Master of Science / Department of Electrical and Computer Engineering / Don M. Gruenbacher / This thesis presents a novel device for delaying Ethernet traffic in a lab setting. Ethernet is the leading standard for communications between computing devices. With the advent of streaming media such as voice over IP phone service and real-time control systems over Ethernet, applications are being rapidly developed that must meet strict communication reliability and timing constraints. Increasingly, these systems must be examined in real world scenarios before actual hardware deployment or protocol release. This increases the demand for both testing equipment and well trained network engineers. Commercial Ethernet delay testing devices are expensive, hardware specific, and not flexible enough for educational purposes. These short-comings make it necessary to design a robust Field Programmable Gate Array (FPGA) based Ethernet delay device that is up to the rigor of educational and research settings. Our approach is based on the inexpensive, high performance Altera Stratix II GX PCI Express development board which can easily be adapted for different delay scenarios. The system's FPGA hardware was developed in Verilog, an industry standard hardware description language, so users will be able to quickly learn, adapt and operate the system. Software for the system's soft processor was developed in C. The device provides a wide range of packet delay from nearly zero up to over fifty milliseconds, as well as providing an easy to use interface with on-the-fly variable delay adjustment. Theoretical throughput was up to 1Gb/s; skew and jitter measurements were comparable with common network switches. These properties allow the device to provide an easy-to-use, inexpensive method to delay Ethernet traffic in lab settings and the device also creates a starting point for future students and researchers to develop high speed traffic delay testbeds. Future work will include 10Gb/s throughput, additional memory capacity and additional software implemented delay profiles.

Master of Engineering by Research / Service-level-agreement (SLA) monitoring measures network Quality-of-Service (QoS) parameters to evaluate whether the service performance complies with the SLAs. It is becoming increasingly important for both Internet service providers (ISPs) and their customers. However, the rapid expansion of the Internet makes SLA monitoring a challenging task. As an efficient method to reduce both complexity and overheads for QoS measurements, sampling techniques have been used in SLA monitoring systems. In this thesis, I conduct a comprehensive study of sampling methods for network QoS measurements. I develop an efficient sampling strategy, which makes the measurements less intrusive and more efficient, and I design a network performance monitoring software, which monitors such QoS parameters as packet delay, packet loss and jitter for SLA monitoring and verification. The thesis starts with a discussion on the characteristics of QoS metrics related to the design of the monitoring system and the challenges in monitoring these metrics. Major measurement methodologies for monitoring these metrics are introduced. Existing monitoring systems can be broadly classified into two categories: active and passive measurements. The advantages and disadvantages of both methodologies are discussed and an active measurement methodology is chosen to realise the monitoring system. Secondly, the thesis describes the most common sampling techniques, such as systematic sampling, Poisson sampling and stratified random sampling. Theoretical analysis is performed on the fundamental limits of sampling accuracy. Theoretical analysis is also conducted on the performance of the sampling techniques, which is validated using simulation with real traffic. Both theoretical analysis and simulation results show that the stratified random sampling with optimum allocation achieves the best performance, compared with the other sampling methods. However, stratified sampling with optimum allocation requires extra statistics from the parent traffic traces, which cannot be obtained in real applications. In order to overcome this shortcoming, a novel adaptive stratified sampling strategy is proposed, based on stratified sampling with optimum allocation. A least-mean-square (LMS) linear prediction algorithm is employed to predict the required statistics from the past observations. Simulation results show that the proposed adaptive stratified sampling method closely approaches the performance of the stratified sampling with optimum allocation. Finally, a detailed introduction to the SLA monitoring software design is presented. Measurement results are displayed which calibrate systematic error in the measurements. Measurements between various remote sites have demonstrated impressively good QoS provided by Australian ISPs for premium services.

Master of Engineering by Research / Service-level-agreement (SLA) monitoring measures network Quality-of-Service (QoS) parameters to evaluate whether the service performance complies with the SLAs. It is becoming increasingly important for both Internet service providers (ISPs) and their customers. However, the rapid expansion of the Internet makes SLA monitoring a challenging task. As an efficient method to reduce both complexity and overheads for QoS measurements, sampling techniques have been used in SLA monitoring systems. In this thesis, I conduct a comprehensive study of sampling methods for network QoS measurements. I develop an efficient sampling strategy, which makes the measurements less intrusive and more efficient, and I design a network performance monitoring software, which monitors such QoS parameters as packet delay, packet loss and jitter for SLA monitoring and verification. The thesis starts with a discussion on the characteristics of QoS metrics related to the design of the monitoring system and the challenges in monitoring these metrics. Major measurement methodologies for monitoring these metrics are introduced. Existing monitoring systems can be broadly classified into two categories: active and passive measurements. The advantages and disadvantages of both methodologies are discussed and an active measurement methodology is chosen to realise the monitoring system. Secondly, the thesis describes the most common sampling techniques, such as systematic sampling, Poisson sampling and stratified random sampling. Theoretical analysis is performed on the fundamental limits of sampling accuracy. Theoretical analysis is also conducted on the performance of the sampling techniques, which is validated using simulation with real traffic. Both theoretical analysis and simulation results show that the stratified random sampling with optimum allocation achieves the best performance, compared with the other sampling methods. However, stratified sampling with optimum allocation requires extra statistics from the parent traffic traces, which cannot be obtained in real applications. In order to overcome this shortcoming, a novel adaptive stratified sampling strategy is proposed, based on stratified sampling with optimum allocation. A least-mean-square (LMS) linear prediction algorithm is employed to predict the required statistics from the past observations. Simulation results show that the proposed adaptive stratified sampling method closely approaches the performance of the stratified sampling with optimum allocation. Finally, a detailed introduction to the SLA monitoring software design is presented. Measurement results are displayed which calibrate systematic error in the measurements. Measurements between various remote sites have demonstrated impressively good QoS provided by Australian ISPs for premium services.

Radio networks are expanding, becoming more advanced, and pushing the limits of what is possible. Services utilizing the radio networks are also being developed in order to provide new functionality to end-users worldwide. When discussing 5G radio networks, concepts such as driverless vehicles, drones and near zero communication delay are recurrent. However, measures of delay are needed in order to verify that such services can be provided -- and measuring this is an extensive task. Ericsson has developed a platform for simulating a radio environment surrounding a radio base station. Using this simulator, this project involved measuring one-way packet delay in a radio network, and performing a Quality of Service evaluation of a radio network with a number of network applications in concern. Application data corresponding to video streams, or Voice over IP conversations, were simulated and packet delay measurements were used to calculate and evaluate the Quality of Service provided by a radio network. One of the main conclusions of this project was that packet delay variations are asymmetric in uplink, which suggests usage of non-conventional jitter measurement techniques.

In a WiMAX network, four traffic types with different priorities are defined. They are Unsolicited Grant Service (UGS), real-time Polling Service (rtPS), non-real-time Polling Service (nrtPS), and Best Effort (BE). In this thesis, we propose a Dynamic Bandwidth Allocation (DBA) scheme for BS to schedule the four above-mentioned traffic types. By adopting Deficit Round Robin (DRR) scheduling, DBA first assigns minimum quantum to each traffic type for transmission. When rtPS packets exceed their delay constraints, without sacrificing the minimum requirements of nrtPS and BE traffic, DBA borrows some quantum from nrtPS and BE to satisfy the delay requirements of rtPS traffic. When nrtPS packets can not reach the minimum transmission rate, without starving the BE traffic, DBA borrows some quantum from BE to support the required throughput of nrtPS traffic. According to the history record of borrowed quantum, DBA dynamically adjusts the assigned quantum for the three traffic types. For the purpose of evaluation, we use NS-2 to simulate the proposed DBA. We adjust the traffic load to analyze the performance in terms of average packet delay, average throughput, and average packet loss ratio. The simulation results show that the DBA, in comparison to a previous work, can promise the delay constraints of rtPS, maintain the average throughput of nrtPS, and avoid the starvation of BE, when the traffic load is high.

The demand for ever larger, more efficient, reliable and cost effective communication networks necessitates new network architectures, such as wireless ad hoc networks, cognitive radio, relaying networks, and wireless sensor networks. The study of such networks requires a fundamental shift from thinking of a network as a collection of independent communication pipes, to a multi-user channel where users cooperate via conferencing, relaying, and joint source-channel coding. The traditional centralized networks, such as cellular networks, include a central controller and a fixed infrastructure, in which every node communicates with each other via a centralized based station (BS). However, for a decentralized network, such as wireless ad hoc networks and wireless sensor networks, there is no infrastructure support and no central controllers. In such multi-user wireless networks, the scheduling algorithm plays an essential role in efficiently assigning channel resources to different users for better system performance, in terms of system throughput, packet-delay, stability and fairness. In this dissertation, our main goal is to develop practical scheduling algorithms in wireless ad hoc networks to enhance system performance, in terms of throughput, delay and stability. Our dissertation mainly consists of three main parts. First, we identify major challenges intrinsic to ad hoc networks that affect the system performance, in terms of throughput limits, delay and stability condition. Second, we develop scheduling algorithms for wireless ad hoc networks, with various considerations of non-cooperative relays and cooperative relays, fixed-rate transmission and adaptive-rate transmission, full-buffer traffic model and finite-buffer traffic model. Specifically, we propose an opportunistic scheduling scheme and study the throughput and delay performance, with fixed-rate transmissions in a two-hop wireless ad hoc networks. In the proposed scheduling scheme, we prove two key inequalities that capture the various tradeoffs inherent in the broad class of opportunistic relaying protocols, illustrating that no scheduling and routing algorithm can simultaneously yield lower delay and higher throughput. We then develop an adaptive rate transmission scheme with opportunistic scheduling, with the constraints of practical assumptions on channel state information (CSI) and limited feedback, which achieves an optimal system throughput scaling order. Along this work with the consideration of finite-buffer model, we propose a Buffer-Aware Adaptive (BAA) scheduler which considers both channel state and buffer conditions to make scheduling decisions, to reduce average packet delay, while maintaining the queue stability condition of the networks. The proposed algorithm is an improvement over existing algorithms with adaptability and bounded potential throughput reduction. In the third part, we extend the methods and analyses developed for wireless ad hoc networks to a practical Aeronautical Communication Networks (ACN) and present the system performance of such networks. We use our previously proposed scheduling schemes and analytical methods from the second part to investigate the issues about connectivity, throughput and delay in ACN, for both single-hop and two-hop communication models. We conclude that the two-hop model achieves greater throughput than the single-hop model for ACN. Both throughput and delay performances are characterized.

abstract: The upstream transmission of bulk data files in Ethernet passive optical networks (EPONs) arises from a number of applications, such as data back-up and multimedia file upload. Existing upstream transmission approaches lead to severe delays for conventional packet traffic when best-effort file and packet traffic are mixed. I propose and evaluate an exclusive interval for bulk transfer (EIBT) transmission strategy that reserves an EIBT for file traffic in an EPON polling cycle. I optimize the duration of the EIBT to minimize a weighted sum of packet and file delays. Through mathematical delay analysis and verifying simulation, it is demonstrated that the EIBT approach preserves small delays for packet traffic while efficiently serving bulk data file transfers. Dynamic circuits are well suited for applications that require predictable service with a constant bit rate for a prescribed period of time, such as demanding e-science applications. Past research on upstream transmission in passive optical networks (PONs) has mainly considered packet-switched traffic and has focused on optimizing packet-level performance metrics, such as reducing mean delay. This study proposes and evaluates a dynamic circuit and packet PON (DyCaPPON) that provides dynamic circuits along with packet-switched service. DyCaPPON provides (i) flexible packet-switched service through dynamic bandwidth allocation in periodic polling cycles, and (ii) consistent circuit service by allocating each active circuit a fixed-duration upstream transmission window during each fixed-duration polling cycle. I analyze circuit-level performance metrics, including the blocking probability of dynamic circuit requests in DyCaPPON through a stochastic knapsack-based analysis. Through this analysis I also determine the bandwidth occupied by admitted circuits. The remaining bandwidth is available for packet traffic and I analyze the resulting mean delay of packet traffic. Through extensive numerical evaluations and verifying simulations, the circuit blocking and packet delay trade-offs in DyCaPPON is demonstrated. An extended version of the DyCaPPON designed for light traffic situation is introduced in this article as well. / Dissertation/Thesis / Ph.D. Electrical Engineering 2014

Vehicular ad-hoc networks (VANETs) are used for vehicle-to-vehicle (V2V) and vehicle-to-infrastructure (V2I) communications. The IEEE 802.11p/WAVE (Wireless Access in Vehicular Environment) and with WAVE Short Messaging Protocol (WSMP) has been proposed as the standard protocol for designing applications for VANETs. This communication protocol must be thoroughly tested before reliable and efficient applications can be built using its protocols. In this paper, we perform on-road experiments in a variety of scenarios to evaluate the performance of the standard. We use commercial VANET devices with 802.11p/WAVE compliant chipsets for both BSM (basic safety messages) as well as video streaming applications using WSMP as a communication protocol. We show that while the standard performs well for BSM application in lightly loaded conditions, the performance becomes inferior when traffic and other performance metric increases. Furthermore, we also show that the standard is not suitable for video streaming due to the bursty nature of traffic and the bandwidth throttling, which is a major shortcoming for V2X applications.

Esta Tese propõe duas métricas objetivas para avaliar a percepção de qualidade de vídeos sujeitos a degradações de transmissão em uma rede de pacotes. A primeira métrica usa apenas o vídeo degradado, enquanto que a segunda usa os vídeos de referência e degradado. Esta última é uma métrica de referência completa (FR - Full Reference) chamada de QCM (Quadratic Combinational Metric) e a primeira é uma métrica sem referência (NR - No Reference) chamada de VQOM (Viewing Quality Objective Metric). Em particular, o procedimento de projeto é aplicado à degradação de variação de atraso de pacotes (PDV - Packet Delay Variation). A métrica NR é descrita por uma spline cúbica composta por dois polinômios cúbicos que se encontram suavemente num ponto chamado de nó. Para o projeto de ambas métricas, colhem-se opiniões de observadores a respeito das sequências de vídeo degradadas que compõem o conjunto. A função objetiva inclui o erro quadrático total entre as opiniões e suas estimativas paramétricas, ainda consideradas como expressões algébricas. Acrescentam-se à função objetiva três condições de igualdades de derivadas tomadas no nó, cuja posição é especificada dentro de uma grade fina de pontos entre o valor mínimo e o valor máximo do fator de degradação. Essas condições são afetadas por multiplicadores de Lagrange e adicionadas à função objetiva, obtendo-se o lagrangiano, que é minimizado pela determinação dos coeficientes subótimos dos polinômios em função de cada valor do nó na grade. Finalmente escolhe-se o valor do nó que produz o erro quadrático mínimo, determinando assim os valores finais para dos coeficientes do polinômio. Por outro lado, a métrica FR é uma combinação não-linear de duas métricas populares, a PSNR (Peak Signal-to-Noise Ratio) e a SSIM (Structural Similarity Index). Um polinômio completo de segundo grau de duas variáveis é usado para realizar a combinação, porque é sensível a ambas métricas constituintes, evitando o sobreajuste em decorrência do baixo grau. Na fase de treinamento, o conjunto de valores dos coeficientes do polinômio é determinado através da minimização do erro quadrático médio para as opiniões sobre a base de dados de treino. Ambas métricas, a VQOM e a QCM, são treinadas e validadas usando uma base de dados, e testadas com outra independente. Os resultados de teste são comparados com métricas NR e FR recentes através de coeficientes de correlação, obtendo-se resultados favoráveis para as métricas propostas. / This dissertation proposes two objective metrics for estimating human perception of quality for video subject to transmission degradation over packet networks. The first metric just uses traffic data while the second one uses both the degraded and the reference video sequences. That is, the latter is a full reference (FR) metric called Quadratic Combinational Metric (QCM) and the former one is a no reference (NR) metric called Viewing Quality Objective Metric (VQOM). In particular, the design procedure is applied to packet delay variation (PDV) impairments, whose compensation or control is very important to maintain quality. The NR metric is described by a cubic spline composed of two cubic polynomials that meet smoothly at a point called a knot. As the first step in the design of either metric, the spectators score a training set of degraded video sequences. The objective function for designing the NR metric includes the total square error between the scores and their parametric estimates, still regarded as algebraic expressions. In addition, the objective function is augmented by the addition of three equality constraints for the derivatives at the knot, whose position is specified within a fine grid of points between the minimum value and the maximum value of the degradation factor. These constraints are affected by Lagrange multipliers and added to the objective function to obtain the Lagrangian, which is minimized by the suboptimal polynomial coefficients determined as a function of each knot in the grid. Finally, the knot value is selected that yields the minimum square error. By means of the selected knot value, the final values of the polynomial coefficients are determined. On the other hand, the FR metric is a nonlinear combination of two popular metrics, namely, the Peak Signal-to-Noise Ratio (PSNR) and the Structural Similarity Index (SSIM). A complete second-degree two-variable polynomial is used for the combination since it is sensitive to both constituent metrics while avoiding overfitting. In the training phase, the set of values for the coefficients of this polynomial is determined by minimizing the mean square error to the opinions over the training database. Both metrics, the VQOM and the QCM, are trained and validated using one database and tested with a different one. The test results are compared with recent NR and FR metrics by means of correlation coefficients, obtaining favorable results for the proposed metrics.

Esta Tese propõe duas métricas objetivas para avaliar a percepção de qualidade de vídeos sujeitos a degradações de transmissão em uma rede de pacotes. A primeira métrica usa apenas o vídeo degradado, enquanto que a segunda usa os vídeos de referência e degradado. Esta última é uma métrica de referência completa (FR - Full Reference) chamada de QCM (Quadratic Combinational Metric) e a primeira é uma métrica sem referência (NR - No Reference) chamada de VQOM (Viewing Quality Objective Metric). Em particular, o procedimento de projeto é aplicado à degradação de variação de atraso de pacotes (PDV - Packet Delay Variation). A métrica NR é descrita por uma spline cúbica composta por dois polinômios cúbicos que se encontram suavemente num ponto chamado de nó. Para o projeto de ambas métricas, colhem-se opiniões de observadores a respeito das sequências de vídeo degradadas que compõem o conjunto. A função objetiva inclui o erro quadrático total entre as opiniões e suas estimativas paramétricas, ainda consideradas como expressões algébricas. Acrescentam-se à função objetiva três condições de igualdades de derivadas tomadas no nó, cuja posição é especificada dentro de uma grade fina de pontos entre o valor mínimo e o valor máximo do fator de degradação. Essas condições são afetadas por multiplicadores de Lagrange e adicionadas à função objetiva, obtendo-se o lagrangiano, que é minimizado pela determinação dos coeficientes subótimos dos polinômios em função de cada valor do nó na grade. Finalmente escolhe-se o valor do nó que produz o erro quadrático mínimo, determinando assim os valores finais para dos coeficientes do polinômio. Por outro lado, a métrica FR é uma combinação não-linear de duas métricas populares, a PSNR (Peak Signal-to-Noise Ratio) e a SSIM (Structural Similarity Index). Um polinômio completo de segundo grau de duas variáveis é usado para realizar a combinação, porque é sensível a ambas métricas constituintes, evitando o sobreajuste em decorrência do baixo grau. Na fase de treinamento, o conjunto de valores dos coeficientes do polinômio é determinado através da minimização do erro quadrático médio para as opiniões sobre a base de dados de treino. Ambas métricas, a VQOM e a QCM, são treinadas e validadas usando uma base de dados, e testadas com outra independente. Os resultados de teste são comparados com métricas NR e FR recentes através de coeficientes de correlação, obtendo-se resultados favoráveis para as métricas propostas. / This dissertation proposes two objective metrics for estimating human perception of quality for video subject to transmission degradation over packet networks. The first metric just uses traffic data while the second one uses both the degraded and the reference video sequences. That is, the latter is a full reference (FR) metric called Quadratic Combinational Metric (QCM) and the former one is a no reference (NR) metric called Viewing Quality Objective Metric (VQOM). In particular, the design procedure is applied to packet delay variation (PDV) impairments, whose compensation or control is very important to maintain quality. The NR metric is described by a cubic spline composed of two cubic polynomials that meet smoothly at a point called a knot. As the first step in the design of either metric, the spectators score a training set of degraded video sequences. The objective function for designing the NR metric includes the total square error between the scores and their parametric estimates, still regarded as algebraic expressions. In addition, the objective function is augmented by the addition of three equality constraints for the derivatives at the knot, whose position is specified within a fine grid of points between the minimum value and the maximum value of the degradation factor. These constraints are affected by Lagrange multipliers and added to the objective function to obtain the Lagrangian, which is minimized by the suboptimal polynomial coefficients determined as a function of each knot in the grid. Finally, the knot value is selected that yields the minimum square error. By means of the selected knot value, the final values of the polynomial coefficients are determined. On the other hand, the FR metric is a nonlinear combination of two popular metrics, namely, the Peak Signal-to-Noise Ratio (PSNR) and the Structural Similarity Index (SSIM). A complete second-degree two-variable polynomial is used for the combination since it is sensitive to both constituent metrics while avoiding overfitting. In the training phase, the set of values for the coefficients of this polynomial is determined by minimizing the mean square error to the opinions over the training database. Both metrics, the VQOM and the QCM, are trained and validated using one database and tested with a different one. The test results are compared with recent NR and FR metrics by means of correlation coefficients, obtaining favorable results for the proposed metrics.