Optimized implementations of blocking and nonblocking collective operations are most important for scalable high-performance applications. Offloading such collective operations into the communication layer can improve performance and asynchronous progression of the operations. However, it is most important that such offloading schemes remain flexible in order to support user-defined (sparse neighbor) collective communications. In this work we propose a design for a collective offload unit. Our hardware design is able to execute dependency graph based representations of collective functions. To cope with the scarcity of memory resources we designed a new point to point messaging protocol which does not need to store information about unexpected messages. The offload unit proposed in this thesis could be integrated into high performance networks such as EXTOLL. Our design achieves a clock frequency of 212 MHz on a Xilinx Virtex6 FPGA, while using less than 10% of the available logic slices and less than 30% of the available memory blocks. Due to the specialization of our design we can accelerate important tasks of the message passing framework, such as message matching by a factor of two, compared to a software implementation running on a CPU with a ten times higher clock speed.

Although healthcare operations management has been an active and popular research direction over the past few years, there is a lack of formal quantitative models to analyze the ambulance o oad delay problem. O oad delays occur when an ambulance arriving at a hospital Emergency Department (ED) is forced to remain in front of the ED until a bed is available for the patient. Thus, the ambulance and the paramedic team are responsible to care for the patient until a bed becomes available inside the ED. But it is not as simple as waiting for a bed, as EDs also admit patients based on acuity levels. While the main cause of this problem is the lack of capacity to treat patients inside the EDs, Emergency Medical Services (EMS) coverage and availability are signi cantly a ected. In this thesis, we develop three network queueing models to analyze the o oad delay problem. In order to capture the main cause of those delays, we construct queueing network models that include both EMS and EDs. In addition, we consider patients arriving to the EDs by themselves (walk-in patients) since they consume ED capacity as well. In the rst model, ED capacity is modeled as the combination of bed, nurse, and doctor. If a patient with higher acuity level arrives to the ED, the current patient's service is interrupted. Thus, the service discipline at the EDs is preemptive resume. We also assume that the time the ambulance needs to reach the patient, upload him into the ambulance, and transfer him to the ED (transit time) is negligible. We develop e cient algorithms to construct the model Markov chain and solve for its steady state probability distribution using Matrix Analytic Methods. Moreover, we derive di erent performance measures to evaluate the system performance under di erent settings in terms of the number of beds at each ED, Length Of Stay (LOS) of patients at an ED, and the number of ambulances available to serve a region. Although capacity limitations and increasing demand are the main drivers for this problem, our computational analysis show that ambulance dispatching decisions have a substantial impact on the total o oad delays incurred. In the second model, the number of beds at each ED is used to model the service capacity. As a result of this modeling approach, the service discipline of patients is assumed to be nonpreemptive priority. We also assume that transit times of ambulances are negligible. To analyze the queueing network, we develop a novel algorithm to construct the system Markov chain by de ning a layer for each ED in a region. We combine the Markov chain layers based on the fact that regional EDs are only connected by the number of available ambulances to serve the region. Using Matrix Analytic Methods, we nd the limiting probabilities and use the results to derive di erent system performance measures. Since each ED's patients are included in the model simultaneously, we solve only for small instances with our current computational resources. In the third model, we decompose the regional network into multiple single EDs. We also assume that patients arriving by ambulances have higher nonpreemptive priority discipline over walk-in patients. Unlike the rst two models, we assume that transit times of ambulances are exponentially distributed. To analyze the decomposed queueing network performance, we construct a Markov chain and solve for its limiting probabilities using Matrix Analytic Methods. While the main objective for the rst two models is performance evaluation, in this model we optimize the steady state dispatching decisions for ambulance patients. To achieve this goal, we develop an approximation scheme for the expected o oad delays and expected waiting times of patients. Computational analysis conducted suggest that larger EDs should be loaded more heavily in order to keep the total o oad delays at minimal levels.

Mobile networks have changed from circuit switched to IP-based mobile wireless packet switched networks. This paradigm shift led to new possibilities and challenges. The development of new capabilities based on IP-based networks is ongoing and raises new problems that have to be tackled, for example, the heterogeneity of current radio access networks and the wide range of data rates, coupled with user requirements and behaviour. A typical example of this shift is the nature of traffic, which is currently mostly data-based; further, forecasts based on market and usage trends indicate a data traffic increase of nearly 11 times between 2013 and 2018. The majority of this data traffic is predicted to be multimedia traffic, such as video streaming and live video streaming combined with voice traffic, all prone to delay, jitter, and packet loss and demanding high data rates and a high Quality of Service (QoS) to enable the provision of valuable service to the end-user. While the demands on the network are increasing, the end-user devices become more mobile and end-user demand for the capability of being always on, anytime and anywhere. The combination of end-user devices mobility, the required services, and the significant traffic loads generated by all the end-users leads to a pressing demand for adequate measures to enable the fulfilment of these requirements. The aim of this research is to propose an architecture which provides smart, intelligent and per end-user device individualised traffic steering for heterogeneous mobile networks to cope with the traffic volume and to fulfil the new requirements on QoS, mobility, and real-time capabilities. The proposed architecture provides traffic steering mechanisms based on individual context data per end-user device enabling the generation of individual commands and recommendations. In order to provide valuable services for the end-user, the commands and recommendations are distributed to the end-user devices in real-time. The proposed architecture does not require any proprietary protocols to facilitate its integration into the existing network infrastructure of a mobile network operator. The proposed architecture has been evaluated through a number of use cases. A proof-of-concept of the proposed architecture, including its core functionality, was implemented using the ns-3 network simulator. The simulation results have shown that the proposed architecture achieves improvements for traffic steering including traffic offload and handover. Further use cases have demonstrated that it is possible to achieve benefits in multiple other areas, such as for example improving the energy efficiency, improving frequency interference management, and providing additional or more accurate data to 3rd party to improve their services.

Cette thèse aborde le problème de la surcharge des réseaux des données des opérateurs mobiles. La croissance des abonnements au haut débit mobile engendre aujourd'hui de nombreux goulots d'étranglement dans ces réseaux. Plus particulièrement, la disponibilité de la bande passante sur les stations de bases est de plus en plus réduite. Pour faire face à cette problématique, les opérateurs mobiles essaient de décharger le trafic des données de leurs infrastructures en déployant des réseaux de substitution à petites cellules, tels que les femtocells ou réseaux WiFi publics. Ces réseaux restent néanmoins très localisés et ne résolvent donc que très partiellement le problème. Ainsi, plus récemment, nous voyons l'émergence des réseaux opportunistes qui visent à transmettent les données en ne se basant que sur les dispositifs mobiles, c-à-d, de dispositif à dispositif. Cette thèse vise à évaluer la faisabilité de décharger le trafic de données mobile à l'aide des hotspots WiFi en étendant leur champs de couverture par l'utilisation des réseaux opportunistes. Pour ce faire, cette thèse propose un cadre pour le déchargement (offloading) de données de façon opportuniste et un mécanisme d'incitation pour encourager la coopération des utilisateurs des dispositifs mobiles. Dans une première partie de cette thèse, nous avons tracé la couverture 3G et WiFi à travers plusieurs lignes de bus à Paris afin d'évaluer la façon dont les utilisateurs et les opérateurs mobiles peuvent bénéficier des réseaux WiFi existants pour le déchargement des données. Nos résultats indiquent que les points d'accès WiFi déployés par les fournisseurs de service Internet peuvent décharger une partie non négligeable du trafic de données, cependant des restrictions telles que le temps de l'association et le processus d'authentification peuvent diminuer la quantité de données transmises. Dans une tentative d'offrir une nouvelle approche pour le déchargement mobile, nous proposons dans un second temps un cadre décisionnel multi-critères, appelé OppLite, pour décharger les données des réseaux de mobiles 3G grâce à des communications dispositif à dispositif opportunistes. Nous avons montré par des simulations qu'un tel déchargement mobile opportuniste peut étendre la couverture et l'efficacité des réseaux cellulaires, permettant un déchargement pouvant aller jusqu'à 36% des données dans certains scénarios. L'efficacité du déchargement mobile par les réseaux opportunistes dépend principalement de la tolérance au délai par l'application et de la coopération des utilisateurs mobiles. Le déchargement opportuniste dépend de la volonté de l'utilisateur d'offrir ses ressources aux autres. Nous avons donc proposé, dans un troisième temps, un mécanisme d'incitation, appelé MINEIRO, qui calcul un rang de réputation basée sur la source des messages reçus par les nœuds intermédiaires. MINEIRO permet à des réseaux composés d'un pourcentage important, allant jusqu'à 60%, de nœuds avec un comportement égoïste sans dégradation des performances dans un scénario de mobilité aléatoire. Au delà de ce pourcentage, MINEIRO permet maintenir un taux de livraison et des délais de livraison constants / This thesis addresses the overload problem of the Wireless Internet service Providers' (WISP) network. The growth of mobile broadband subscription has been leading several bottlenecks to WISPs, such as, bandwidth availability and resource sharing of over a single cellular cell. WISPs can move off data traffic from its infrastructure by deploying small cells, such as femtocells, to public WiFi networks or, more recently, to device-to-device opportunistic networks. This work evaluates the feasibility to offload mobile data traffic using WiFi hotspots, proposes a framework to opportunistic data offloading and an incentive mechanism to encourage users cooperation. We mapped 3G and WiFi coverage through several bus routes in Paris in order to evaluate how users and WISPs can benefit from the existing infrastructure. Our results indicate that the deployed WISPs access points can offload part of the data traffic, however restrictions such as association time and the authentication process may reduce the amount of offloaded data. We propose a multi-criteria decision-making framework, called OppLite, to offload data from 3G networks using opportunistic device-to-device communications. Trace-driven simulations showed that opportunistic mobile offloading can expand coverage and network efficiency, offloading up to 36% of data in certain scenarios. Thus, the effectiveness of opportunistic mobile offloading depends mainly of the delay tolerance of the applications and whether the user cooperates. Since opportunistic offloading depends on the user's willingness to offer his/her resources to others, we propose a message-based incentive mechanism that builds a reputation rank based on the source of messages received by the forwarding nodes, called MINEIRO. The network supports up to 60% of nodes with selfish behavior without performance degradation in a random mobility scenario. After this threshold, MINEIRO kept the delivery rate and the delay constant. Meanwhile, in a scenario with social-based mobility, selfish behavior degrades the network performance quickly

The development of the mobile phone has been rapid. From being a device mainly usedfor phone calls and writing text messages the mobile phone of today, or commonlyreferred to as the smartphone, has become a multi-purpose device. Because of its size andthermal constraints there are certain limitations in areas of battery life andcomputational capabilities. Some say that cloud computing is just another buzzword, away to sell already existing technology. Others claim that it has the potential to transformthe whole IT-industry. This thesis is covering the intersection of these two fields byinvestigating if it is possible to increase the speed of mobile phones by offloadingcomputational heavy mobile phone application functions by using cloud computing. Amobile phone application was developed that conducts three computational heavy tests.The tests were run twice, by not using cloud computing offloading and by using it. Thetime taken to carry out the tests were saved and later compared to see if it is faster to usecloud computing in comparison to not use it. The results showed that it is not beneficial touse cloud computing to carry out these types of tasks; it is faster to use the mobile phone.

No description available.

The focus of this thesis is development of an electronic subsystem for reduced gravity impact simulator. The research part of this thesis firstly covers methods used for simulation of microgravity or reduced gravity and compares them. Another part of research focuses on three selected potential approaches to creating the electronic subsystem for this device. The second, practical, part of this thesis describes the design and development of the electronic subsystem. The foundation of the electronic subsystem is a control unit which has been developed on the basis of an STMicroelectronics microcontroller. Using the control unit, the electronic subsystem measures pressure, position, acceleration and force in the system. Another task of the control unit is control of a stepper motor. Integration of individual elements into the electronic subsystem is described both in terms of software and hardware. Furthermore, a graphic user interface program for PC has been developed as a means to interact with the system. In the final part of the thesis, the operation of the electronic subsystem is described and lastly, the electronic subsystem is evaluated.

The diploma thesis „Mechanisms modelling of multi access into mobile wireless network“ is focusing the wireless network. The diploma thesis contains basic network topology of wireless standard 802.11g and utilizes key identificators of mobile node in dependency on the distance and collision controlling function for simulation purposes. In the next part of this thesis is created LTE mobile network topology, which serves for finding key identificators. In the last part is created offload topology containing both - 802.11g and LTE network. As the result are implemented offloading algorithms, which ensure data traffic switching based on comparing 802.11g and LTE key identificators. Automatically generated and shown figures providing the key statistics are the main output of this thesis.

Optimized implementations of blocking and nonblocking collective operations are most important for scalable high-performance applications. Offloading such collective operations into the communication layer can improve performance and asynchronous progression of the operations. However, it is most important that such offloading schemes remain flexible in order to support user-defined (sparse neighbor) collective communications. In this work we propose a design for a collective offload unit. Our hardware design is able to execute dependency graph based representations of collective functions. To cope with the scarcity of memory resources we designed a new point to point messaging protocol which does not need to store information about unexpected messages. The offload unit proposed in this thesis could be integrated into high performance networks such as EXTOLL. Our design achieves a clock frequency of 212 MHz on a Xilinx Virtex6 FPGA, while using less than 10% of the available logic slices and less than 30% of the available memory blocks. Due to the specialization of our design we can accelerate important tasks of the message passing framework, such as message matching by a factor of two, compared to a software implementation running on a CPU with a ten times higher clock speed.:1. Task Description 1.1. Theses 2. Introduction 2.1. Motivation 2.2. Outline of this Thesis 2.3. Related Work 2.3.1. NIC Based Packet Forwarding 2.3.2. Hardware Barrier Implementations 2.3.3. ConnectX2 CORE-Direct Collective Offload Support 2.3.4. Collective Offload Support in the Portals 4 API 2.4. Group Operation Assembly Language 2.4.1. GOAL API 2.4.2. Scratchpad Buffer 2.4.3. Schedule Execution 2.5. The EXTOLL Network 2.6. Field Programmable Gate Arrays 3. Dealing with Constrained Resources 3.1. Hardware Limitations 3.2. Common Collective Functions in GOAL 3.3. Schedule Representation for the Hardware GOAL Interpreter 3.4. Executing Large Schedules using a small amount of Memory 3.4.1. Limits of Previously Suggested Approaches 3.4.2. Testing for Deadlocks in Schedules 3.4.3. Transforming Process Local Schedules into Global Schedules 3.4.4. Predetermined Buffer Locations 3.5. Queueing Active Operations in Hardware 3.6. Designing a Low-Memory-Footprint Point to Point Protocol 3.6.1. Arrival Times 3.6.2. Eager Protocol 3.6.3. Rendezvous Protocol 3.6.4. A Protocol without an Unexpected Queue 3.7. Protocol Verification 3.7.1. Capabilities of the Model Checker SPIN 3.7.2. Modeling the Protocol 3.7.3. Limitations of the Basic Protocol 4. The Matching Problem 4.1. Matching on the Host CPU 4.2. Implementation Methodology 4.3. Matching Unit Interface 4.4. Matching Unit Implementation 4.4.1. Slot Management Unit 4.4.2. The Input Consumer 4.4.3. The Output Generator 4.4.4. The Matching Unit 4.5. Slot Management Unit for Non-synchronous Transfers 5. The GOAL Interpreter 5.1. Schedule Interpreter Design 5.1.1. The Active Queue 5.1.2. The Dependency Resolver 5.2. Transceiver Interface 5.3. The Starter 5.3.1. Starting Operations 5.3.2. Processing Incoming Packets 5.3.3. Incoming Non-synchronous Packets 5.3.4. Presorting the Active Queue 5.3.5. Arbitration Units 5.3.6. IN-Filter 5.3.7. Outcommand Manager 5.3.8. Non-synchronous Protocol 5.3.9. Send Protocol 5.3.10. Receive Protocol 5.3.11. Local Operations on FPGA 6 Evaluation 6.1. Performance Analysis 6.2. Future Work 6.3. Conclusions Bibliography

Optimized implementations of blocking and nonblocking collective operations are most important for scalable high-performance applications. Offloading such collective operations into the communication layer can improve performance and asynchronous progression of the operations. However, it is most important that such offloading schemes remain flexible in order to support user-defined (sparse neighbor) collective communications. In this work, we describe an operating system kernel-based architecture for implementing an interpreter for the flexible Group Operation Assembly Language (GOAL) framework to offload collective communications. We describe an optimized scheme to store the schedules that define the collective operations and show an extension to profile the performance of the kernel layer. Our microbenchmarks demonstrate the effectiveness of the approach and we show performance improvements over traditional progression in user-space. We also discuss complications with the design and offloading strategies in general.