[pt] Este trabalho é um estudo sobre o desempenho da arquitetura Diffserv proposta para permitir oferecimento de serviços com QoS na Internet. A ênfase do estudo é no desempenho de fluxos que utilizam o protocolo de transporte TCP.Inicialmente são revistos conceitos básicos, características do protocolo TCP em suas diferentes versões e os principais mecanismos de controle de tráfego. É feita, então, uma modelagem da arquitetura Diffserv visando a avaliação de seu desempenho. Procurou-se resumir os principais resultados sobre desempenho desta arquitetura disponíveis na Literatura e estabelecer um modelo de referência que permitisse a realização de avaliações específicas. Este modelo de oferecimento de serviços com QoS por um provedor é avaliado através do programa de simulação NS. Através de diversos casos analisados são feitas comparações de desempenho de fluxos TCP em diferentes classes de serviço,levando em conta o tipo de implementação e os mecanismos de encaminhamento de pacotes. / [en] This work presents a Diffserv architecture study to provide QoS in the Internet. The emphasis of the study is on the performance of flows that use the TCP transport protocol. Initially, basic concepts, TCP characteristics in its different versions and the main mechanisms of traffic control are reviewed. A modeling of the Diffserv architecture, aiming at the performance evaluation, is made. The main results on performance of this architecture in the available Literature were summarized and a reference model that allowed the accomplishment of specific evaluations was proposed. This model of QoS offering for a provider is evaluated through the NS simulation program. Through several analyzed cases, comparisons performance of TCP flows are made in different classes of service, talking into account the type of TCP implementation and the mechanisms of packet forwarding.

Fundamental properties of cloud computing such as resource sharing and on-demand self-servicing is driving a growing adoption of the cloud for hosting both legacy and new application services. A consequence of this growth is that the increasing scale and complexity of the underlying cloud infrastructure as well as the fluctuating service workloads is inducing performance incidents at a higher frequency than ever before with far-reaching impact on revenue, reliability, and reputation. Hence, effectively managing performance incidents with emphasis on timely detection, diagnosis and resolution has thus become a necessity rather than luxury. While other aspects of cloud management such as monitoring and resource management are experiencing greater automation, automated management of performance incidents remains a major concern. Given the volume of operational data produced by cloud datacenters and services, this thesis focus on how data analytics techniques can be used in the aspect of cloud performance management. In particular, this work investigates techniques and models for automated performance anomaly detection and prevention in cloud environments. To familiarize with developments in the research area, we present the outcome of an extensive survey of existing research contributions addressing various aspects of performance problem management in diverse systems domains. We discuss the design and evaluation of analytics models and algorithms for detecting performance anomalies in real-time behaviour of cloud datacenter resources and hosted services at different resolutions. We also discuss the design of a semi-supervised machine learning approach for mitigating performance degradation by actively driving quality of service from undesirable states to a desired target state via incremental capacity optimization. The research methods used in this thesis include experiments on real virtualized testbeds to evaluate aspects of proposed techniques while other aspects are evaluated using performance traces from real-world datacenters. Insights and outcomes from this thesis can be used by both cloud and service operators to enhance the automation of performance problem detection, diagnosis and resolution. They also have the potential to spur further research in the area while being applicable in related domains such as Internet of Things (IoT), industrial sensors as well as in edge and mobile clouds. / Grundläggande egenskaper för datormoln såsom resursdelning och självbetjäning driver ett växande nyttjande av molnet för internettjänster. En följd av denna tillväxt är att den underliggande molninfrastrukturens ökande storlek och komplexitet samt fluktuerade arbetsbelastning orsakar prestandaincidenter med högre frekvens än någonsin tidigare. En konsekvens av detta blir omfattande inverkan på intäkter, tillförlitlighet och rykte för de som äger tjänsterna. Det har därför blivit viktigt att snabbt och effektivt hantera prestandaincidenter med avseende på upptäckt, diagnos och korrigering. Även om andra aspekter av resurshantering för datormoln, som övervakning och resursallokering, på senare tid automatiserats i allt högre grad så är automatiserad hantering av prestandaincidenter fortfarande ett stort problem. Denna avhandling fokuserar på hur prestandahanteringen i molndatacenter kan förbättras genom användning av dataanalystekniker på de stora datamängder som produceras i de system som monitorerar prestanda hos datorresurser och tjänster. I synnerhet undersöks tekniker och modeller för automatisk upptäckt och förebyggande av prestandaanomalier i datormoln. För att kartlägga utvecklingen inom forskningsområdet presenterar vi resultatet av en omfattande undersökning av befintliga forskningsbidrag som behandlar olika aspekter av hantering av prestandaproblem inom i relevanta tillämpningsområden. Vi diskuterar design och utvärdering av analysmodeller och algoritmer för att upptäcka prestandaanomalier i realtid hos resurser och tjänster. Vi diskuterar också utformningen av ett maskininlärningsbaserat tillvägagångssätt för att mildra prestandaförluster genom att aktivt driva tjänsternas kvalitet från oönskade tillstånd till ett önskat målläge genom inkrementell kapacitetoptimering. Forskningsmetoderna som används i denna avhandling innefattar experiment på verkliga virtualiserade testmiljöer för att utvärdera aspekter av föreslagna tekniker medan andra aspekter utvärderas med hjälp av belastningsmönster från verkliga datacenter. Insikter och resultat från denna avhandling kan användas av både moln- och tjänsteoperatörer för att bättre automatisera detekteringen av prestandaproblem, inklusive dess diagnos och korrigering. Resultaten har också potential att uppmuntra vidare forskning inom området samtidigt som de är användbara inom relaterade områden som internet-av-saker, industriella sensorer, och storskaligt distribuerade moln eller telekomnätverk. / Cloud Control / eSSENCE

La Qualité de Service (Quality of Service - QoS) est une garantie de paramètres réseau (bande passante élevée, délai court, etc.). Dans un réseaux inter-domaine, elle peut être assurée par des contrats entre domaines appelés Service Level Agreements (SLA). Dans cette thèse, nous nous intéressons d’abord à l’étape de négociation de SLA : la sélection des SLA proposés par un domaine. Nous proposons des méthodes exactes et approchées permettant aux domaines de proposer les SLA qui maximisent leurs revenus. Nous étudions également l'impact de la réputation des domaines sur cette négociation. Au niveau de l’instanciation des SLA, nous nous intéressons au calcul de chemins qui prennent en compte les encapsulations de protocoles (afin de pallier l’hétérogénéité technologique des domaines). En utilisant des outils de théorie des langages, nous proposons la première solution polynomiale au calcul de chemins dans un tel contexte / Quality of Service (QoS) is a guarantee of network parameters (high bandwidth, short delay, etc.). In an inter-domaine network, it can be provided by contracts between domains, called Service Level Agreements (SLAs). In this thesis, we first focus on the SLA negotiation step, i.e., the selection of SLAs proposed by a domain. We provide exact and approximate methods that allow the domains to propose the SLAs that maximize their revenues. We also study the impact of the domain reputation on the negotiation process. Regarding the SLA instantiation step, we focus on path computation processes that take into account the encapsulation of protocols (in order to overcome the domain technological heterogeneity). Using tools from Language Theory, we provide the first polynomial solution to the path computation problem in this context.

Orientador: Lee Luan Ling / Tese (doutorado) - Universidade Estadual de Campinas, Faculdade de Engenharia Eletrica e de Computação / Made available in DSpace on 2018-08-08T01:26:33Z (GMT). No. of bitstreams: 1 Vieira_FlavioHenriqueTeles_D.pdf: 4214611 bytes, checksum: 755dfe9865aff1214f8e551afde7541d (MD5) Previous issue date: 2006 / Resumo: A modelagem multifractal generaliza os modelos de tráfego existentes na literatura e se mostra apropriada para descrever as características encontradas nos fluxos de tráfego das redes atuais. A presente tese investiga abordagens para alocação de banda, predição de tráfego e estimação de probabilidade de perda de bytes considerando as características multifractais de tráfego. Primeiramente, um Modelo Multifractal baseado em Wavelets (MMW) é proposto. Levando em consideração as propriedades deste modelo, são derivados o parâmetro de escala global, a função de autocorrelação e a banda efetiva para processos multifractais. A capacidade de atualização em tempo real do MMW aliada à banda efetiva proposta permite o desenvolvimento de um algoritmo de estimação adaptativa de banda efetiva. Através deste algoritmo é introduzido um esquema de provisão adaptativo de banda efetiva. Estuda-se também a alocação de banda baseada em predição de tráfego. Para este fim, propõe-se um preditor adaptativo fuzzy de tráfego, o qual é aplicado em uma nova estratégia de alocação de banda. O preditor fuzzy adaptativo proposto utiliza funções de base ortonormais baseadas nas propriedades do MMW. Com relação à probabilidade de perda para tráfego multifractal, derivase uma expressão analítica para a estimação da probabilidade de perda de bytes considerando que o tráfego obedece ao MMW. Além disso, uma caracterização mais completa do comportamento de fila é efetuada pela obtenção de limitantes para a probabilidade de perda e para a ocupação média do buffer em termos da banda efetiva do MMW. Por fim, é apresentado um esquema de controle de admissão usando o envelope efetivo proposto para o MMW oriundo do cálculo de rede estatístico, que garante que os fluxos admitidos obedeçam simultaneamente aos requisitos de perda e de retardo. As simulações realizadas evidenciam a relevância das propostas apresentadas / Abstract: Multifractal modeling generalizes the existing traffic models and is believed to be appropriate to describe the characteristics of traffic flows of modern communication networks. This thesis investigates some novel approaches for bandwidth allocation, traffic prediction and byte loss probability estimation, by considering the multifractal characteristics of the network traffic. Firstly, a Wavelet based Multifractal Model (WMM) is proposed. Taking into account the properties of this multifractal model, we derive the global scaling parameter, the autocorrelation function and the effective bandwidth for multifractal processes. The real time updating capacity of the WMM in connection with our effective bandwidth proposal allows us to develop an algorithm for adaptive effective bandwidth estimation. Then, through this algorithm, an adaptive bandwidth provisioning scheme is introduced. In this work, we also study a prediction-based bandwidth allocation case. For this end, we develop an adaptive fuzzy predictor, which is incorporated into a novel bandwidth allocation scheme. The proposed adaptive fuzzy predictor makes use of orthonormal basis functions based on the properties of the WMM. Additionally, we derive an analytical expression for the byte loss probability estimation assuming that the traffic obeys the MMW. Besides, a more complete characterization of the queuing behavior is carried out through the estimation of the bounds for the loss probability and mean queue length in buffer in terms of the WMM based effective bandwidth. Finally, an admission control scheme is presented that uses the WMM based effective envelope derived through the statistical network calculus, guaranteeing that the admitted flows simultaneously attend the loss and delay requirements. The computer simulation results confirm the relevance of the presented proposals / Doutorado / Telecomunicações e Telemática / Doutor em Engenharia Elétrica

Este trabalho de mestrado apresenta a modelagem, a prototipação e os resultados do desenvolvimento de um middleware para composição dinâmica de Web Services denominado DWSC-M (Dynamic Web service Composition Middleware). O objetivo principal desse middleware é permitir que serviços sejam compostos dinamicamente considerando aspectos de qualidade de serviço na escolha dos serviços que fazem parte de um fluxo de composição. Para complementar o funcionamento do middleware DWSC-M foram propostos dois algoritmos para seleção de Web Services: o primeiro utiliza seleção aleatória e o segundo utiliza distância Euclidiana para seleção de serviços e considera, para tal finalidade, os parâmetros de QoS enviados pela requisição do cliente do serviço / This work presents the modeling, the prototype and the results of the developing of a middleware for dynamic composition of Web Services named DWSC-M (Dynamic Web Service Composition Middleware). The main focus of this middleware is to allow services being dynamically composed, considering aspects of quality of service in the choice of services that are part of the composite flow. Complementing the operation of the DWSC-M middleware two algorithms were proposed for Web Services selection : the first one uses Random selection and the second one uses Euclidean Distance for the selection of services and considers for this purpose the QoS parameters sent by the client service request

Este trabalho propõe políticas para servidores Web baseados em sessões, visando ao oferecimento de garantias de qualidade de serviço (Quality of Service - QoS) a seus usuários. Para o fornecimento de QoS, dois tipos de políticas são consideradas: as responsáveis pela diferenciação de serviços e as responsáveis pelo controle de admissão. Alguns algoritmos de escalonamento foram desenvolvidos com o objetivo de oferecer melhores serviços para as sessões prioritárias: o algoritmo de escalonamento baseado em sessão (SBSA) e o algoritmo de escalonamento baseado em sessão com diferenciação de serviços (SBSA-DS). Quanto ao controle de admissão, foram consideradas duas novas políticas, as quais utilizam diferentes parâmetros para a tomada de decisão e têm como objetivo garantir a finalização do maior número de sessões prioritárias. A partir dos resultados obtidos nos experimentos realizados por meio de simulação, constatou-se que o emprego dos algoritmos e das políticas desenvolvidos melhora o atendimento das sessões para as quais deve ser oferecido um serviço preferencial. Portanto, em servidores Web preocupados com a garantia de QoS para seus usuários, os quais interagem com o sistema por meio de sessões, o uso dessas novas técnicas se mostra adequado / This work considers policies for Web servers based on sessions, aiming at the guarantees of Quality of Service (QoS) to the systems users. Two types of policies are considered: the responsible for the differentiation of services and the responsible for the admission control. Some scheduling algorithms had been developed to offer better services for the priority sessions: the Session Based Scheduling Algorithm (SBSA) and the Session Based Scheduling Algorithm with Differentiation of Services (SBSA-DS). In relation to the admission control, two new policies had been considered, which use different parameters to help the decisions and have as objective guarantee the finalization of as many priority sessions as possible. From the obtained results of the experiments realized by mean of simulation, it is evidenced that the use of the developed algorithms and policies improves the attendance of the sessions that deserve preferential service. Therefore, in Web servers worried in the guarantee the QoS for their users, which interact with the system by sessions, the use of these new techniques reveals adequate

Esta dissertação de mestrado apresenta o projeto, implementação e teste de um gerador de carga de trabalho para servidores Web que permite a caracterização de sessões e que representa uma carga real. O gerador desenvolvido gera carga visando à alimentar programas que simulam o comportamento de servidores Web. Para a definição dos parâmetros utilizados no gerador de carga sintética desenvolvido, diversos estudos sobre a carga em servidores Web encontrados na literatura, logs de servidores Web reais e benchmarks para servidores Web, foram utilizados. Vários algoritmos de controle de admissão e de escalonamento de requisições descritos na literatura foram implementados para teste e validação do gerador desenvolvido. Visando a utilização de sessões, é proposto nesta dissertação um algoritmo de controle de admissão, RED-SB (Random Early Detection - Session Based) que se baseia em um algoritmo (RED) que não considera sessão. Esses algoritmos foram executados utilizando-se um modelo de simulação de servidor Web com suporte a diferenciação de serviços, chamado SWDS (Servidor Web com Diferenciação de Serviços) / This MSc dissertation presents the design, implementation and tests of a workload generator for Web servers that allows the characterization of sessions and represents a real workload. The generator developed generates a workload to be used as an input to programs that simulate the behavior of Web servers. For the definition of the parameters used in the synthetic workload generator developed, several studies about the web server load described in the literature, actual web server logs and benchmarks for Web servers were used. Several algorithms for admission control and scheduling of requests described in the literature were implemented for testing and validation of the generator developed. Aiming at the utilization of sessions, this dissertation proposes an admission control algorithm RED-SB (Random Early Detection - Session based) which is based on an algorithm (RED) that does not consider session. These algorithms were implemented using a simulation model ofWeb server that supports service differentiation, called SWDS (Web Server with Service Differentiation)

Dans les domaines critiques tels que l’avionique, le ferroviaire ou encore l’automobile, il faut, afin de pouvoir certifier un système, démontrer qu’il réalise la fonction pour laquelle il a été conçu, selon des exigences temporelles spécifiées. En effet, un rendu temporel trop long peut rendre des données erronées, et ainsi mettre en danger la sûreté des personnes. Aujourd’hui, la plupart des approches proposent d’assurer ces exigences de Qualité de service au niveau des couches basses, e.g., au travers d’une bande passante déterministe, d’allocation statique d’intervalles de temps, et d’un ordonnancement prédéfini. Ces contraintes assurent que les applications ne peuvent dépasser le temps d’exécution alloué ; les applications récupèrent de ce fait des exigences qui sont découplées de leur fonctionnalité. En revanche, il faut aussi pouvoir certifier des exigences temporelles spécifiques à une application. De là, les garanties au niveau des couches basses ne sont plus suffisantes. Il faudrait pouvoir prendre en compte ces exigences dès la phase de conception des applications. Aujourd’hui, la plupart des approches existant dans ce domaine se concentrent sur le support de QoS à des phases isolées du processus de développement logiciel, empêchant la traçabilité des exigences. Cette thèse propose une approche dirigée par la conception pour supporter les exigences de QoS tout au long du processus de développement logiciel, intégrée dans une méthodologie outillée, appelée DiaSuite. L’extension de QoS enrichit le langage de conception DiaSpec avec la capacité d’instancier les exigences de QoS sur les composants logiciels. Un support de surveillance à l’exécution de ces exigences temporelles est ensuite généré, directement à partir de la spécification. Cette thèse intègre uniformément les concepts temporels avec les concepts de gestion d’erreurs, au travers de la méthodologie DiaSuite, afin de proposer une couche de supervision qui puisse effectuer une reconfiguration applicative, dans le cas de violation de contrat de QoS. Les contributions de cette thèse sont évaluées au regard du respect des critères de cohérence et de conformité, illustrés au travers d’une étude de cas dans le domaine avionique. / In critical domains such as avionics, railways or automotive, to certify a system, it is required to demonstrate that it achieves its function, with respect to specified timing requirements. Indeed, longer-than-predicted function computing can make data erroneous, leading potentially to endanger people lives. Today, most approaches propose to ensure these Quality of Service requirements at platform level, e.g., through deterministic bandwidth, static time slots allocation and predefined scheduling. These constraints ensure applications can’t overpass allocated time slots; applications are then fed with requirements decoupled to their functionality. However, it shall be possible to certify timing requirements, dedicated to an application. Hence, guarantees at platform-level are not sufficient anymore. It should be possible to take into account these requirements from the stage of application design. Today, most of existing approaches in this domain, focus on supporting QoS at individual stages of the software development process, preventing requirements traceability. This thesis proposes a design-driven approach to supporting QoS throughout software development process, integrated in a tool-based methodology, namely DiaSuite. The QoS extension enriches the DiaSpec design language, with the capability to instantiate QoS requirements onto software components. A runtime execution support to monitoring these timing requirements, is then generated, directly from the specification. This thesis uniformly integrates timing concepts with error ones, around DiaSuite methodology, to propose a supervision layer that could lead to application reconfiguration in case of QoS contract violation. Contributions of this thesis are evaluated through respect of coherence and conformance critera, illustrated through a case study in avionics.

Diplomová práca sa zaoberá vývojom emulátoru prenosových parametrov dátovych sietí nad protokolovou sadou TCP/IP. Práca opisuje špecifiká sieťovej časti Linuxového jadra ako aj realizáciu emulátoru pomocou userspace programov ako tc, netem, ifb a iptables ovládajúcich jadro. Program pre emuláciu je realizovaný v jazyku Python a je ovládaný cez grafické rozhranie vytvorené s PyQt5. Aplikácia bola otestovaná a verifikovaná, že je dostatočne presnou reprezentáciou špecifikovanej WAN siete a že taktiež využíva prostriedky efektívne.

The main object of this thesis was to design a computer network for small and medium companies which will be made among others from network server providing simple network control. The next task was to provide internet connection to subscribers of local area network, their security and access to common services. There had been discussed the choice of distribution of Linux operation system with regarding to demanded services, stability and long lasting operation in theoretical part. One part of the work is a theoretic preliminary to separate services, to the purpose of their using and to their weaknesses. The practical part deals with an installation and configuration of Debian operating system, launching the base services and the setting of selected monitoring programs.