In this paper in order to improve the performance of the IEEE 802.11e WLAN (Wireless Local Area Network), we propose a simple MAC layer access control mechanism SCAS (Self-Conscious Congestion Avoidance Scheme). During the contention period, if the traffic load is high, we adaptively temporary stop some STAs (stations) contending for the channel to reduce the collision rate according to the network conditions. SCAS is on the MAC layer. Hence, it can not only be applied to the IEEE 802.11e medium access method (EDCA [2]) but also be applied to AEDCF [12]. Besides, each STA can operate SCAS by itself not through AP (Access Point) or higher layer. Finally we evaluate the performance of SCAS on EDCA and AEDCF through simulations. Results show that SCAS can not only reduce collision rate of EDCA and AEDCF efficiently but also improve mean delay time and system throughput especially in high traffic load conditions.

International Telemetering Conference Proceedings / October 18-21, 2004 / Town & Country Resort, San Diego, California / Sensor network applications typically require continuous monitoring of physical phenomena for extended periods of time under severe energy resource constraints. Accordingly, design considerations for sensor Media Access Control (MAC) protocols depart significantly from those of traditional wireless MAC protocols that largely ignore the energy factor. In this paper, we reexamine the design space of wireless sensor MAC protocols and modify IEEE 802.11 Distributed Coordination Function (DCF) to incorporate energy-adaptive contention mechanisms for prolonging sensor lifetime. Performance of the proposed schemes is evaluated with DCF as a baseline and results indicate the benefits of energy-aware mechanisms for sensor MAC protocols.

Data access is an essential part of any program, and is especially critical to the performance of parallel computing systems. The objective of this work is to investigate factors that affect data access parallelism in parallel computing systems, and design/evaluate methods to improve such parallelism - and thereby improving the performance of corresponding parallel systems. We focus on data access contention and network resource contention in representative parallel and distributed systems, including transactional memory system, Geo-replicated transactional systems and MapReduce systems. These systems represent two widely-adopted abstractions for parallel data accesses: transaction-based and distributed-system-based. In this thesis, we present methods to analyze and mitigate the two contention issues. We first study the data contention problem in transactional memory systems. In particular, we present a queueing-based model to evaluate the impact of data contention with respect to various system configurations and workload parameters. We further propose a profiling-based adaptive contention management approach to choose an optimal policy across different benchmarks and system platforms. We further develop several analytical models to study the design of transactional systems when they are Geo-replicated. For the network resource contention issue, we focus on data accesses in distributed systems and study opportunities to improve upon the current state-of-art MapReduce systems. We extend the system to better support map task locality for dual-map-input applications. We also study a strategy that groups input blocks within a few racks to balance the locality of map and reduce tasks. Experiments show that both mechanisms significantly reduce off-rack data communication and thus alleviate the resource contention on top-rack switch and reduce job execution time. In this thesis, we show that both the data contention and the network resource contention issues are key to the performance of transactional and distributed data access abstraction and our mechanisms to estimate and mitigate such problems are effective. We expect our approaches to provide useful insight on future development and research for similar data access abstractions and distributed systems.

A l’heure des développements technologiques majeurs dans le domaine de la radiochirurgie, il est important de s’intéresser aux précisions des différentes options qui se présentent lors de la préparation d’un traitement. Dans ce travail, ce sont les systèmes de contention utilisés pour les traitements radiochirurgicaux délivrés à l’aide de l’appareil Gamma Knife qui sont au centre de l’attention.Il a été mis en évidence par les études cliniques et expérimentales que chaque système de contention présente une capacité différente à maintenir l’immobilité du patient au cours du traitement. Les analyses des différentes mesures effectuées ont permis d’établir la capacité de chaque système à garantir une correspondance entre les coordonnées théoriques et effectives du traitement, ce qui se traduit par la précision attribuable à chaque système. Les points forts et les faiblesses de chacun des systèmes de contention ont été mis en évidence. Dans certainessituations, il a été possible de proposer une méthode permettant de corriger en totalité ou en partie les imprécisions observées. Ceci n’est néanmoins pas possible actuellement pour tous les systèmes de contention. Cela a permis d’établir une proposition d’algorithme décisionnel afin d’aider au choix du système de contention optimal à utiliser en fonction de la situation rencontrée. Enfin, ce travail met en évidence les limitations de chaque technique et énonce des pistes d’amélioration et de développement futur dans ce domaine. / Doctorat en Sciences biomédicales et pharmaceutiques (Médecine) / info:eu-repo/semantics/nonPublished

Mémoire numérisé par la Direction des bibliothèques de l'Université de Montréal.

Au cours de cette étude, l expression des gènes codants de quatre sous types de récepteurs centraux (GABAA, 5-HT1A, m-opioïdes et a1-adrénergiques) impliqués dans la modulation de l anxiété a été évaluée. L impact de l anxiété trait sur les réponses immunitaires cellulaire et humorale et sur le niveau du statut oxydatif au niveau du SNC (neurones et cellules gliales) et au niveau périphérique (lymphocytes, granulocytes et monocytes) a été mis en évidence. En même temps, l effet de l anxiété trait couplée au modèle de stress expérimental anxiogène par contention, sur les réponses immunitaires cellulaire et humorale a été évalué. En effet, le niveau élevé d anxiété induit d une part, un effet dépressif significatif sur l immunité cellulaire (lymphocytes totaux, TCD4+ et TCD8+) et humorale (IgA et IgE) et d autre part, une augmentation significative du niveau intracellulaire des espèces réactives de l oxygène (ERO) des neurones et des cellules gliales au niveau du cortex cérébral, du cervelet et de l hippocampe et des granulocytes, des lymphocytes et des monocytes au niveau périphérique. De la même manière, l anxiété trait couplée au stress aigu et chronique par contention a provoqué une dépression de certains paramètres de l immunité cellulaire (lymphocytes totaux, TCD4+, TCD8+ et NK) et humorale (IgA, E et G) et une stimulation des autres paramètres (granulocytes et monocytes). Ces travaux ont permis ainsi de valider scientifiquement le caractère anxiogène du modèle du stress par contention, d établir de manière valide et reproductible le lien et la corrélation entre le niveau élevé d anxiété chez des animaux et leur statut oxydatif inducteur d une cytotoxicité ainsi que le rôle de l expression des gènes codants de 4 sous types de récepteurs dans l expression de cette anxiété élevée et d un statut oxydatif important au niveau des cellules périphériques du système immunitaire et des neurones et cellules gliales au niveau central. / In this study, genes expression from four central receptors (GABAA, 5-HT1A, m-opioïdes and a1-adrenergic) involved in the modulation of anxiety was assessed. The impact of anxiety on the cellular and humoral immunity and on the oxidative status at the SNC (neurons and glial cells) and peripheral (lymphocytes, granulocytes and monocytes) level was highlighted. At the same time, the effect of anxiety coupled with an anxiogenic restraint stress, on the cellular and humoral immunity was also evaluated. Indeed, the high level of anxiety induced firstly, a significant depressive effect on cellular (total lymphocytes, TCD4+ and TCD8+) and humoral (IgA and IgE) immunity, and secondly, a significant increase of the level of intracellular reactive oxygen species (ROS) of neurons and glial cells in the cerebral cortex, the cerebellum and the hippocampus and in the peripheral blood granulocytes, monocytes and lymphocytes. In the same way, the anxiety coupled to acute and chronic restraint stress provoked, a depression of some parameters of cellular (total lymphocytes, TCD4+, TCD8+ and NK) and humoral (IgA, E and G) immunity, and a stimulation of others (granulocytes and monocytes). These works thus made it possible to validate scientifically the anxiogenic character of the model of restraint stress, to establish in a valid and reproducible way the bond and the correlation between the high level of anxiety in animals and their oxidative status inductive of a cytotoxicity as well as the role of the expression of coding genes of 4 receptors in the expression of this high anxiety and of a significant oxidative status at the level of the peripheral cells of the immune system and the neurons and glial cells at the central level

Chip Multi-Processor (CMP) platforms, well-established in the server, desktop and embedded domain, succeeded in overcoming the power consumption and heat dissipation bottlenecks by integrating multiple cores, less complex and powerful than their single-core ancestors, in a single die. A major issue induced by the design of the CMPs is contention for the shared resources of the platform, Last Level Cache (LLC) and main memory bandwidth. Applications, running concurrently on the cores, compete with each other for the shared resources, and are subject to performance degradation. The way applications are assigned to the CMP, is crucial for the overall performance of the system. A scheduling policy that accounts for contention will bring high performance speed-ups, whereas an agnostic one will generate unpredictable contention conditions. For this reason the significance of the scheduler has been elevated, as it is the component that determines which applications utilize the resources each time period.In this thesis, we address cross-core interference on CMP platforms, by designing scheduling policies that improve performance and fairness. We deal with contention in three ways. In our first approach, we incorporate the notion of progress in order to balance unfairness among the applications of the workload. Performance degradation is not evenly distributed and progress greatly varies among them. In order to provide a fair execution environment, we monitor, at run-time, applications assigned to the CPU and prioritize them based on the extent at which they are affected by contention.In our second approach, we target performance by mitigating contention on shared resources. It is necessary to decide, out of all the possible application schedules, the one that generates the least amount of resource interference. To achieve that, the first indispensable step is to extract an interference profile for the applications executed on the CMP. We accomplish that by applying pressure to all levels of memory hierarchy and identifying the point at which performance is compromised. From our analysis, we understand that shared resources can tolerate pressure of certain amount; applications can be grouped together if the overall generated pressure does not reach the saturation point of the shared resources. Having extracted this information, we proceed to the placement of the application in such a way that overall resource requirements are as balanced as possible across the execution.Finally, we design a policy in order to improve performance and fairness at the same time. Applications that heavily rely on the LLC are separated from those with high main memory bandwidth, in order to avoid the destructive effects caused by the LLC thrashing behavior of the latter. The group executed on the CPU is determined based on the key observation that the overall requirements of the group should not exceed the saturation limits of the CMP. Additionally, during execution, the progress for each application is estimated and those with the least accumulated progress are prioritized.Our proposed policies are evaluated in an Intel Xeon E5-2620 v3 processor. A variety of benchmark suites were utilized to generate mixes of diverse characteristics. Our methodologies are implemented in user-space and can be deployed on Linux-based systems. Experimental results show the benefits of tackling contention in shared resources. We achieve throughput gains of up to 16% and unfairness is reduced by 2.37x on average compared to Linux scheduler.

No description available.

Les architectures multi-cœurs utilisant des mémoire bloc-notes sont des architectures attrayantes pour l'exécution des applications embarquées temps-réel, car elles offrent une grande capacité de calcul. Cependant, les systèmes temps-réel nécessitent de satisfaire des contraintes temporelles, ce qui peut être compliqué sur ce type d'architectures à cause notamment des ressources matérielles physiquement partagées entre les cœurs. Plus précisément, les scénarios de pire cas de partage du bus de communication entre les cœurs et la mémoire externe sont trop pessimistes. Cette thèse propose des stratégies pour réduire ce pessimisme lors de l'ordonnancement d'applications sur des architectures multi-cœurs. Tout d'abord, la précision du pire cas des coûts de communication est accrue grâce aux informations disponibles sur l'application et l'état de l'ordonnancement en cours. Ensuite, les capacités de parallélisation du matériel sont exploitées afin de superposer les calculs et les communications. De plus, les possibilités de superposition sont accrues par le morcellement de ces communications. / Multi-core architectures using scratch pad memories are very attractive to execute embedded time-critical applications, because they offer a large computational power. However, ensuring that timing constraints are met on such platforms is challenging, because some hardware resources are shared between cores. When targeting the bus connecting cores and external memory, worst-case sharing scenarios are too pessimistic. This thesis propose strategies to reduce this pessimism. These strategies offer to both improve the accuracy of worst-case communication costs, and to exploit hardware parallel capacities by overlapping computations and communications. Moreover, fragmenting the latter allow to increase overlapping possibilities.

Broadcasting is an essential and effective data propagation mechanism with several important applications, such as route discovery, address resolution and many other network services. Though data broadcasting has many advantages, it can also cause a high degree of contention, collision and congestion, leading to what is known as 'broadcast storm problems'. Broadcasting has traditionally been based on the flooding protocol, which simply overflows the network with a high number of rebroadcast messages until these reach all the network nodes. A good probabilistic broadcast protocol can achieve high saved rebroadcast (SRB), low collision and a lower number of relays. When a node is in a sparse region of the network, rebroadcasting is relatively more important while the potential redundancy of rebroadcast is low because there are few neighbours which might rebroadcast the packet unnecessarily. Further, in such a situation, contention over the wireless medium resulting from Redundant broadcasts is not as serious as in scenarios with medium or high density node populations. This research proposes a dynamic probabilistic approach that dynamically fine-tunes the rebroadcast probability according to the number of neighbouring nodes distributed in the ad-hoc network for routing request packets (RREQs) without requiring the assistance of distance measurements or location-determination devices. The main goal of this approach is to reduce the number of rebroadcast packets and collisions in the network. The performance of the proposed approach is investigated and compared with simple AODV, fixed-probabilistic and adjusted-probabilistic flooding [1] schemes using the GloMoSim network simulator and a number of important MANET parameters, including node speed, traffic load and node density under a Random Waypoint (RWP) mobility model. Performance results reveal that the proposed approach is able to achieve higher SRB and less collision as well as a lower number of relays than fixed probabilistic, simple AODV and adjusted-probabilistic flooding. In this research, extensive simulation experiments have been conducted in order to study and analyse the proposed dynamic probabilistic approach under different mobility models. The mobility model is designed to describe the movement pattern of mobile customers, and how their position, velocity and acceleration change over time. In this study, a new enhanced dynamic probabilistic flooding scheme is presented. The rebroadcast probability p will be calculated dynamically and the rebroadcasting decision will be based on the average number of nodes in the ad-hoc networks. The performance of the new enhanced algorithm is evaluated and compared to the simple AODV, fixed-probabilistic, adjusted-probabilistic and dynamic-probabilistic flooding schemes. It is demonstrated that the new algorithm has superior performance characteristics in terms of collision, relays and SRB. Finally, the proposed schemes are tested and evaluated through a set of experiments under different mobility models to demonstrate the relative merits and capabilities of these schemes.