Task scheduling in operating system area is well known problem on traditional system architectures (NUMA, SMP). However, it does not perform well on emerging many-core architectures, especially on Intel Xeon Phi. We have collected all publicly available information about the Xeon Phi's architecture. After that we have benchmarked the Xeon Phi to find missing information about its architecture. We were especially curious in architecture of cores and memory controllers. These parts are most important while designing scheduler. Based on the measured results we have proposed improvements to scheduling algorithm in the Bobox - experimental streaming system. Powered by TCPDF (www.tcpdf.org)

Task scheduling in operating system area is a well-known problem on traditional system architectures (NUMA, SMP). Unfortunately, it does not perform well on emerging many-core architectures, especially on Intel Xeon Phi. We collected all publicly available information about the architecture of Xeon Phi. After that, we benchmarked the Xeon Phi in order to find the missing information about its architecture. We focused especially on the information about cores and memory controllers. These are the most important parts when designing a scheduler. Based on the results, we proposed improvements for scheduling algorithm in the Bobox (an experimental streaming system). However, we found that the biggest problem is not in the scheduling algorithm, but in the design of operators' parallelization. Therefore, we proposed improvements to the parallelization and tested one of the proposals.

Peer-to-peer (P2P) streaming has recently received much research attention, with successful commercial systems showing its viability in the Internet. Despite the remarkable popularity in real-world systems, the fundamental properties and limitations are not yet well understood from a theoretical perspective, as there exists a significant gap between the fundamental limits and the performance achieved in practice. In this thesis, we seek to provide an in-depth analytical understanding of fundamental properties and limitations of P2P streaming systems, with a particular spotlight on the performance gap. We first identify the major problem in existing streaming protocols and show that this problem accounts for most of the gap separating the actual and optimal performances of the streaming systems. We then propose a remedy based on network coding to address this problem and show that the gap to the fundamental limits can be significantly reduced.

Peer-to-peer (P2P) streaming has recently received much research attention, with successful commercial systems showing its viability in the Internet. Despite the remarkable popularity in real-world systems, the fundamental properties and limitations are not yet well understood from a theoretical perspective, as there exists a significant gap between the fundamental limits and the performance achieved in practice. In this thesis, we seek to provide an in-depth analytical understanding of fundamental properties and limitations of P2P streaming systems, with a particular spotlight on the performance gap. We first identify the major problem in existing streaming protocols and show that this problem accounts for most of the gap separating the actual and optimal performances of the streaming systems. We then propose a remedy based on network coding to address this problem and show that the gap to the fundamental limits can be significantly reduced.

<div>The proliferation of GPS-enabled devices has led to the development of numerous location-based services. These services need to process massive amounts of spatial data in real-time with high-throughput and low response time. The current scale of spatial data cannot be handled using centralized systems. This has led to the development of distributed spatial streaming systems. The performance of distributed streaming systems relies on how even the workload is distributed among their machines. However, the real-time streamed spatial data and query follow non-uniform spatial distributions that are continuously changing over time. Therefore, Distributed spatial streaming systems need to track the changes in the distribution of spatial data and queries and redistribute their workload accordingly. This thesis addresses the challenges of adapting to workload changes in distributed spatial streaming systems to improve the performance while preserving the system's security. </div><div>The thesis proposes TrioStat, an online workload estimation technique that relies on a probabilistic model for estimating the cost of partitions and machines of distributed spatial streaming systems. TrioStat has a decentralised technique to collect and maintain the required statistics in real-time with minimal overhead. In addition, this thesis introduces SWARM, a light-weight adaptive load-balancing protocol that continuously monitors the data and query workloads across the distributed processes of spatial data streaming systems, and redistribute the workloads soon as performance bottlenecks get detected. SWARM uses TrioStat to estimate the workload of the system's machines. Although using adaptive load-balancing techniques significantly improves the performance of distributed streaming systems, they make the system vulnerable to attacks. In this thesis, we introduce a novel attack model that targets adaptive load-balancing mechanisms of distributed streaming systems. The attack reduces the throughput and the availability of the system by making it stay in a continuous state of rebalancing. The thesis proposes Guard, a component that detects and blocks attacks that target the adaptive load balancing of distributed streaming systems. Guard is deployed in SWARM to develop an attack-resilient adaptive load balancing mechanism for Distributed spatial streaming systems.<br></div>

A Internet tem se tornado uma plataforma importante para interação e compartilhamento de arquivos, o que motivou uma crescente demanda por serviços eficientes. Sistemas de distribuição de conteúdo (CDS) precisaram ser criados visando modernidade e robustez. No contexto desta tese, CDS são definidos como sistemas usados para compartilhar qualquer tipo de conteúdo na Internet. Duas categorias de CDS se destacam como as mais populares: compartilhamento de arquivos e sistemas de mídia contínua. Arquiteturas par-a-par (P2P) surgiram como potenciais soluções para o aprimoramento da disseminação de conteúdo nos CDS. Nesse contexto, a popularização das arquiteturas P2P motivou a comunidade científica a investigar alguns aspectos de pesquisa desafiadores, e.g., otimização de topologias de redes, mecanismos de inicialização de sistemas e serviços de descoberta de recursos. Um desafio com interesse especial a esta tese diz respeito a mecanismos para conciliar a preferência dos usuários aos conteúdos publicados. Esse aspecto é importante para garantir uma boa qualidade de experiência (QoE) aos usuários dos sistemas, uma vez que podem existir divergências entre opiniões na descrição dos conteúdos e ações maliciosas. Esforços de pesquisa constantes têm sido feitos para combater poluição de conteúdo em CDS. Abordagens buscam construir uma base de conhecimento sobre poluidores e conteúdos poluídos para identificar e isolar conteúdos suspeitos depois que eles são publicados. Entretanto, o tempo de reação dessas abordagens até considerar um conteúdo poluído é consideravelmente longo, permitindo uma ampla disseminação de poluição. Além disso, algumas abordagens anteriores buscam polarizar conteúdos entre poluídos ou não, desconsiderando a intrínseca subjetividade acerca da classificação dos conteúdos compartilhados. O objetivo principal desta tese é propor um mecanismo para prover uma boa QoE aos usuários – agindo proativamente durante as fases iniciais da publicação dos conteúdos – e reduzir os efeitos de interferências maliciosas. Para alcançar tal objetivo, três passos principais guiaram o trabalho de pesquisa apresentado nesta tese. Primeiro, propusemos uma estratégia inovadora que opera de forma conservadora para conter a disseminação de poluição. Segundo, estendemos nossa solução para lidar com a subjetividade acerca das descrições dos conteúdos. Terceiro, tratamos o ataque de poluição como um ataque massivo. Para avaliar a solução, experimentos foram executados utilizando testes reais e simulações. Resultados ressaltaram a importância de adotar medidas de segurança para combater comportamentosmaliciosos em CDS. Na ausência de mecanismos de contramedida, pequenas proporções (10%) de atacantes foram capazes de comprometer o sistema. A instanciação da estratégia conservadora proposta nesta tese demonstrou a eficácia em atrasar usuários para contornar ataques massivos. / The Internet has become a large platform where users can interact and share personal files or third-party productions. Considering the increasing demand for efficient content sharing, modern and robust content distribution systems (CDS) need to be deployed and maintained. In the context of this thesis, CDS are defined as systems used for sharing any kind of content on the Internet. Two categories of CDS are underscored as the most popular ones: file sharing and streaming systems. Peer-to-peer (P2P) architectures have emerged as a potential solution to improve content dissemination in CDS. The popularization of P2P architectures, in the context of CDS,motivated the scientific community to investigate some challenging problems, namely network topology optimization, bootstrap mechanisms, and service discovery. One particular interesting challenge, in the context of this thesis, is related to mechanisms to approximate users to their personal interests. This is important to guarantee good quality of experience (QoE) to users when searching for content. Imprecise descriptions are likely to happen due to different users’ opinion or malicious behavior. Substantial research has been carried out to fight content pollution in CDS. Proposed approaches try to identify and isolate suspicious content after publication. The rationale is to build a base of knowledge about polluters and fake content. However, the reaction time until a content is considered polluted is considerably long, which allows pollution to get widely disseminated. Furthermore, some previous approaches attempt to polarize contents in either polluted or not, not taking into account the inherent subjectivity behind the evaluation of shared contents. The main objective of this thesis is to devise a mechanism to provide users a good QoE – by acting proactively in the early stages of content distribution life cycle – and reduce the effect of malicious interferences. To achieve that, three main steps guided the research work presented in this thesis. First, we proposed a novel strategy that operates conservatively to avoid wide pollution dissemination. Second, we extended our previous solution to cope with the subjectivity regarding content descriptions. Third, and last, we address the pollution attack as a massive attack. To evaluate our solution, a set of experiments was carried out using both real tests and simulations. Results showed the importance of adopting security measures to mitigate malicious behavior in CDS. In the absence of countermeasure mechanisms, even a small proportion (10%) of attackers was able to subvert the system. The introduction of a conservative strategy in this thesis demonstrated the efficacy of delaying users in circumventing massive attacks.

A Internet tem se tornado uma plataforma importante para interação e compartilhamento de arquivos, o que motivou uma crescente demanda por serviços eficientes. Sistemas de distribuição de conteúdo (CDS) precisaram ser criados visando modernidade e robustez. No contexto desta tese, CDS são definidos como sistemas usados para compartilhar qualquer tipo de conteúdo na Internet. Duas categorias de CDS se destacam como as mais populares: compartilhamento de arquivos e sistemas de mídia contínua. Arquiteturas par-a-par (P2P) surgiram como potenciais soluções para o aprimoramento da disseminação de conteúdo nos CDS. Nesse contexto, a popularização das arquiteturas P2P motivou a comunidade científica a investigar alguns aspectos de pesquisa desafiadores, e.g., otimização de topologias de redes, mecanismos de inicialização de sistemas e serviços de descoberta de recursos. Um desafio com interesse especial a esta tese diz respeito a mecanismos para conciliar a preferência dos usuários aos conteúdos publicados. Esse aspecto é importante para garantir uma boa qualidade de experiência (QoE) aos usuários dos sistemas, uma vez que podem existir divergências entre opiniões na descrição dos conteúdos e ações maliciosas. Esforços de pesquisa constantes têm sido feitos para combater poluição de conteúdo em CDS. Abordagens buscam construir uma base de conhecimento sobre poluidores e conteúdos poluídos para identificar e isolar conteúdos suspeitos depois que eles são publicados. Entretanto, o tempo de reação dessas abordagens até considerar um conteúdo poluído é consideravelmente longo, permitindo uma ampla disseminação de poluição. Além disso, algumas abordagens anteriores buscam polarizar conteúdos entre poluídos ou não, desconsiderando a intrínseca subjetividade acerca da classificação dos conteúdos compartilhados. O objetivo principal desta tese é propor um mecanismo para prover uma boa QoE aos usuários – agindo proativamente durante as fases iniciais da publicação dos conteúdos – e reduzir os efeitos de interferências maliciosas. Para alcançar tal objetivo, três passos principais guiaram o trabalho de pesquisa apresentado nesta tese. Primeiro, propusemos uma estratégia inovadora que opera de forma conservadora para conter a disseminação de poluição. Segundo, estendemos nossa solução para lidar com a subjetividade acerca das descrições dos conteúdos. Terceiro, tratamos o ataque de poluição como um ataque massivo. Para avaliar a solução, experimentos foram executados utilizando testes reais e simulações. Resultados ressaltaram a importância de adotar medidas de segurança para combater comportamentosmaliciosos em CDS. Na ausência de mecanismos de contramedida, pequenas proporções (10%) de atacantes foram capazes de comprometer o sistema. A instanciação da estratégia conservadora proposta nesta tese demonstrou a eficácia em atrasar usuários para contornar ataques massivos. / The Internet has become a large platform where users can interact and share personal files or third-party productions. Considering the increasing demand for efficient content sharing, modern and robust content distribution systems (CDS) need to be deployed and maintained. In the context of this thesis, CDS are defined as systems used for sharing any kind of content on the Internet. Two categories of CDS are underscored as the most popular ones: file sharing and streaming systems. Peer-to-peer (P2P) architectures have emerged as a potential solution to improve content dissemination in CDS. The popularization of P2P architectures, in the context of CDS,motivated the scientific community to investigate some challenging problems, namely network topology optimization, bootstrap mechanisms, and service discovery. One particular interesting challenge, in the context of this thesis, is related to mechanisms to approximate users to their personal interests. This is important to guarantee good quality of experience (QoE) to users when searching for content. Imprecise descriptions are likely to happen due to different users’ opinion or malicious behavior. Substantial research has been carried out to fight content pollution in CDS. Proposed approaches try to identify and isolate suspicious content after publication. The rationale is to build a base of knowledge about polluters and fake content. However, the reaction time until a content is considered polluted is considerably long, which allows pollution to get widely disseminated. Furthermore, some previous approaches attempt to polarize contents in either polluted or not, not taking into account the inherent subjectivity behind the evaluation of shared contents. The main objective of this thesis is to devise a mechanism to provide users a good QoE – by acting proactively in the early stages of content distribution life cycle – and reduce the effect of malicious interferences. To achieve that, three main steps guided the research work presented in this thesis. First, we proposed a novel strategy that operates conservatively to avoid wide pollution dissemination. Second, we extended our previous solution to cope with the subjectivity regarding content descriptions. Third, and last, we address the pollution attack as a massive attack. To evaluate our solution, a set of experiments was carried out using both real tests and simulations. Results showed the importance of adopting security measures to mitigate malicious behavior in CDS. In the absence of countermeasure mechanisms, even a small proportion (10%) of attackers was able to subvert the system. The introduction of a conservative strategy in this thesis demonstrated the efficacy of delaying users in circumventing massive attacks.

A Internet tem se tornado uma plataforma importante para interação e compartilhamento de arquivos, o que motivou uma crescente demanda por serviços eficientes. Sistemas de distribuição de conteúdo (CDS) precisaram ser criados visando modernidade e robustez. No contexto desta tese, CDS são definidos como sistemas usados para compartilhar qualquer tipo de conteúdo na Internet. Duas categorias de CDS se destacam como as mais populares: compartilhamento de arquivos e sistemas de mídia contínua. Arquiteturas par-a-par (P2P) surgiram como potenciais soluções para o aprimoramento da disseminação de conteúdo nos CDS. Nesse contexto, a popularização das arquiteturas P2P motivou a comunidade científica a investigar alguns aspectos de pesquisa desafiadores, e.g., otimização de topologias de redes, mecanismos de inicialização de sistemas e serviços de descoberta de recursos. Um desafio com interesse especial a esta tese diz respeito a mecanismos para conciliar a preferência dos usuários aos conteúdos publicados. Esse aspecto é importante para garantir uma boa qualidade de experiência (QoE) aos usuários dos sistemas, uma vez que podem existir divergências entre opiniões na descrição dos conteúdos e ações maliciosas. Esforços de pesquisa constantes têm sido feitos para combater poluição de conteúdo em CDS. Abordagens buscam construir uma base de conhecimento sobre poluidores e conteúdos poluídos para identificar e isolar conteúdos suspeitos depois que eles são publicados. Entretanto, o tempo de reação dessas abordagens até considerar um conteúdo poluído é consideravelmente longo, permitindo uma ampla disseminação de poluição. Além disso, algumas abordagens anteriores buscam polarizar conteúdos entre poluídos ou não, desconsiderando a intrínseca subjetividade acerca da classificação dos conteúdos compartilhados. O objetivo principal desta tese é propor um mecanismo para prover uma boa QoE aos usuários – agindo proativamente durante as fases iniciais da publicação dos conteúdos – e reduzir os efeitos de interferências maliciosas. Para alcançar tal objetivo, três passos principais guiaram o trabalho de pesquisa apresentado nesta tese. Primeiro, propusemos uma estratégia inovadora que opera de forma conservadora para conter a disseminação de poluição. Segundo, estendemos nossa solução para lidar com a subjetividade acerca das descrições dos conteúdos. Terceiro, tratamos o ataque de poluição como um ataque massivo. Para avaliar a solução, experimentos foram executados utilizando testes reais e simulações. Resultados ressaltaram a importância de adotar medidas de segurança para combater comportamentosmaliciosos em CDS. Na ausência de mecanismos de contramedida, pequenas proporções (10%) de atacantes foram capazes de comprometer o sistema. A instanciação da estratégia conservadora proposta nesta tese demonstrou a eficácia em atrasar usuários para contornar ataques massivos. / The Internet has become a large platform where users can interact and share personal files or third-party productions. Considering the increasing demand for efficient content sharing, modern and robust content distribution systems (CDS) need to be deployed and maintained. In the context of this thesis, CDS are defined as systems used for sharing any kind of content on the Internet. Two categories of CDS are underscored as the most popular ones: file sharing and streaming systems. Peer-to-peer (P2P) architectures have emerged as a potential solution to improve content dissemination in CDS. The popularization of P2P architectures, in the context of CDS,motivated the scientific community to investigate some challenging problems, namely network topology optimization, bootstrap mechanisms, and service discovery. One particular interesting challenge, in the context of this thesis, is related to mechanisms to approximate users to their personal interests. This is important to guarantee good quality of experience (QoE) to users when searching for content. Imprecise descriptions are likely to happen due to different users’ opinion or malicious behavior. Substantial research has been carried out to fight content pollution in CDS. Proposed approaches try to identify and isolate suspicious content after publication. The rationale is to build a base of knowledge about polluters and fake content. However, the reaction time until a content is considered polluted is considerably long, which allows pollution to get widely disseminated. Furthermore, some previous approaches attempt to polarize contents in either polluted or not, not taking into account the inherent subjectivity behind the evaluation of shared contents. The main objective of this thesis is to devise a mechanism to provide users a good QoE – by acting proactively in the early stages of content distribution life cycle – and reduce the effect of malicious interferences. To achieve that, three main steps guided the research work presented in this thesis. First, we proposed a novel strategy that operates conservatively to avoid wide pollution dissemination. Second, we extended our previous solution to cope with the subjectivity regarding content descriptions. Third, and last, we address the pollution attack as a massive attack. To evaluate our solution, a set of experiments was carried out using both real tests and simulations. Results showed the importance of adopting security measures to mitigate malicious behavior in CDS. In the absence of countermeasure mechanisms, even a small proportion (10%) of attackers was able to subvert the system. The introduction of a conservative strategy in this thesis demonstrated the efficacy of delaying users in circumventing massive attacks.