Este trabalho de pesquisa apresenta um protótipo de servidor Web distribuído com diferenciação de serviços baseado em QoS relativa e absoluta. São implementados e comparados diferentes algoritmos de escalonamento. Um dos algoritmos avaliados é o EBS. O objetivo é transpor a teoria da política para o mundo real, e comparar o seu comportamento com os resultados das simulações, utilizando o tempo de resposta como medida de desempenho. Além do EBS, outros algoritmos são avaliados, tais como, Round-Roubin e Weighted Round Robin / This work presents a prototype of a distributed Web server with service differentiation based on relative and absolute QoS. Are compared different schedule algorithms. One of the policies to be evaluated is the EBS. The goal is to implement the policy theory to the real world, comparing their behavior with the simulation results, using the response time as a performance measure. In addition to the EBS, others policies are evaluated, such as Round-Roubin and Weight Round Robin

Este trabalho de pesquisa apresenta um protótipo de servidor Web distribuído com diferenciação de serviços baseado em QoS relativa e absoluta. São implementados e comparados diferentes algoritmos de escalonamento. Um dos algoritmos avaliados é o EBS. O objetivo é transpor a teoria da política para o mundo real, e comparar o seu comportamento com os resultados das simulações, utilizando o tempo de resposta como medida de desempenho. Além do EBS, outros algoritmos são avaliados, tais como, Round-Roubin e Weighted Round Robin / This work presents a prototype of a distributed Web server with service differentiation based on relative and absolute QoS. Are compared different schedule algorithms. One of the policies to be evaluated is the EBS. The goal is to implement the policy theory to the real world, comparing their behavior with the simulation results, using the response time as a performance measure. In addition to the EBS, others policies are evaluated, such as Round-Roubin and Weight Round Robin

Este trabalho apresenta um estudo, implementação e validação em ambiente simulado de uma política de escalonamento de tempo real para provisão de QoS absoluto em serviço Web. Sintetizando características de escalonamento de tempo real, com baixa latência e de modelo re-alimentado, a política proposta permite um ajuste ponderado pela quantificação da exigência à qual o sistema está submetido por meio de suas classes. A meta é oferecer ações imediatas às requisições mais urgentes, sem, entretanto, degradar a qualidade do sistema como um todo. Verificou-se que a estratégia de escalonamento baseada em exigência (EBS - Exigency-Based Scheduling) é benéfica para o controle da qualidade de serviço oferecida. Escalonar de forma a evitar demasiado peso imposto ao sistema permite que o servidor tenha mais condições de cumprir os requisitos contratuais. Também foi alvo do estudo a criação de uma métrica de avaliação da satisfação de atendimento por parte dos usuários dos serviços. Os resultados alcançados com o emprego da política EBS sinalizam uma melhoria em termos de qualidade de serviço e melhor satisfação dos clientes de forma balanceada / The present work presents a study, implementation and validation in a simulated environment of a real time scheduling policy to provide absolute QoS for web services. Synthesizing characteristic from real time scheduling, low latency and feedback scheduling, the proposed policy allows an adjustment weighed by the quantification of the exigency which the system is exposed through its classes. The goal is to offer immediate actions to most urgent requests, without decreasing the system quality as a whole. It was verified that the scheduling strategy based on exigency (EBS - Exigency-Based Scheduling) helps to control the quality of service offered. Scheduling in order to avoid imposing a heavy load to the system gives more condition to the server to fulfill the requirements agreed. Another goal of this work is the creation of a metrics to evaluate the client satisfaction. The results achieved with the EBS policy indicate a higher quality of service and better client satisfaction

Este trabalho apresenta uma avaliação do algoritmo EBS, uma política de escalonamento proposta para sistemas de tempo real flexíveis com qualidade de serviço baseado em limites superiores para tempos médios de resposta. Experimentos têm demonstrado propriedades vantajosas da política EBS em servidores Web com diferenciação de serviço. O objetivo do presente estudo é compreender o comportamento da EBS em relação à diferentes parâmetros que descrevem a carga de trabalho. Esse conhecimento é útil para obtenção de um melhor aproveitamento computacional. São apresentados experimentos e resultados que analisam a influência de cada um dos fatores considerados na qualidade do serviço oferecido. A partir desses resultados são tecidas conclusões acerca de abordagens para o dimensionamento de carga e de capacidade do servidor / This Master degree project has presented an evaluation of the EBS algorithm, a scheduling policy proposed for soft real-time systems with quality of service based on upper limits for average response times. Experiments have shown advantageous properties of the EBS policy on Web servers with service differentiation. The aim of this study is to understand the behavior of the EBS in relation to different parameters that describe the workload. This knowledge is useful for obtaining a better use of computing. Experiments and results are presented analyzing the influence of each factor considering the quality of service offered. From these results, conclusions are woven about approaches to the design load and server capacity

通訊與資訊科技的大幅進步，電信自由化帶來的激烈競爭，以及網際網路的蓬勃發展，刺激大量多媒體網路資訊的流通，為了因應此種趨勢，網路提供者已趨向合併數據及電信網路朝單一的All-IP網路方向發展。為了保證時效性服務在All-IP網路上的品質，網路服務品質(QoS)已成為All-IP網路的主要研究議題。不同的網路應用各有不同的特性與需求；對於那些比較不注重傳輸延遲時間的應用，增加網路頻寬或許就已足夠應付需求，但是對於那些具有互動特性 (interactive)、重視傳輸延遲時間的應用，像VoIP，除了增加網路頻寬外，All-IP網路必須提供服務品質保證才能獲得網路營運者的支持。本研究團隊設計一個管理架構，在此架構上提供完整的End-to-End QoS保證，以符合All-IP網路上各種不同服務需求。本文中另外提出以預先批購頻寬的方式進行核心網路資源規劃，根據需求預測，考量批購成本期望值，決定出適當之頻寬預購值。於執行時段提出數個允入控制資源不足解決方案，並且配合執行時段頻寬管理機制，掌握頻寬使用情形，以達到順利允入網路訊務之目的。最後於NS2平台以實驗模擬的方式，評估本文中所提出之預先批購頻寬與執行時段頻寬管理機制，從結果中我們發現本研究所提出之頻寬預購方法可以有效預防因為預測誤差所造成之資源不足現象，配合執行時段頻寬管理機制可根據執行時期之資源使用狀況，在資源缺乏時提前進行頻寬補充，以順利允入訊務，提升使用率。 / Because of great progress of communication and computer technology, aggressive deployment of broadband fiber optical network, advance of Internet technology, and the global standardization of IP technology, the telecommunication industry is moving toward a converged network, which uses a single global IP based packet-switching network to carry all types of network traffics. In these types of network traffics, different traffic types require corresponding service to ensure end- to-end quality. For carrying all types of network traffics on All-IP network, BBQ (Budget-Based QoS) research group propose a QoS system architecture to provide end-to-end QoS guarantee. In this thesis, basing on BBQ QoS system, we propose resource pre-planning mechanism to management core network. According to demand forecast from historical data and considering pre-planning cost, pre-planning mechanism will find optimal policy to reduce management cost. Besides resource pre- planning, we also propose several solutions to avoid resource shortage at run-time. Through intensive evaluation in network simulatior-2(ns2), we demonstrate that our resource pre-planning can minimize resource cost and cover some forecasting error. And run-time resource management can maintain reasonable in-hand resource at run-time to reduce the effect of resource shortage.

隨著IEEE 802.11無線傳輸頻率的快速發展，而傳輸距離卻隨著縮短，使得我們考慮多跳接的傳輸方式來擴展通訊的範圍。但是在IEEE 802.11 的多跳接無線網路中並沒有服務品質的保證，而且在傳輸上的效能不高，因此我們希望在IEEE 802.11 多跳接的無線網路中改善傳輸效能且提供服務品質保證。在這篇論文中，我們提出了一個安排MAC存取的方法，來達到服務品質保證。我們使用一個樹狀架構來建構這個無線網路並且運作訊標繞樹的動作來建立我們的分離集合，而分離集合提供我們安排無衝突傳輸。當訊務流在這個網路中產生，我們的方法能夠有效的安排MAC存取並且保證不同類型訊務之品質。我們同時也提供一種適應性安排機制，來有效利用整個網路的效能，並且避免重新安排整個網路的MAC存取規畫，來維持網路品質在適當的滿意度。實驗結果顯示我們在多跳接的無線網路中有效的達到服務品質保證。 / With fast development of radio frequency, the wireless communication in IEEE 802.11 is growing up to high transmission rate (802.11g). However the distance of transmission decreases when wave frequency increases. Thus there is growing need to extend communication distance through multihop transmissions. But there is no QoS in IEEE 802.11 Multihop Wireless Networks; it has poor performance on transmission. Thus we are motivated to provide QoS in IEEE 802.11 Multihop Wireless Networks. In This Thesis, we propose a MAC Access Scheduling Scheme to support guaranteed QoS in IEEE 802.11 Multihop Wireless Networks. We use a tree-based architecture to construct the network and form the operation of token traversal. Token traversal operation builds the Disjoint Set to support scheduling collision-free transmission. When traffic is generated to the network, our scheme provides efficient schedule of MAC access and guarantee the QoS of different types of traffic. We also provide a adaptive scheduling to utilize the network performance and avoid the overhead of re-schedule to maintain the network quality at a satisfied level. The simulation results show we achieve the network QoS efficiently and provide guaranteed QoS.

Within recent years, the concept of Next Generation Networks (NGN) has become widely accepted within the telecommunication area, in parallel with the migration of telecommunication networks from traditional circuit-switched technologies such as ISDN (Integrated Services Digital Network) towards packet-switched NGN. In this context, SIP (Session Initiation Protocol), originally developed for Internet use only, has emerged as the major signalling protocol for multimedia sessions in IP (Internet Protocol) based NGN. One of the traditional limitations of IP when faced with the challenges of real-time communications is the lack of quality support at the network layer. In line with NGN specification work, international standardisation bodies have defined a sophisticated QoS (Quality of Service) architecture for NGN, controlling IP transport resources and conventional IP QoS mechanisms through centralised higher layer network elements via cross-layer signalling. Being able to centrally control QoS conditions for any media session in NGN without the imperative of a cross-layer approach would result in a feasible and less complex NGN architecture. Especially the demand for additional network elements would be decreased, resulting in the reduction of system and operational costs in both, service and transport infrastructure. This thesis proposes a novel framework for QoS optimisation for media sessions in SIP-based NGN without the need for cross-layer signalling. One key contribution of the framework is the approach to identify and logically group media sessions that encounter similar QoS conditions, which is performed by applying pattern recognition and clustering techniques. Based on this novel methodology, the framework provides functions and mechanisms for comprehensive resource-saving QoS estimation, adaptation of QoS conditions, and support of Call Admission Control. The framework can be integrated with any arbitrary SIP-IP-based real-time communication infrastructure, since it does not require access to any particular QoS control or monitoring functionalities provided within the IP transport network. The proposed framework concept has been deployed and validated in a prototypical simulation environment. Simulation results show MOS (Mean Opinion Score) improvement rates between 53 and 66 percent without any active control of transport network resources. Overall, the proposed framework comes as an effective concept for central controlled QoS optimisation in NGN without the need for cross-layer signalling. As such, by either being run stand-alone or combined with conventional QoS control mechanisms, the framework provides a comprehensive basis for both the reduction of complexity and mitigation of issues coming along with QoS provision in NGN.

Smartphones have become crucial enablers for users to exploit online services such as learning, leisure, communicating, and socializing. The user-perceived quality of applications and services is an important factor to consider, in order to achieve lean resource management, to prevent user churn and revenue depletion of service or network providers. This is often studied within the scope of Quality of Experience (QoE), which has attracted researchers both in academia and industry. The objective of this thesis is to study the most important factors influencing QoE on smartphones and synthesize solutions for intervention. The temporal impairments during a real-time energy-hungry video streaming are studied. The aim is to quantify the influence of temporal impairments on the user-perceived video QoE at the network and application level together with energy measurements, and also to propose solutions to reduce smartphone energy consumption without degrading the user’s QoE on the smartphone for both user-interactive, e.g., video, and non-interactive cases. QoE measurements on smartphones are performed throughout in-the-wild user studies. A set of quantitative Quality of Experience (QoE) assessment tools are implemented and deployed for automatic data logging at the network- and application-level. Online momentary survey, Experience Sampling Method (ESM) software, and Day Reconstruction Method (DRM) along weekly face-to-face user interviews are employed. The subjective QoE is obtained through qualitative feedback including Mean Opinion Score (MOS) as well as in-situ indications of poor experiences by users. Additionally, energy measurements on smartphones are conducted in controlled-lab environment with the Monsoon device. The QoE of smartphone applications and services perceived by users depends on many factors including anomalies in the network, application, and also the energy consumption. At the network-level, high packet delay variation causes long video freezes that eventually impact negatively the end-user perceived quality. The freezes can be quantified as large time gaps in-between the displayed pictures during a video stream at the application-level. We show that the inter-picture time in cellular-based video stream can be represented via two-state exponential ON/OFF models. We show models representing the non-linear relationship between the QoE and the mean inter-picture time. It is shown that energy measurements help to reveal the temporal impairments in video stream enabling energy consumption as a QoE indicator. Next, energy waste and saving during temporal impairments are identified. Additionally, other video streaming use cases, e.g., “download first and watch later”, are studied and appropriate energy-saving download scheduling mechanisms are recommended. The possibility for decreasing energy consumption when the smartphone screen is OFF, while maintaining QoE, is revealed. We first show exponential models to represent user’s interaction with smartphone, then propose a NyxEnergySaver software, to control the cellular network interface in a personalized manner to save smartphone energy. According to our findings, more than 30% smartphone energy can be saved without impacting the user-perceived QoE.

Este trabalho teve por objetivo estudar aspectos de Qualidade de Serviço (QoS) na transmissão de áudio e vídeo por redes de grande abrangência geográfica (WANs). Dois cenários distintos foram estudados. No primeiro foi feita uma proposta de um método de gerenciamento que permite que computadores de uma rede local (LAN) equipados com interfaces Ethernet possam criar canais dinamicamente através de um backbone ATM para tráfego de voz e vídeo. Uma implementação do modelo foi avaliada para ilustrar a viabilidade do método proposto. No segundo cenário foi mostrado através de extensas simulações computacionais que o princípio da deflexão pode ser usado com êxito para oferecer QoS a aplicações em tempo real. Nestas simulações foi usado como exemplo prático uma aplicação de voz sobre IP (VoIP).

The growing popularity of multi-tenant, cloud-based computing platforms is increasing interest in resource allocation models that permit flexible sharing of the underlying infrastructure. This thesis introduces a novel IO resource allocation model that better captures the requirements of paying tenants sharing a physical infrastructure. The model addresses a major concern regarding application performance stability when clients migrate from a dedicated to a shared platform. Specifically, while clients would like their applications to behave similarly in both situations, traditional models of fairness, like proportional share allocation, do not exhibit this behavior in the context of modern multi-tiered storage architectures. We also present a scheduling algorithm, the Reward Scheduler, that implements the new allocation policy, by rewarding clients with better runtime characteristics, resulting in benefits to both the clients and the service provider. Moreover, the Reward scheduler also supports weight-based capacity allocation subject to a minimum reservation and maximum limitation on the IO allocation for each task. Experimental results indicate that the proposed algorithm proportionally allocates the system capacity in proportion to their entitlements.