Ces dernières années, de nouvelles problématiques sont nées au vu des considérations écologiques de plus en plus présentes dans notre société. Dans le domaine de la technologie de l'Information, les centres de calcul consomment actuellement environ 1.5% de l'électricité mondiale. Cela ne cesse d’augmenter en raison de l'évolution de nombreux domaines et particulièrement du Cloud Computing. Outre cet aspect environnemental, le contrôle de la consommation d’énergie fait désormais partie intégrante des paramètres de Qualité de Service (QoS) incombant aux fournisseurs de services de Cloud Computing. En effet, ces fournisseurs de services à la demande proposent à leurs utilisateurs un contrat de QoS, appelé SLA (Service Level Agreement), qui définit de manière précise la qualité de service qu’ils s’engagent à respecter. Le niveau de QoS proposé influence directement la qualité d’utilisation des services par les utilisateurs, mais aussi la consommation et le rendement général de l’ensemble des ressources de calcul utilisées, impactant fortement les bénéfices des fournisseurs de services.Le Cloud Computing étant intrinsèquement lié à la virtualisation des ressources de calcul, une élaboration de modèles d’architecture matérielle et logicielle est proposée afin de définir les caractéristiques de l’environnement considéré. Ensuite, une modélisation détaillée de paramètres de QoS en termes de performance, de sûreté de fonctionnement, de sécurité des données et de coûts est proposée. Des métriques associées à ces paramètres sont définies afin d’étendre les possibilités d'évaluation des SLA. Ces modélisations constituent la première contribution de cette thèse.Il convient alors de démontrer comment l’utilisation et l’interprétation de plusieurs métriques de QoS ouvrent la possibilité d'une analyse plus complexe et plus fine de la perspicacité des algorithmes de placement. Cette approche multi-critères leur apporte des informations importantes sur l’état de leur système qu’ils peuvent analyser afin de gérer le niveau de chaque paramètre de QoS. Ainsi, quatre métriques antagonistes, incluant la consommation énergétique, ont été sélectionnées et utilisées conjointement dans plusieurs algorithmes de placement de manière à montrer leur pertinence, l’enrichissement qu’elles apportent à ces algorithmes, et comment un fournisseur de service peut tirer profit des résultats d’une optimisation multi-objectifs. Cette seconde contribution présente un algorithme génétique (GA) ainsi que deux algorithmes gloutons. L’analyse du comportement de l'algorithme génétique a permis de démontrer différents intérêts d’une optimisation multi-critères appliquée à des métriques de QoS habituellement ignorées dans les études dédiées au Cloud Computing.La troisième contribution de cette thèse propose une étude de l’impact de l'utilisation des métriques de QoS sur l’ordonnancement de machines virtuelles au cours du temps. Pour cela, le simulateur CloudSim a été exploité et étendu afin d'améliorer ses fonctionnalités de gestion de consommation énergétique. Tout d’abord par l’ajout du DVFS (Dynamic Voltage & Frequency Scaling) apportant une gestion dynamique très précise des fréquences de fonctionnement CPU, puis la possibilité de reconfiguration de machines virtuelles et enfin par la gestion dynamique des évènements. Les simulations effectuées mettent en jeu l'ensemble de ces outils énergétiques ainsi que les algorithmes de placement et évaluent chacune des métriques de QoS sélectionnées. Ces simulations donnent une vision temporelle de l’évolution de celles-ci, en fonction des algorithmes utilisés et de plusieurs configurations d’optimisation du GA. Cela permet d'analyser sous différents angles le comportement des algorithmes gloutons, l'impact des optimisations du GA, et l'influence des métriques les unes par rapport aux autres.Une collaboration a pu être établie avec le laboratoire CLOUDS Laborartory de Melbourne, dirigé par Prof. Rajkumar Buyya. / In recent years, new issues have arisen in environmental considerations, increasingly pointed out in our society. In the field of Information Technology, data centers currently consume about 1.5% of world electricity. This increasing is due to changes in many areas, especially in Cloud Computing. Besides this environmental aspect, the management of energy consumption has become an important field of Quality of Service (QoS), in the responsibility of Cloud providers. These providers propose a QoS contract called SLA (Service Level Agreement), which specify the level of QoS given to users. The level of QoS offered directly influences the quality of the users' utilization, but also the overall energy consumption and performance of computing resources, which strongly affect profits of the Cloud providers. Cloud computing is intrinsically linked to the virtualization of computing resources. A model of hardware and software architecture is proposed in order to define the characteristics of the environment considered. Then, a detailed modeling of QoS parameters in terms of performance, dependability, security and cost is proposed. Therefore, QoS metrics, associated to these parameters are defined in order to extend the possibilities for evaluating the SLA. These models represent the first contribution of this thesis. Then, it is necessary to illustrate how the use and interpretation of several QoS metrics open the possibility of a more complex and precise analysis of algorithms' insight. This multi-criteria approach, that provides useful informations about the system's status can be analyzed to manage the QoS parameters' level. Thus, four antagonists metrics, including energy consumption, are selected and used together in several scheduling algorithms which allow to show their relevance, the enrichment given to these algorithms, and how a Cloud provider can take advantage of the results of this kind of multi-objective optimization. The second contribution presents a genetic algorithm (GA) and two greedy algorithms. The analysis of the genetic algorithm behavior allows to show different interests of a multi-criteria optimization applied to QoS metrics, usually ignored in studies dedicated to Cloud Computing. The third contribution of this thesis proposes a study of the impact of the use of QoS metrics in virtual machines scheduling. The simulator CloudSim has been used and expanded to improve its energy-aware tools. The DVFS (Dynamic Voltage & Frequency Scaling), providing a highly accurate dynamic management of CPU frequencies, the virtual machines reconfiguration, and the dynamic management of events have been included. The simulations involve all of these energy tools and placement algorithms, and evaluate each selected QoS metrics. These simulations allow to see the evolution in time of these metrics, depending on the algorithms used and the behavior of the GA in different optimizations configurations. This allows to analyze from different angles the behavior of greedy algorithms, the impact of optimizations GA, and the influence of these metrics one against the others.

Data stream processing in the industrial as well as in the academic field has gained more and more importance during the last years. Consider the monitoring of industrial processes as an example. There, sensors are mounted to gather lots of data within a short time range. Storing and post-processing these data may occasionally be useless or even impossible. On the one hand, only a small part of the monitored data is relevant. To efficiently use the storage capacity, only a preselection of the data should be considered. On the other hand, it may occur that the volume of incoming data is generally too high to be stored in time or–in other words–the technical efforts for storing the data in time would be out of scale. Processing data streams in the context of this thesis means to apply database operations to the stream in an on-the-fly manner (without explicitly storing the data). The challenges for this task lie in the limited amount of resources while data streams are potentially infinite. Furthermore, data stream processing must be fast and the results have to be disseminated as soon as possible. This thesis focuses on the latter issue. The goal is to provide a so-called Quality-of-Service (QoS) for the data stream processing task. Therefore, adequate QoS metrics like maximum output delay or minimum result data rate are defined. Thereafter, a cost model for obtaining the required processing resources from the specified QoS is presented. On that basis, the stream processing operations are scheduled. Depending on the required QoS and on the available resources, the weight can be shifted among the individual resources and QoS metrics, respectively. Calculating and scheduling resources requires a lot of expert knowledge regarding the characteristics of the stream operations and regarding the incoming data streams. Often, this knowledge is based on experience and thus, a revision of the resource calculation and reservation becomes necessary from time to time. This leads to occasional interruptions of the continuous data stream processing, of the delivery of the result, and thus, of the negotiated Quality-of-Service. The proposed robustness concept supports the user and facilitates a decrease in the number of interruptions by providing more resources.

[pt] Este estudo examinou três determinantes das dimensões da qualidade de serviço de refeição (ambiente físico, comida, e serviço) na imagem do restaurante, na satisfação do cliente e nas intenções comportamentais por meio de um questionário on-line auto-administrado coletado de clientes do Outback Steakhouse Brasil. A modelagem dos dados, realizada via equações estruturais, mostrou que a qualidade do ambiente físico, comida e serviço foram determinantes significativos da imagem do restaurante, e que a imagem do restaurante foi determinante significativo da satisfação do consumidor. Além disso, os resultados reforçam que a satisfação do consumidor afeta significativamente as intenções comportamentais. As descobertas podem fornecer aos administradores de restaurantes uma orientação de como melhorar a satisfação do cliente e o nível de intenção comportamental. / [en] This study examined three determinants of foodservice quality dimensions (physical environment, food, and service) on restaurant image, customer satisfaction, and behavioral intentions via a self-administered online questionnaire collected from customers of Outback Steakhouse Brazil. Structural equation modeling showed that the quality of the physical environment, food, and service were significant determinants of restaurant image, and that restaurant image was significant determinant of customer satisfaction. In addition, the results reinforced that customer satisfaction is a significant predictor of behavioral intentions. The findings may provide restaurateurs with a guideline for enhancing customer satisfaction and behavioral intention level.

Network traffic forecasting estimates future network traffic based on historical traffic observations. It has a wide range of applications, and substantial attention has been dedicated to this research area. Cellular networks serve as the backbone of modern-day communication systems, which support billions of users throughout the world and can help improve the quality of urban life to a great extent. Therefore, accurate traffic prediction is becoming more important for network planning, control management, and the Quality of Service. Diverse methods, including neural network-based methods and data mining methods, have been used for this goal. The Recurrent Neural family is well known for time series data modeling, which predicts the future time series based on the historical data being fed as input to neural nets which may have large time lags with variable lengths. RNN includes several network architectures, such as vanilla RNN and Long Short Term memory (LSTM), that can learn temporal patterns and long-term dependencies in vast sequences of arbitrary length. This paper proposes three models based on LSTM architecture, a multi-layer LSTM with Auto-Encoder, and an AE-LSTM combined with a Multi-Layer Perceptron neural network. The results of each model are discussed in the paper. Simulation outcomes were implemented in Python and compared to existing algorithms, demonstrating high efficacy and performance.

No / An investigation is carried out on the nature of QoS measures for queues with correlated traffic in both discrete and continuous time domains. The study focuses on the single server GI(G)/M-[x]/1/N and GI(G)/Geo([x])/1/N queues with finite capacity, N, a general batch renewal arrival process (BRAP), GI(G) and either batch Poisson, M-[x] or batch geometric, Geo([x]) service times with general batch sizes, X. Closed form expressions for QoS measures, such as queue length and waiting time distributions and blocking probabilities are stochastically derived and showed to be, essentially, time domain invariant. Moreover, the sGGeo(sGGo)/Geo/l/N queue with a shifted generalised geometric (sGGeo) distribution is employed to assess the adverse impact of varying degrees of traffic correlations upon basic QoS measures and consequently, illustrative numerical results are presented. Finally, the global balance queue length distribution of the M-Geo/M-Geo/1/N queue is devised and reinterpreted in terms of information theoretic principle of entropy maximisation. (C) 2014 Elsevier Inc. All rights reserved.

No / Over the past decade, wireless communications has experienced tremendous growth, and this growth is likely to multiply in the near future. The proliferation of mobile users and an ever increasing demand for multimedia services has resulted in greater capacity requirements. Radio frequency spectrum is scarce and cannot meet this ever increasing demand and the required Quality of Service (QoS) will no longer be achieved if efficient Radio Resource Management (RRM) solutions are not found. Conventional Radio Access Networks (RAN) have standalone Base Stations (BS) with capacity preconfigured for peak loads. These RANs have high call blocking and dropping rates since BSs resources cannot be shared. Cloud based RANs (C-RAN) have been proposed as a cost and energy efficient way of meeting high capacity demand of future wireless access networks by consolidating BSs to the cloud. Instead of relying on rejection of new call requests due to limited BS resources, C-RAN takes benefit of the cloud elasticity, which allows dynamic provisioning of cloud BS resources. This paper presents a novel C-RAN Call Admission Control (C-RAN CAC) to ensure Grade of Service (GoS) by improving blocking probability and improvement of call waiting times. Call blocking probability, call average waiting time and system utilization are used to evaluate the performance of the proposed CAC algorithm.

Yes / Unmanned Aerial Vehicles (UAVs) are considered, nowadays, as a futuristic and robust paradigm for 5G wireless networks, in terms of providing Internet connectivity services onto infrastructure cellular networks. In this paper, the interference regime caused by multiple downlink aerial wireless transmission beams has been highlighted. This has been introduced by estimating the UAVs coverage area that is analytically derived in a tractable closed-form expression. The rationale of the analysed coverage approach relies on observing and adapting the joint aerial distance between the aerial base stations. This can minimize the intra-overlapped coverage and ultimately maximize the overall coverage performance for a better quality of service demands. The novelty of our approach brings useful design insights for UAVs system-level performance that technically helps in aerial coverage computations without the need of performing an aerial deployment setup. To the end, the performance effectiveness of our methodology has been tested under an urban propagation environment conditions, in which the original probabilistic channel model approximation has been taken into account. Moreover, this paper identifies the interference issue of such an aerial network as a shrinkage or distortion phenomenon.

Traditionally the Internet is used for the following applications: FTP, e-mail and Web traffic. However in the recent years the Internet is increasingly supporting emerging applications such as IP telephony, video conferencing and online games. These new applications have different requirements in terms of throughput and delay than traditional applications. For example, interactive multimedia applications, unlike traditional applications, have more strict delay constraints and less strict loss constraints. Unfortunately, the current Internet offers only a best-effort service to all applications without any consideration to the applications specific requirements. In this thesis three existing Active Queue Management (AQM) mechanisms are modified by incorporating into these a control function to condition routers for better Quality of Service (QoS). Specifically, delay is considered as the key QoS metric as it is the most important metric for real-time multimedia applications. The first modified mechanism is Drop Tail (DT), which is a simple mechanism in comparison with most AQM schemes. A dynamic threshold has been added to DT in order to maintain packet queueing delay at a specified value. The modified mechanism is referred to as Adaptive Drop Tail (ADT). The second mechanism considered is Early Random Drop (ERD) and, iii in a similar way to ADT, a dynamic threshold has been used to keep the delay at a required value, the main difference being that packets are now dropped probabilistically before the queue reaches full capacity. This mechanism is referred to as Adaptive Early Random Drop (AERD). The final mechanism considered is motivated by the well known Random Early Detection AQM mechanism and is effectively a multi-threshold version of AERD in which packets are dropped with a linear function between the two thresholds and the second threshold is moveable in order to change the slope of the dropping function. This mechanism is called Multi Threshold Adaptive Early Random Drop (MTAERD) and is used in a similar way to the other mechanisms to maintain delay around a specified level. The main focus with all the mechanisms is on queueing delay, which is a significant component of end-to-end delay, and also on reducing the jitter (delay variation) A control algorithm is developed using an analytical model that specifies the delay as a function of the queue threshold position and this function has been used in a simulation to adjust the threshold to an effective value to maintain the delay around a specified value as the packet arrival rate changes over time. iv A two state Markov Modulated Poisson Process is used as the arrival process to each of the three systems to introduce burstiness and correlation of the packet inter-arrival times and to present sudden changes in the arrival process as might be encountered when TCP is used as the transport protocol and step changes the size of its congestion window. In the investigations it is assumed the traffic source is a mixture of TCP and UDP traffic and that the mechanisms conserved apply to the TCP based data. It is also assumed that this consists of the majority proportion of the total traffic so that the control mechanisms have a significant effect on controlling the overall delay. The three mechanisms are evaluated using a Java framework and results are presented showing the amount of improvement in QoS that can be achieved by the mechanisms over their non-adaptive counterparts. The mechanisms are also compared with each other and conclusions drawn.

As the number of people using wireless networks is increasing, the need to reduce the vulnerability of wireless networks from node or link failures that cause loss of data is becoming a priority. Also the present techniques and topologies used for wireless networking are not sufficient to handle the traffic load even if we solve the issues of reliability and resilience. Packet loss or delay is increasingly likely due to the increase in the number of packets as technology is evolving and more video and voice packets along with the data packets are being transmitted. Only the efficient and intelligent use of the shared medium can solve the problem and help in avoiding the collision or delay among the packets using a newly proposed intelligent topology. Wireless technology offers the potential to replace wires from many applications, particularly for the rapid deployment of networks for permanent or temporary use. Fiber_ optic metropolitan area networks (WAN) provide security and resilience. A target of the research was to match this in the wireless environment. This research investigates the suitability if using wireless technology for the establishment of a MAN by adding features to enhance resilience. We proposed a mechanism that may be rapidly deployed and provide automatic configuration. Research work and simulation design has been used to develop a new wireless network topology for an efficient and intelligent packet transmission by identifying reliable routes. This novel idea will help give wireless as well as mobile technology a clear edge over wired technology, not only in the case of mobility but also in the case of security of data and other services. A decision based routing tree mechanism has been developed, that can establish routes in an ad-hoc network which may than be configured into a logical dual ring. At the same time the proposed system proposes to embed the quality of service mechanisms, resilience, and reliability features from RPR. The simulations were created using Microsoft Visual Studio.Net for the Decision based routing algorithm. The results were compared with an existing LAR algorithm. We have obtained 95% confidence intervals on all the performance analysis results to indicate accuracy.

The exponential growth of Internet applications has created new challenges for the control and administration of large-scale networks, which consist of heterogeneous elements under dynamically changing traffic conditions. These emerging applications need guaranteed service levels, beyond those supported by best-effort networks, to deliver the intended services to the end user. Several models have been proposed for a Quality of Service (QoS) framework that can provide the means to transport these services. It is desirable to find efficient routing strategies that can meet the strict routing requirements of these applications. QoS routing is considered as one of the major components of the QoS framework in communication networks. In QoS routing, paths are selected based upon the knowledge of resource availability at network nodes and the QoS requirements of traffic. Several QoS routing schemes have been proposed that differ in the way they gather information about the network state and the way they select paths based on this information. The biggest downside of current QoS routing schemes is the frequent maintenance and distribution of global state information across the network, which imposes huge communication and processing overheads. Consequently, scalability is a major issue in designing efficient QoS routing algorithms, due to the high costs of the associated overheads. Moreover, inaccuracy and staleness of global state information is another problem that is caused by relatively long update intervals, which can significantly deteriorate routing performance. Localized QoS routing, where source nodes take routing decisions based solely on statistics collected locally, was proposed relatively recently as a viable alternative to global QoS routing. It has shown promising results in achieving good routing performance, while at the same time eliminating many scalability related problems. In localized QoS routing each source¿destination pair needs to determine a set of candidate paths from which a path will be selected to route incoming flows. The goal of this thesis is to enhance the scalability of QoS routing by investigating and developing new models and algorithms based on the localized QoS routing approach. For this thesis, we have extensively studied the localized QoS routing approach and demonstrated that it can achieve a higher routing performance with lower overheads than global QoS routing schemes. Existing localized routing algorithms, Proportional Sticky Routing (PSR) and Credit-Based Routing (CBR), use the blocking probability of candidate paths as the criterion for selecting routing paths based on either flow proportions or a crediting mechanism, respectively. Routing based on the blocking probability of candidate paths may not always reflect the most accurate state of the network. This has motivated the search for alternative localized routing algorithms and to this end we have made the following contributions. First, three localized bandwidth-constrained QoS routing algorithms have been proposed, two are based on a source routing strategy and the third is based on a distributed routing strategy. All algorithms utilize the quality of links rather than the quality of paths in order to make routing decisions. Second, a dynamic precautionary mechanism was used with the proposed algorithms to prevent candidate paths from reaching critical quality levels. Third, a localized delay-constrained QoS routing algorithm was proposed to provide routing with an end-to-end delay guarantee. We compared the performance of the proposed localized QoS routing algorithms with other localized and global QoS routing algorithms under different network topologies and different traffic conditions. Simulation results show that the proposed algorithms outperform the other algorithms in terms of routing performance, resource balancing and have superior computational complexity and scalability features. / Umm AlQura University, Saudi Arabia