which are very similar to hiring above the median. The connections between hiring above the m-th best and the notion of m-records, and also the seating plan (0,m) of the CRP are investigated here. We report preliminary results for the number of hired candidates for a generalization of hiring above the median; called hiring above the alpha-quantile (of the hired staff). The explicit results for the number of hired candidates enable us to design an estimator, called RECORDINALITY, for the number of distinct elements in a large sequence of data which may contain repetitions; this problem is known in the literature as cardinality estimation problem. We show that another hiring parameter, the score of best discarded candidate, can also be used to design a new cardinality estimator, which we call DISCARDINALITY. Most of the results presented here have been published or submitted for publication. The thesis leaves some open questions, as well as many promising ideas for future work. One interesting question is how to compare two different strategies; that requires a suitable definition of the notion of optimality, which is still missing in the context of the hiring problem. We are also interested in investigating other variants of the problem like probabilistic hiring strategies, that is when the hiring criteria is not deterministic, unlike all the studied strategies.

This thesis reports research on mapping, path planning, and autonomous exploration. These are classical problems in robotics, typically studied independently, and here we link such problems by framing them within a common SLAM approach, adopting Pose SLAM as the basic state estimation machinery. The main contribution of this thesis is an approach that allows a mobile robot to plan a path using the map it builds with Pose SLAM and to select the appropriate actions to autonomously construct this map. Pose SLAM is the variant of SLAM where only the robot trajectory is estimated and where landmarks are only used to produce relative constraints between robot poses. In Pose SLAM, observations come in the form of relative-motion measurements between robot poses. With regards to extending the original Pose SLAM formulation, this thesis studies the computation of such measurements when they are obtained with stereo cameras and develops the appropriate noise propagation models for such case. Furthermore, the initial formulation of Pose SLAM assumes poses in SE(2) and in this thesis we extend this formulation to SE(3), parameterizing rotations either with Euler angles and quaternions. We also introduce a loop closure test that exploits the information from the filter using an independent measure of information content between poses. In the application domain, we present a technique to process the 3D volumetric maps obtained with this SLAM methodology, but with laser range scanning as the sensor modality, to derive traversability maps. Aside from these extensions to Pose SLAM, the core contribution of the thesis is an approach for path planning that exploits the modeled uncertainties in Pose SLAM to search for the path in the pose graph with the lowest accumulated robot pose uncertainty, i.e., the path that allows the robot to navigate to a given goal with the least probability of becoming lost. An added advantage of the proposed path planning approach is that since Pose SLAM is agnostic with respect to the sensor modalities used, it can be used in different environments and with different robots, and since the original pose graph may come from a previous mapping session, the paths stored in the map already satisfy constraints not easy modeled in the robot controller, such as the existence of restricted regions, or the right of way along paths. The proposed path planning methodology has been extensively tested both in simulation and with a real outdoor robot. Our path planning approach is adequate for scenarios where a robot is initially guided during map construction, but autonomous during execution. For other scenarios in which more autonomy is required, the robot should be able to explore the environment without any supervision. The second core contribution of this thesis is an autonomous exploration method that complements the aforementioned path planning strategy. The method selects the appropriate actions to drive the robot so as to maximize coverage and at the same time minimize localization and map uncertainties. An occupancy grid is maintained for the sole purpose of guaranteeing coverage. A significant advantage of the method is that since the grid is only computed to hypothesize entropy reduction of candidate map posteriors, it can be computed at a very coarse resolution since it is not used to maintain neither the robot localization estimate, nor the structure of the environment. Our technique evaluates two types of actions: exploratory actions and place revisiting actions. Action decisions are made based on entropy reduction estimates. By maintaining a Pose SLAM estimate at run time, the technique allows to replan trajectories online should significant change in the Pose SLAM estimate be detected. The proposed exploration strategy was tested in a common publicly available dataset comparing favorably against frontier based exploration

Propositional Proof Complexity is the area of Computational Complexity that studies the length of proofs in propositional logic. One of its main questions is to determine which particular propositional formulas have short proofs in a given propositional proof system. In this thesis we present several results related to this question, all on proof systems that are extensions of the well-known resolution proof system. The first result of this thesis is that TQBF, the problem of determining if a fully-quantified propositional CNF-formula is true, is PSPACE-complete even when restricted to instances of bounded tree-width, i.e. a parameter of structures that measures their similarity to a tree. Instances of bounded tree-width of many NP-complete problems are tractable, e.g. SAT, the boolean satisfiability problem. We show that this does not scale up to TQBF. We also consider Q-resolution, a quantifier-aware version of resolution. On the negative side, our first result implies that, unless NP = PSPACE, the class of fully-quantified CNF-formulas of bounded tree-width does not have short proofs in any proof system (and in particular in Q-resolution). On the positive side, we show that instances with bounded respectful tree-width, a more restrictive condition, do have short proofs in Q-resolution. We also give a natural family of formulas with this property that have real-world applications. The second result concerns interpretability. Informally, we say that a first-order formula can be interpreted in another if the first one can be expressed using the vocabulary of the second, plus some extra features. We show that first-order formulas whose propositional translations have short R(const)-proofs, i.e. a generalized version of resolution with DNF-formulas of constant-size terms, are closed under a weaker form of interpretability (that with no extra features), called definability. Our main result is a similar result on interpretability. Also, we show some examples of interpretations and show a systematic technique to transform some Sigma_1-definitions into quantifier-free interpretations. The third and final result is about a relativized weak pigeonhole principle. This says that if at least 2n out of n^2 pigeons decide to fly into n holes, then some hole must be doubly occupied. We prove that the CNF encoding of this principle does not have polynomial-size DNF-refutations, i.e. refutations in the generalized version of resolution with unbounded DNF-formulas. For this proof we discuss the existence of unbalanced low-degree bipartite expanders satisfying a certain robustness condition.