The modern world has opened the gates to a lot of advancements in cloud computing, particularly in the field of Cloud Gaming. The most recent development made in this area is the open-source cloud gaming system called GamingAnywhere. The relationship between the CPU and GPU is what is the main object of our concentration in this thesis paper. The Graphical Processing Unit (GPU) performance plays a vital role in analyzing the playing experience and enhancement of GamingAnywhere. In this paper, the virtualization of the GPU has been concentrated on and is suggested that the acceleration of this unit using NVIDIA CUDA, is the key for better performance while using GamingAnywhere. After vast research, the technique employed for NVIDIA CUDA has been chosen as gVirtuS. There is an experimental study conducted to evaluate the feasibility and performance of GPU solutions by VMware in cloud gaming scenarios given by GamingAnywhere. Performance is measured in terms of bitrate, packet loss, jitter and frame rate. Different resolutions of the game are considered in our empirical research and our results show that the frame rate and bitrate have increased with different resolutions, and the usage of NVIDIA CUDA enhanced GPU.

Cloud computing is a set of technologies that provide efficient and effective usage of information technology resources. The application of cloud computing in games is a new market that is currently growing. Applying cloud computing to games results in a new gaming platform for the users, referred to as cloud gaming. This new platform has a set of different features that affect the user acceptance of this new platform. Problem There is a lack of understanding on how the users perceive cloud gaming. Investigat-ing the factors that affect the users' acceptance of cloud gaming is crucial in deter-mining the future of this new platform. A lack of awareness regarding these factors may lead to the users‟ rejecting the new technology.

Cloud Gaming systems are among the most challenging networked-applications, since they deal with streaming high-quality and bulky video in real-time to players’ devices. While all industry solutions today are centralized, we introduce an AI-assisted hybrid networking architecture that, in addition to the central cloud servers, also uses some players’ computing resources as additional points of service. We describe the problem, its mathematical formulation, and potential solution strategy. Edge computing is a promising paradigm that brings servers closer to users, leading to lower latencies and enabling latency-sensitive applications such as cloud gaming, virtual/augmented reality, telepresence, and telecollaboration. Due to the high number of possible edge servers and incoming user requests, the optimum choice of user-server matching has become a difficult challenge, especially in the 5G era where the network can offer very low latencies. In this thesis, we introduce the problem of fair server selection as not only complying with an application's latency threshold but also reducing the variance of the latency among users in the same session. Due to the dynamic and rapidly evolving nature of such an environment and the capacity limitation of the servers, we propose as solution a Reinforcement Learning method in the form of a Quadruple Q-Learning model with action suppression, Q-value normalization, and a reward function that minimizes the variance of the latency. Our evaluations in the context of a cloud gaming application show that, compared to a existing methods, our proposed method not only better meets the application's latency threshold but is also more fair with a reduction of up to 35\% in the standard deviation of the latencies while using the geo-distance, and it shows improvements in fairness up to 18.7\% compared to existing solutions using the RTT delay especially during resource scarcity. Additionally, the RL solution can act as a heuristic algorithm even when it is not fully trained. While designing this solution, we also introduced action suppression, Quadruple Q-Learning, and normalization of the Q-values, leading to a more scalable and implementable RL system. We focus on algorithms for distributed applications and especially esports, but the principles we discuss apply to other domains and applications where fairness can be a crucial aspect to be optimized.

As an increasing position of the world’s communication moves towards the cloud and wireless solutions the requirement for good throughput and low delay increases. One step towards meeting higher requirements is the move from 4G Long Term Evolution (LTE) to 5G New Radio (NR). In order to utilize the potential of 5G NR, software needs to be improved. With the goal to lower the delay for delay critical applications and services when using 5G NR this thesis studies a new scheduler inspired by Earliest Deadline First (EDF) as a soft real-time system scheduler. This new scheduler called Real-Time Inspired Hybrid Scheduler (RTIHS) is proposed where two different schedulers are used depending on if the network traffic is delay critical or not. Delay critical traffic is served by the new Deadline Inspired Scheduler (DIS) scheduler and other traffic by a traditional Round Robin (RR) scheduler. The transmissions that are delay critical are prioritized differently with the use of a constant that has two functions. To determine if the transmission is in time or not, and how close the transmission is to the fixed deadline. Up until the deadline the priority of that transmission is increased with a factor that is affected by how close the current time is to the deadline. If the deadline has been missed the priority is, however, decreased with respect to how much it missed the deadline.RTIHS is implemented and tested in a state-of-the-art system simulator where services such as;Cloud Gaming (CG), Video on Demand (VoD), and web browsing are evaluated. An already existing technology named Low Latency Low Loss Scalable Throughput (L4S) is included in the evaluations to investigate how RTIHS scales. The performance of RTIHS is then compared to a Delay Scheduler(DS) and a RR scheduler that act as the baseline. The results show that RTIHS performs better for CG in terms of delay and nominal rate than the baseline, especially when the network is under high load. Using RTIHS compared to a DS shows an average increase of nominal rate by roughly 6 % and an average decrease of delay by 16 % for the average users. At the same time RTIHS does show higher delay and lower throughput for services such as VoD and web browsing, making its performance for those service worse than the baseline, especially when the network load is high. With the same comparison as before, RTIHS shows an average 17 % higher delay for web browsing and 1 % lower requested video rate than the DS for the average users. This is due to how the evaluated services are prioritized with the limited resources available. Since RTIHS prioritizes CG more, less resources remain for the other services. The baseline is not as biased towards CG and therefore has a lower nominal rate and higher delay for said service but better for the other services in comparison. Activating L4S mitigates the underwhelming performance of RTIHS for VoD and web browsing further improves the performance for CG. This is also true for the baseline which increases performance for CG and further increases the performance for other services with a small amount. With L4S activated the average increase of nominal rate in CG for RTIHS compared to the DS is 13 % for the average users and the average delay decrease is 9 %. Meanwhile the average requested video rate for RTIHS is less than 1 % lower than that of the DS for the average users and the delay for web browsing is 10 % higher than the DS for the average users.

Submitted by Rosa Assis (rosa_assis@yahoo.com.br) on 2017-11-14T18:05:03Z No. of bitstreams: 2 VICTOR PERAZZOLO BARROS.pdf: 24134660 bytes, checksum: 8761000aa9ba093f81a14b3c2368f2b7 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Approved for entry into archive by Paola Damato (repositorio@mackenzie.br) on 2017-11-27T12:14:38Z (GMT) No. of bitstreams: 2 VICTOR PERAZZOLO BARROS.pdf: 24134660 bytes, checksum: 8761000aa9ba093f81a14b3c2368f2b7 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) / Made available in DSpace on 2017-11-27T12:14:38Z (GMT). No. of bitstreams: 2 VICTOR PERAZZOLO BARROS.pdf: 24134660 bytes, checksum: 8761000aa9ba093f81a14b3c2368f2b7 (MD5) license_rdf: 0 bytes, checksum: d41d8cd98f00b204e9800998ecf8427e (MD5) Previous issue date: 2017-02-06 / The advances in Cloud Computing and communication technologies enabled the concept of Cloud Gaming to become a reality. Through PCs, consoles, smartphones, tablets, smart TVs and other devices, people can access and use games via data streaming, regardless the computing power of these devices. The Internet is the fundamental way of communication between the device and the game, which is hosted and processed on an environment known as Cloud. In the Cloud Gaming model, the games are available on demand and offered in large scale to the users. The players' actions and commands are sent to servers that process the information and send the result (reaction) back to the players. The volume of data processed and stored in these Cloud environments exceeds the limits of analysis and manipulation of conventional tools, but these data contains information about the players' profile, its singularities, actions, behavior and patterns that can be valuable when analyzed. For a proper comprehension and understanding of this raw data and to make it interpretable, it is necessary to use appropriate techniques and platforms to manipulate this amount of data. These platforms belong to an ecosystem that involves the concepts of Big Data. The model known as Big Data Analytics is an effective and capable way to, not only work with these data, but understand its meaning, providing inputs for assertive analysis and predictive actions. This study searches to understand how these technologies works and propose a method capable to analyze and identify patterns in players' behavior and characteristics on a virtual environment. By knowing the patterns of different players, it is possible to group and compare information, in order to optimize the user experience, revenue for developers and raise the level of control over the environment in a way that players' actions can be predicted. The results presented are based on different analysis modeling using the Hadoop technology combined with data visualization tools and information from open data sources in a dataset of the World of Warcraft game. Fraud detection, users' game patterns, churn prevention inputs and relations with game attractiveness elements are examples of modeling used. In this research, it was possible to map and identify the players' behavior patterns and create a prediction of its frequency and tendency to evade or stay in the game. / Os avanços das tecnologias de Computacão em Nuvem (Cloud Computing) e comunicações possibilitaram o conceito de Jogos em Nuvem (Cloud Gaming) se tornar uma realidade. Por meio de computadores, consoles, smartphones, tablets, smart TVs e outros equipamentos é possível acessar via streaming e utilizar jogos independentemente da capacidade computacional destes dispositivos. Os jogos são hospedados e executados em um ambiente computacional conhecido como Nuvem, a Internet é o meio de comunicação entre estes dispositivos e o jogo. No modelo conhecido como Cloud Gaming, compreendesse que os jogos são disponibilizados sob demanda para os usuários e podem ser oferecidos em larga escala. Os comandos e ações dos jogadores são enviados para servidores que processam a informação e enviam o resultado (reação) para o jogador. A quantidade de dados que são processados e armazenados nestes ambientes em Nuvem superam os limites de análise e manipulação de plataformas convencionais, porém tais dados contém informacões sobre o perfil dos jogadores, suas particularidades, ações, comportamentos e padrões que podem ser importantes quando analisados. Para uma devida compreensão e lapidação destes dados brutos, a fim de torná-los interpretáveis, se faz necessário o uso de técnicas e plataformas apropriadas para manipulação desta quantidade de dados. Estas plataformas fazem parte de um ecossistema que envolvem os conceitos de Big Data. Arquiteturas e ferramentas de Big Data, mais especificamente, o modelo denominado Big Data Analytics, são instrumentos eficazes e capazes de não somente trabalhar com estes dados, mas entender seu significado, fornecendo insumos para análise assertiva e predição de acões. O presente estudo busca compreender o funcionamento destas tecnologias e fornecer um método capaz de identificar padrões nos comportamentos e características dos jogadores em ambiente virtual. Conhecendo os padrões de diferentes usuários é possível agrupar e comparar as informações, a fim de otimizar a experiência destes usuários no jogo, aumentar a receita para os desenvolvedores e elevar o nível de controle sobre o ambiente ao ponto que seja possível de prever ações futuras dos jogadores. Os resultados obtidos são derivados de diferentes modelagens de análise utilizando a tecnologia Hadoop combinada com ferramentas de visualização de dados e informações de fontes de dados abertas, em um dataset do jogo World of Warcraft. Detecção de fraude, padrões de jogo dos usuários, insumos para prevencão de churn e relações com elementos de atratividade no jogo, são exemplos de modelagens abordadas. Nesta pesquisa foi possível mapear e identificar os padrões de comportamento dos jogadores e criar uma previsão e tendência de assiduidade sobre evasão ou permanencia de usuários no jogo.

Cloud computing is a way to deliver high-performance services to clients who would not usually be able to handle the computations on their own. They rely on computers in the cloud performing the calculations and therefore ease the load on the client-side. The goal of this thesis is to find what factors affect the players' experience and how they affect the player. To handle this, we have done a user-based study on cloud gaming. The users get to play a fast-paced single-player game and a fast-paced multiplayer game against each other, while we collect data about their experiences. During the tests, we manipulated the players' network conditions, and afterward, they answered questions regarding their quality of experience. From the data collected, we can see that the frame age is the most important measurement for determining the players' in-game performance as well as the quality of experience. We are also able to see that from the quality of service measurements manipulated, the latency is the one affecting the player the most. Results from the multiplayer test would indicate that we can equalize the skill difference between the players without affecting the players quality of experience too much. These results are based on the advantage in ping time as well as frame age. From a developers perspective, this thesis emphasized the need to take frame age into account, and to try to manipulate the different parts of the frame age. The goal would be to ultimately lower the frame age and make the gaming experience more enjoyable for the player.

<p>The accessibility and flexibility of mobile devices make them an advantageous platform for gaming, but there are hardware limitations that impede the rendering of high-quality graphics. Rendering complex graphics on a mobile device typically results in a delayed image, also known as latency, and is a great discomfort for users of any real-time rendering experience. This study tests the image stream optimization View Bypass within a cloud gaming architecture, surpassing this imposing limitation by processing the high-quality game render on a remote computational server. A two sample for means test is performed to determine significance between two treatments: the control group without the View Bypass algorithm and the experimental group rendering with the View Bypass algorithm. A SSIM index score is calculated comparing the disparity between the remote server image output and the final mobile device image output after optimizations have been performed. This score indicates the overall image structural integrity difference between the two treatments and determines the quality and effectiveness of the tested algorithm.</p>

VR gaming is an emerging market with relatively new technology when compared to PC and console gaming. Streamed VR gaming is when the VR game is being streamed from another device to the VR headset. This experiment will show how the VR hardware and software performs under network constraints such as RTT and packet loss for streamed VR gaming specifically. User QoE (quality of experience) will be gathered and compiled into MOS (mean opinion score) for later use in graphs to determine negative trends. All collected data will also be sent to the Infovista technical team so they can make theoretical models on how user QoE is affected by network conditions. This thesis will show how the QoS (quality of service) factors, such as latency and packet loss, impact the users overall QoE and the VR headsets FPS. Before the subjective experiment could begin a testbed was chosen and setup, a VR game was decided on, and a streaming service was tested on. During the experiment the game Serious Sam VR: The Last Hope was played and NVIDIA’s CloudXR streaming service was used. The testbed setup consisted of two PCs, one router, one access point, a tablet used for user reviews, and a VR headset with two handheld controllers. The software program ALTRUIST was used for managing the user data and applying the desired network conditions such as latency. Each user experiment tested 28 different network conditions with varying values of RTT (round trip time), Packet loss, RTT with packet loss, and RTT with random jitter. 30 users were gathered for the actual experiment where the Calendly application was used to manage the user appointments. The results consisted of objective and subjective data gathered during the experiments. From the objective data it could be determined that RTT primarily affects the FPS in the VR headset. Packet loss affected the FPS but to a minor degree compared to RTT. RTT of 175 ms received an average FPS of 53 while 24% packet loss yielded approximately 65 FPS. Random jitter had close to no impact on the FPS according to the results. The subjective data suggests that RTT of 75 ms and below yields a good user MOS and RTT of 175 ms gives an acceptable user MOS at 3 out of 5. Packet loss of 6% and below gives a good user MOS while 12% and above results in poor MOS. RTT with random jitter performs average to poorly with Standard deviation of 3 and above, for low RTT values. For future experiments it is advisableto use ALTRUIST or a similar software to help the researcher during the experiment.

In recent years, there has been an increased number of people that are engaged in gaming - a phenomenon connected to constantly developing internet infrastructure and a growing middleclass. The phenomenon of Cloud gaming has received increasing attention since it was introduced some years ago; in 2022 all Cloud gaming services combined had 31,7 million paying users. The number that is estimated to triple by 2025. Given this information, we proceeded to identify a gap in previous literature within the topic of Cloud gaming. Previous research within the topic of Cloud gaming has predominantly discussed Cloud gaming as a business model rather than focusing on consumers. We sought to seize the opportunity to fill this research gap by studying consumers' motivations to engage in this new technology and examine which factors can influence the customer's decision-making process in the path to purchase. We also wanted this study to contribute to marketing practice by examining important current research priorities within the field of marketing. In order to investigate our research purposes, we adopted an exploratory qualitative study where we, with the help of online focus groups, investigated the customers' perceptions of Cloud gaming as a service and their motivations to engage in it. Another purpose of this study was to find out how the gaming industry may be affected by the increased usage of cloud gaming, therefore we wanted to conduct a semi-structured interview with an industry representative. The most interesting result of this study indicates that a somewhat odd situation arises for the Cloud gaming companies where the respondents who spent the least amount of money and engage the least in traditional gaming, showed the greatest motivation to engage in Cloud gaming. Therefore we suggest that there are a number of different measures that must be kept in mind by marketers within the industry.

Cloud gaming has emerged as a popular solution to meet the increasing hardware de-mands of modern video games, allowing players with dated or non-sufficient hardwareto access high-quality gaming experiences. However, the growing reliance on cloud ser-vices has led to heightened concerns regarding user privacy and the risk of fingerprintingattacks. In this paper, we investigate the effects of varying send rates on cloud gamingQoS and QoE metrics when applying a fixed-rate fingerprinting defense, BuFLO. Findingsshow that lower send rates impact both client-side and host-side applied defense differ-ently. Based on the results, specific send rates are suggested for maintaining a stable cloudgaming experience. The research offers insights into the trade-offs between security andperformance in cloud gaming and provides recommendations for mitigating fingerprintingattacks. Future work may investigate alternative defenses, device types, and connectionmethods.