With the emergence of computer networks as one of the primary platforms of communication, and with their adoption for an increasingly broad range of applications, there is a growing need for high-quality network traffic measurements to better understand, characterize and engineer the network behaviors. Due to the inherent lack of fine-grained measurement capabilities in the original design of the Internet, it does not have enough data or information to compute or even approximate some traffic statistics such as traffic matrices and per-link delay. While it is possible to infer these statistics from indirect aggregate measurements that are widely supported by network measurement devices (e.g., routers), how to obtain the best possible inferences is often a challenging research problem. We name this as "too little data" problem after its root cause. Interestingly, while "too little data" is clearly a problem, "too much data" is not a blessing either. With the rapid increase of network link speeds, even to keep sampled summarized network traffic (for inferring various network statistics) at low sample rates results in too much data to be stored, processed, and transmitted over measurement devices. In summary high-quality measurements in today's Internet is very challenging due to resource limitations and lack of built-in support, manifested as either "too little data" or "too much data". We present some new practices and proposals to alleviate these two problems.The contribution is four fold: i) designing universal methodologies towards ideal network traffic measurements; ii) providing accurate estimations for several critical traffic statistics guided by the proposed methodologies; iii) offering multiple useful and extensible building blocks which can be used to construct a universal network measurement system in the future; iv) leading to some notable mathematical results such as a new large deviation theorem that finds applications in various areas.

ABSTRACT EXPLORING THE RELATIONSHIP BETWEEN SUPPLY NETWORK CONFIGURATION, INTER-ORGANIZATIONAL INFORMATION SHARING AND PERFORMANCE By MARCIA DALEY August 2008 Committee Chair: Dr. Subhashish Samaddar Major Department: Decision Science Critical to the success of a firm is the ability of managers to coordinate the complex network of business relationships that can exist between business partners in the supply network. However many managers are unsure on how best to leverage their resources to capitalize on the information sharing opportunities that are available in such networks. Although there is significant research on information sharing, the area of inter-organizational information sharing (IIS) is still evolving and there is limited research on IIS in relation to systemic factors within supply networks. To help fill this gap in the literature, a primary focus of this dissertation is on the relationship between the design of the supply network and IIS. The design of the supply network is characterized by the supply network configuration which is comprised of (1) the network pattern, (2) the number of stages in the supply network, and (3) where the firm is located in that supply network. Four different types of IIS are investigated, herein. These types of IIS are a function of the frequency with which information is shared and the scope of information shared. Type 1 (Type 2) IIS is the low (high) frequency state where only operational information is shared. Similarly, Type 3 (Type 4) is the low (high) frequency state where strategic information is shared. The argument is that the type of IIS varies depending on the configuration of the supply network and that this relationship is influenced by the coordination structure established between firms in the network. The second focus of this dissertation deals with the relationship between IIS and performance. Research findings on the benefits to be gained from IIS have been ambiguous, with some researchers claiming reduced cost in the supply network with IIS, and others finding minimal or no benefits. To add clarity to these findings, the role that uncertainty plays in the relationship between IIS and performance is examined. The thesis presented is that the positive relationship between IIS types and the performance of the supply network is impacted by process uncertainty (i.e. the variability in process outcomes and production times), and partner uncertainty. Social network theory and transaction cost economics provide the theoretical lens for this dissertation. A model is developed and will be empirically validated in a cross-sectional setting, utilizing a sampling frame randomly selected and comprised of supply management executives from various industries within the United States.

The objective of the proposed research is to design and evaluate end-to-end solutions to support the Quality of Experience (QoE) for the Internet Protocol Television (IPTV) service. IPTV is a system that integrates voice, video, and data delivery into a single Internet Protocol (IP) framework to enable interactive broadcasting services at the subscribers. It promises significant advantages for both service providers and subscribers. For instance, unlike conventional broadcasting systems, IPTV broadcasts will not be restricted by the limited number of channels in the broadcast/radio spectrum. Furthermore, IPTV will provide its subscribers with the opportunity to access and interact with a wide variety of high-quality on-demand video content over the Internet. However, these advantages come at the expense of stricter quality of service (QoS) requirements than traditional Internet applications. Since IPTV is considered as a real-time broadcast service over the Internet, the success of the IPTV service depends on the QoE perceived by the end-users. The characteristics of the video traffic as well as the high-quality requirements of the IPTV broadcast impose strict requirements on transmission delay. IPTV framework has to provide mechanisms to satisfy the stringent delay, jitter, and packet loss requirements of the IPTV service over lossy transmission channels with varying characteristics. The proposed research focuses on error recovery and channel change latency problems in IPTV networks. Our specific aim is to develop a content delivery framework that integrates content features, IPTV application requirements, and network characteristics in such a way that the network resource utilization can be optimized for the given constraints on the user perceived service quality. To achieve the desired QoE levels, the proposed research focuses on the design of resource optimal server-based and peer-assisted delivery techniques. First, by analyzing the tradeoffs on the use of proactive and reactive repair techniques, a solution that optimizes the error recovery overhead is proposed. Further analysis on the proposed solution is performed by also focusing on the use of multicast error recovery techniques. By investigating the tradeoffs on the use of network-assisted and client-based channel change solutions, distributed content delivery frameworks are proposed to optimize the error recovery performance. Next, bandwidth and latency tradeoffs associated with the use of concurrent delivery streams to support the IPTV channel change are analyzed, and the results are used to develop a resource-optimal channel change framework that greatly improves the latency performance in the network. For both problems studied in this research, scalability concerns for the IPTV service are addressed by properly integrating peer-based delivery techniques into server-based solutions.

The heavy-tailed traffic from wireless users, caused by the emerging Internet and multimedia applications, induces extremely dynamic and variable network environment, which can fundamentally change the way in which wireless networks are conceived, designed, and operated. This thesis is concerned with modeling, analysis, and optimization of wireless networks in the presence of heavy tails. First, a novel traffic model is proposed, which captures the inherent relationship between the traffic dynamics and the joint effects of the mobility variability of network users and the spatial correlation in their observed physical phenomenon. Next, the asymptotic delay distribution of wireless users is analyzed under different traffic patterns and spectrum conditions, which reveals the critical conditions under which wireless users can experience heavy-tailed delay with significantly degraded QoS performance. Based on the delay analysis, the fundamental impact of heavy-tailed environment on network stability is studied. Specifically, a new network stability criterion, namely moment stability, is introduced to better characterize the QoS performance in the heavy-tailed environment. Accordingly, a throughput-optimal scheduling algorithm is proposed to maximize network throughput while guaranteeing moment stability. Furthermore, the impact of heavy-tailed spectrum on network connectivity is investigated. Towards this, the necessary conditions on the existence of delay-bounded connectivity are derived. To enhance network connectivity, the mobility-assisted data forwarding scheme is exploited, whose important design parameters, such as critical mobility radius, are derived. Moreover, the latency in wireless mobile networks is analyzed, which exhibits asymptotic linearity in the initial distance between mobile users.

In this thesis, we analyze the bandwidth requirements of MPEG-4 AVC video traffic and then propose and evaluate the accuracy of new MPEG-4 AVC video traffic models. First, we analyze the bandwidth requirements of the videos by comparing the statistical characteristics of the different frame types. We analyze their coefficient of variability, autocorrelation, and crosscorrelation in both short and long term. The Hurst parameter is also used to investigate the long range dependence of the video traces. We then provide an insight into B-frame dropping and its impact on the statistical characteristics of the video trace. This leads us to design two algorithms that predict the size of the B-frame and the size of the group of pictures (GOP) in the short-term. To evaluate the accuracy of the prediction, a model for the error is proposed. In a broadband cable network, B-frame size prediction can be employed by a cable headend to provision video bandwidth efficiently or more importantly, reduce bit rate variability and bandwidth requirements via selective B-frame dropping, thereby minimizing buffering requirements and packet losses at the set top box. It will be shown that the model provides highly accurate prediction, in particular for movies encoded in high quality resolution. The GOP size prediction can be used to provision bandwidth. We then enhance the B-frame and GOP size prediction models using a new scene change detector metric. Finally, we design an algorithm that predicts the size of different frame types in the long-term. Clearly, a long-term prediction algorithm may suffer degraded prediction accuracy and the higher complexity may result in higher latency. However, this is offset by the additional time available for long-term prediction and the need to forecast bandwidth usage well ahead of time in order to minimize packet losses during periods of peak bandwidth demands. We also analyze the impact of the video quality and the video standard on the accuracy of the model.

Thesis (MSc (Mathematical Sciences. Computer Science))--University of Stellenbosch, 2009. / ENGLISH ABSTRACT: Since 2002 data tra c has overtaken voice tra c in volume [1]. Telecom / Network operators still generate most of their income carrying voice tra c. There is however a huge revenue potential in delivering reliable guaranteed data services. Network survivability and recovery from network failures are integral to network reliability. Due to the nature of the Internet, recovery from link failures needs to be distributed and dynamic in order to be scalable. Link failure recovery schemes are evaluated in terms of the survivability of the network, the optimal use of network resources, scalability, and the recovery time of such schemes. The need for recovery time to be improved is highlighted by real-time data tra c such as VoIP and video services carried over the Internet. The goal of this thesis is to examine existing link failure recovery schemes and evaluate the need for their extension, and to evaluate the performance of the proposed link failure recovery schemes. i / AFRIKAANSE OPSOMMING: Sedert 2002 het data verkeer die stem verkeer in volume verbygesteek [1]. Telekommunikasie / netwerk operateurs genereer egter steeds die meeste van hul inkomste met stem verkeer. Netwerk oorlewing en die herstel van netwerk mislukkings is integraal tot netwerk stabiliteit. Die samestelling van die Internet noodsaak dat die herstel van skakel mislukkings verspreid en dinamies van natuur moet wees. Die herstel-skema van skakel mislukkings word evalueer in terme van die oorleefbaarheid van die netwerk, die mees e ektiewe benutting van network bronne, aanpasbaarheid, en die herstel tydperk van die skema. Die vinnig moontlikste herstel tydperk word genoodsaak deur oombliklike data verkeer soos VoIP en beeld dienste wat oor die Internet gedra word. The doel van hierdie tesis is om bestaande skakel mislukking herstel skemas te evalueer, en dan verder ondersoek in te stel na hul uitbreiding. Daarna word die voorgestelde skakel mislukking skema se e ektiwiteit gemeet.

A crescente complexidade das redes de comunicação de dados tem como conseqüência direta tornar cada vez mais complexas as tarefas de projetar seu dimensionamento e evolução. Um passo preliminar vitalmente importante no projeto de uma rede é a coleta de dados relacionados ao uso esperado da rede. Os padrões de tráfego estimados são usados nos cálculos de dimensionamento dos recursos. Fortemente relacionada a estas atividades está a previsão da performance em termos de throughput da rede, tempos de resposta, probabilidade de congestionamento, entre outras. Uma análise de performance efetiva depende de se representar precisamente a configuração da rede e da carga de tráfego a ela submetida. Técnicas de modelagem analítica ou de simulação podem ser usadas para determinar a performance esperada do sistema. Neste trabalho a usada a técnica de simulação para modelar o tráfego da rede usando dados coletados em uma rede real por monitoração. Trabalhando-se com o modelo validado e possível ajustar a carga de tráfego para representar mudanças esperadas nos volumes de tráfego, e testar o efeito na performance de diferentes configurações de rede. Isso possibilita que os processos de planejamento e projeto sejam executados com confiança, alem dos custos da rede poderem ser otimizados. O sistema apresentado neste trabalho e orientado à gerência de performance de redes, apoiando as atividades de gerência e planejamento de capacidade. Defende-se a idéia de que, a partir do perfil dos usuários de uma rede, derivado do tráfego por eles gerado, pode-se construir um modelo de tráfego significativamente representativo para a análise e previsão do desempenho da rede. O modelo foi desenvolvido para redes de comutação de pacotes com serviço orientado a conexão, e validado sobre o tráfego de uma rede X.25 através de um modelo de simulação, implementado em GPSS. O protótipo do sistema é composto por três processos principais que incorporam as etapas da análise de performance: monitoração do tráfego, análise do tráfego e simulação do tráfego. Na monitoração, são contabilizadas uma série de variáveis para cada conexão gerada na rede. Para cada tipo de aplicação são calculados os respectivos fatores de carga de tráfego. As aplicações são classificadas em dois grandes grupos: interativas e não-interativas. O processo de análise do tráfego gera os fatores de carga de tráfego na forma de tabelas e gráficos. As conexões e as métricas de performance produzidas na simulação do tráfego podem ser analisadas e validadas graficamente. O sistema proposto foi especificado em SDL (Specification and Description Language). / The increasing complexity of communication networks turns each time more complex the task to plan its expansion and evolution. A vitally important preliminar step in computer network design is the gathering of data relating to the expected use of the network. The estimated traffic patterns are used in resource usage computations. Closely related to these activities is the estimation of performance in terms of network throughput, response times, congestion probability, among others. An efective performance analysis requires the accurate representation of computer network configuration and traffic load. Analytical or simulation techniques can be used to establish the expected system performance. This work makes use of simulation to model the network traffic, based on collected data of a real network. network. Working with the validated model is possible to adjust the traffic load to represent future demand, and to test the performance impact under distinct network configurations. This enable a trustful execution of network planning and design, besides of the cost optimization. The system presented at this work is network performance management oriented, giving support to capacity management and capacity planning. The central idea is that, through network users profile achieved from generated traffic, it's possible to build a representative traffic model to aim at the network performance analysis and forecast. The model was developed to connection oriented packet switched networks, and validated by means of a simulation model to X.25 network traffic, implementd in GPSS. The system prototype comprises three process: traffic monitoring, traffic analysis and traffic simulation. In monitoring, a set of variables are registered for each network connection generated. The traffic load factors are computed according the application type. The applications are classified in two groups: interactive and non-interactive. The traffic analysis process produces traffic load factors in form of tables and graphics. Connections and performance metrics generated at traffic simulation can be graphically analyzed and validated. The proposed system was specified in SDL (Specification and Description Language).

A crescente complexidade das redes de comunicação de dados tem como conseqüência direta tornar cada vez mais complexas as tarefas de projetar seu dimensionamento e evolução. Um passo preliminar vitalmente importante no projeto de uma rede é a coleta de dados relacionados ao uso esperado da rede. Os padrões de tráfego estimados são usados nos cálculos de dimensionamento dos recursos. Fortemente relacionada a estas atividades está a previsão da performance em termos de throughput da rede, tempos de resposta, probabilidade de congestionamento, entre outras. Uma análise de performance efetiva depende de se representar precisamente a configuração da rede e da carga de tráfego a ela submetida. Técnicas de modelagem analítica ou de simulação podem ser usadas para determinar a performance esperada do sistema. Neste trabalho a usada a técnica de simulação para modelar o tráfego da rede usando dados coletados em uma rede real por monitoração. Trabalhando-se com o modelo validado e possível ajustar a carga de tráfego para representar mudanças esperadas nos volumes de tráfego, e testar o efeito na performance de diferentes configurações de rede. Isso possibilita que os processos de planejamento e projeto sejam executados com confiança, alem dos custos da rede poderem ser otimizados. O sistema apresentado neste trabalho e orientado à gerência de performance de redes, apoiando as atividades de gerência e planejamento de capacidade. Defende-se a idéia de que, a partir do perfil dos usuários de uma rede, derivado do tráfego por eles gerado, pode-se construir um modelo de tráfego significativamente representativo para a análise e previsão do desempenho da rede. O modelo foi desenvolvido para redes de comutação de pacotes com serviço orientado a conexão, e validado sobre o tráfego de uma rede X.25 através de um modelo de simulação, implementado em GPSS. O protótipo do sistema é composto por três processos principais que incorporam as etapas da análise de performance: monitoração do tráfego, análise do tráfego e simulação do tráfego. Na monitoração, são contabilizadas uma série de variáveis para cada conexão gerada na rede. Para cada tipo de aplicação são calculados os respectivos fatores de carga de tráfego. As aplicações são classificadas em dois grandes grupos: interativas e não-interativas. O processo de análise do tráfego gera os fatores de carga de tráfego na forma de tabelas e gráficos. As conexões e as métricas de performance produzidas na simulação do tráfego podem ser analisadas e validadas graficamente. O sistema proposto foi especificado em SDL (Specification and Description Language). / The increasing complexity of communication networks turns each time more complex the task to plan its expansion and evolution. A vitally important preliminar step in computer network design is the gathering of data relating to the expected use of the network. The estimated traffic patterns are used in resource usage computations. Closely related to these activities is the estimation of performance in terms of network throughput, response times, congestion probability, among others. An efective performance analysis requires the accurate representation of computer network configuration and traffic load. Analytical or simulation techniques can be used to establish the expected system performance. This work makes use of simulation to model the network traffic, based on collected data of a real network. network. Working with the validated model is possible to adjust the traffic load to represent future demand, and to test the performance impact under distinct network configurations. This enable a trustful execution of network planning and design, besides of the cost optimization. The system presented at this work is network performance management oriented, giving support to capacity management and capacity planning. The central idea is that, through network users profile achieved from generated traffic, it's possible to build a representative traffic model to aim at the network performance analysis and forecast. The model was developed to connection oriented packet switched networks, and validated by means of a simulation model to X.25 network traffic, implementd in GPSS. The system prototype comprises three process: traffic monitoring, traffic analysis and traffic simulation. In monitoring, a set of variables are registered for each network connection generated. The traffic load factors are computed according the application type. The applications are classified in two groups: interactive and non-interactive. The traffic analysis process produces traffic load factors in form of tables and graphics. Connections and performance metrics generated at traffic simulation can be graphically analyzed and validated. The proposed system was specified in SDL (Specification and Description Language).

Today’s technology has a strong tendency towards automation. As a matter of fact, the tremendous improvement of science in the recent years brought new ideas regarding to accelerate the scientific process that is not separated of automation. This paper also deals with automation of manual tests that were used to analyze packet loss and packet loss duration in a network. These two metrics were chosen based on their importance they have in the communication technology, also based on the weak points that were found in the manual processes. This experiment is done in the Ixia platform that was an appropriate test bed to design an automation framework.After a comprehensive research on network and communication we could choose packet loss and packet loss duration as two important parameters that are under test several times per day. Therefore, based on the properties that are used for automation, these two metrics had the priority compare to other metrics. We could create a framework that works correspond to the manual test process. For this purpose, Tcl programming language is used. The script that was written with this high-level language can communicate with the graphical user interface, configuring all the connected devices, measuring mentioned metrics and ultimately save the result in a csv file.Finally, we could reach to the main objective of this project which was to show how positively automatic method can affect on the quality of test in terms of accuracy, time and manpower saving.

Eleven years ago Cisco introduced three types of applications on the same converged network and named it AVVID (Architecture for Voice, Video and Integrated Data). In spite of the initial interest and hype surrounding AVVID, the feature was and still is unable to confront problems such as: data priority, load balancing and network congestion. The work in this report addressed these issues within the network. Different Cisco Internetwork Operating System (IOS) methods: routing protocols, Cisco IOS QoS (including LLQ, LFI and Header Compression), Path Control and Cisco Performance Routing (PfR) were all tested to see which feature would work best at enhancing network performance. A practical network was created in which these features were carefully implemented one after another. In every step, the results were observed and recorded and if necessary repeated in order to check their validity. The results for RTT (Round Trip Time) and MOS (Mean Opinion Score) showed that in comparison to Path Control and Cisco IOS QoS (Quality of Service) mechanisms, Cisco PfR was able to generate routes on the fly during times of network congestion. This cannot be achieved with the use of Cisco QoS and Path Control which makes Cisco Performance Routing a far more superior feature. Although the focus of the work was VoIP (Voice over Internet Protocol) data, it is possible to construct a network with any application as Cisco PfR is able to reroute data. Cisco PfR is an easily accessible IOS feature which, once implemented, requires very little updating and maintenance.