No / The traffic pattern has significant impact on the performance of network-on-chip. Many recent studies have shown that multimedia applications can be supported in on-chip interconnects. Driven by the motivation of evaluating on-chip interconnects in multimedia embedded systems, a new analytical model is proposed to investigate the performance of the fat-tree based on-chip interconnection network under bursty multimedia traffic and nonuniform message destinations. Extensive simulation experiments are conducted to validate the accuracy of the model, which is then adopted as a cost-efficient tool to investigate the effects of bursty multimedia traffic with nonuniform destinations on the network performance.

Yes / The implementation of Reverse Osmosis (RO) technology is noticeably increased to produce freshwater from brackish and seawater resources. In this work, performance analysis of a multistage multi pass medium-sized spiral wound brackish water RO (BWRO) desalination plant (1200 m³/day) of Arab Potash Company (APC) located in Jordan is evaluated using modelling and simulation. For this purpose, a mathematical model for the spiral wound RO process based on the principles of solution diffusion model is developed. The model is then used to simulate the operating conditions of low-salinity brackish water RO (BWRO) desalination plant. The results obtained are then compared against the real industrial data of BWRO desalination plant of APC which shows a high-level of consistency. Finally, the model is used to analysis the impact of the operating parameters such as salinity, pressure, temperature, and flow rate on the plant performance. The sensitivity analysis confirms that both feed flow rate and operating pressure as the critical parameters that positively affect the product salinity.

Systems controlled by embedded computers become indispensable in our lives and can be found in avionics, automotive industry, home appliances, medicine, telecommunication industry, mecatronics, space industry, etc. Fast, accurate and flexible performance estimation tools giving feedback to the designer in every design phase are a vital part of a design process capable to produce high quality designs of such embedded systems. In the past decade, the limitations of models considering fixed task execution times have been acknowledged for large application classes within soft real-time systems. A more realistic model considers the tasks having varying execution times with given probability distributions. No restriction has been imposed in this thesis on the particular type of these functions. Considering such a model, with specified task execution time probability distribution functions, an important performance indicator of the system is the expected deadline miss ratio of tasks or task graphs. This thesis proposes two approaches for obtaining this indicator in an analytic way. The first is an exact one while the second approach provides an approximate solution trading accuracy for analysis speed. While the first approach can efficiently be applied to monoprocessor systems, it can handle only very small multi-processor applications because of complexity reasons. The second approach, however, can successfully handle realistic multiprocessor applications. Experiments show the efficiency of the proposed techniques. / <p>Report code: LiU-Tek-Lic-2002:58.</p>

In this study the maximum entropy formalism [JAYN 57] is suggested as an alternative theory for general queueing systems of computer performance analysis. The motivation is to overcome some of the problems arising in this field and to extend the scope of the results derived in the context of Markovian queueing theory. For the M/G/l model a unique maximum entropy solution., satisfying locALl balance is derived independent of any assumptions about the service time distribution. However, it is shown that this solution is identical to the steady state solution of the underlying Marko-v process when the service time distribution is of the generalised exponential (CE) type. (The GE-type distribution is a mixture of an exponential term and a unit impulse function at the origin). For the G/M/1 the maximum entropy solution is identical in form to that of the underlying Markov process, but a GE-type distribution still produces the maximum overall similar distributions. For the GIG11 model there are three main achievements: first, the spectral methods are extended to give exaft formulae for the average number of customers in the system for any G/G/l with rational Laplace transform. Previously, these results are obtainable only through simulation and approximation methods. (ii) secondly, a maximum entropy model is developed and used to obtain unique solutions for some types of the G/G/l. It is also discussed how these solutions can be related to the corresponding stochastic processes. (iii) the importance of the G/GE/l and the GE/GE/l for the analysis of general networks is discussed and some flow processes for these systems are characterised. For general queueing networks it is shown that the maximum entropy solution is a product of the maximum entropy solutions of the individual nodes. Accordingly, existing computational algorithms are extended to cover general networks with FCFS disciplines. Some implementations are suggested and a flow algorithm is derived. Finally, these results are iised to improve existing aggregation methods. In addition, the study includes a number of examples, comparisons, surveys, useful comments and conclusions.

Decomposition methods based on the hierarchical partitioning of the state space of queueing network models offer powerful evaluation tools for the performance analysis of computer systems and communication networks. These methods being conventionally implemented capture the exact solution of separable queueing network models but their credibility differs when applied to general queueing networks. This thesis provides a universal information theoretic framework for the implementation of hierarchical decomposition schemes, based on the principle of minimum relative entropy given fully decomposable subset and aggregate utilization, mean queue length and flow-balance constraints. This principle is used, in conjuction with asymptotic connections to infinite capacity queues, to derive new closed form approximations for the conditional and marginal state probabilities of general queueing network models. The minimum relative entropy solutions are implemented iteratively at each decomposition level involving the generalized exponential (GE) distributional model in approximating the general service and asymptotic flow processes in the network. It is shown that the minimum relative entropy joint state probability, subject to mean queue length and flow-balance constraints, is identical to the exact product-form solution obtained as if the network was separable. An investigation into the effect of different couplings of the resource units on the relative accuracy of the approximation is carried out, based on an extensive experimentation. The credibility of the method is demonstrated with some illustrative examples involving first-come-first-served general queueing networks with single and multiple servers and favourable comparisons against exact solutions and other approximations are made.

Analytical modelling and characterization of Weighted Fair Queueing (WFQ) have recently received considerable attention by several researches since WFQ offers the minimum delay and optimal fairness guarantee. However, all previous work on WFQ has focused on developing approximations of the scheduler with an infinite buffer because of supposed scalability problems in the WFQ computation. The main aims of this thesis are to study WFQ system, by providing an analytical WFQ model which is a theoretical construct based on a form of processor sharing for finite capacity. Furthermore, the solutions for classes with Poisson arrivals and exponential service are derived and verified against global balance solution. This thesis shows that the analytical models proposed can give very good results under particular conditions which are very close to WFQ algorithms, where accuracy of the models is verified by simulations of WFQ model. Simulations were performed with QNAP-2 simulator. In addition, the thesis presents several performance studies signifying the power of the proposed analytical model in providing an accurate delay bounds to a large number of classes. These results are not able to cover all unsolved issues in the WFQ system. They represent a starting point for the research activities that the Author will conduct in the future. The author believes that the most promising research activities exist in the scheduler method to provide statistical guarantees to multi-class services. The author is convinced that alternative software, for example, on the three class model buffer case, is able to satisfy the large number of buffer because of the software limitation in this thesis. While they can be a good topic for long-term research, the short-medium term will show an increasing interest in the modification of the WFQ models to provide differentiated services. / Ministry of Higher Education

Sensor nodes are deployed over sensing fields for the purpose of monitoring certain phenomena of interest. The sensor nodes perform specific measurements, process the sensed data, and send the data to a base station over a wireless channel. The base station collects data from the sensor nodes, analyses this data, and reports it to the users. Wireless sensor networks are different from traditional networks, because of the following constraints. Typically, a large number of sensor nodes need to be randomly deployed and, in most cases, they are deployed in unreachable environments; however, the sensor nodes may fail, and they are subject to power constraints. Energy is one of the most important design constraints of wireless sensor networks. Energy consumption, in a sensor node, occurs due to many factors, such as: sensing the environment, transmitting and receiving data, processing data, and communication overheads. Since the sensor nodes behave as router nodes for data propagation, of the other sensor nodes to the base station, network connectivity decreases gradually. This may result in disconnected sub networks of sensor nodes. In order to prolong the network¿s lifetime, energy efficient protocols should be designed for the characteristics of the wireless sensor network. Sensor nodes in different regions of the sensing field can collaborate to aggregate the data that they gathered. Data aggregation is defined as the process of aggregating the data from sensor nodes to reduce redundant transmissions. It reduces a large amount of the data traffic on the network, it requires less energy, and it avoids information overheads by not sending all of the unprocessed data throughout the sensor network. Grouping sensor nodes into clusters is useful because it reduces the energy consumption. The clustering technique can be used to perform data aggregation. The clustering procedure involves the selection of cluster heads in each of the cluster, in order to coordinate the member nodes. The cluster head is responsible for: gathering the sensed data from its cluster¿s nodes, aggregating the data, and then sending the aggregated data to the base station. An adaptive clustering protocol was introduced to select the heads in the wireless sensor network. The proposed clustering protocol will dynamically change the cluster heads to obtain the best possible performance, based on the remaining energy level of sensor nodes and the average energy of clusters. The OMNET simulator will be used to present the design and implementation of the adaptive clustering protocol and then to evaluate it. This research has conducted extensive simulation experiments, in order to fully study and analyse the proposed energy efficient clustering protocol. It is necessary for all of the sensor nodes to remain alive for as long as possible, since network quality decreases as soon as a set of sensor nodes die. The goal of the energy efficient clustering protocol is to increase the lifetime and stability period of the sensor network. This research also introduces a new bidirectional data gathering protocol. This protocol aims to form a bidirectional ring structure among the sensor nodes, within the cluster, in order to reduce the overall energy consumption and enhance the network¿s lifetime. A bidirectional data gathering protocol uses a source node to transmit data to the base station, via one or more multiple intermediate cluster heads. It sends data through energy efficient paths to ensure the total energy, needed to route the data, is kept to a minimum. Performance results reveal that the proposed protocol is better in terms of: its network lifetime, energy dissipation, and communication overheads.

Mobile Ad hoc Networks (MANETs) are of great interest to researchers and have become very popular in the last few years. One of the great challenges is to provide a routing protocol that is capable of offering the shortest and most reliable path in a MANET in which users are moving continuously and have no base station to be used as a reference for their position. This thesis proposes some new routing protocols based on the angles (directions) of the adjacent mobile nodes and also the node density. In choosing the next node in forming a route, the neighbour node with the closest heading angle to that of the node of interest is selected, so the connection between the source and the destination consists of a series of nodes that are moving in approximately the same direction. The rationale behind this concept is to maintain the connection between the nodes as long as possible. This is in contrast to the well known hop count method, which does not consider the connection lifetime. We propose three enhancements and modifications of the Ad-hoc on demand distance vector (AODV) protocol that can find a suitable path between source and destination using combinations and prioritization of angle direction and hop count. Firstly, we consider that if there are multiple routing paths available, the path with the minimum hop count is selected and when the hop counts are the same the path with the best angle direction is selected. Secondly, if multiple routing paths are available the paths with the best angle direction are chosen but if the angles are the same (fall within the same specified segment), the path with minimum hop count is chosen. Thirdly, if there is more than one path available, we calculate the average of all the heading angles in every path and find the best one (lowest average) from the source to the destination. In MANETs, flooding is a popular message broadcasting technique so we also propose a new scheme for MANETS where the value of the rebroadcast packets for every host node is dynamically adjusted according to the number of its neighbouring nodes. A fixed probabilistic scheme algorithm that can dynamically adjust the rebroadcasting probability at a given node according to its ID is also proposed; Fixed probabilistic schemes are one of the solutions to reduce rebroadcasts and so alleviate the broadcast storm problem. Performance evaluation of the proposed schemes is conducted using the Global Mobile Information System (GloMoSim) network simulator and varying a number of important MANET parameters, including node speed, node density, number of nodes and number of packets, all using a Random Waypoint (RWP) mobility model. Finally, we measure and compare the performance of all the proposed approaches by evaluating them against the standard AODV routing protocol. The simulation results reveal that the proposed approaches give relatively comparable overall performance but which is better than AODV for almost all performance measures and scenarios examined.

The InfiniBand Architecture (IBA) network has been proposed as a new industrial standard with high-bandwidth and low-latency suitable for constructing high-performance interconnected cluster computing systems. This architecture replaces the traditional bus-based interconnection with a switch-based network for the server Input-Output (I/O) and inter-processor communications. The efficient Quality-of-Service (QoS) mechanism is fundamental to ensure the import at QoS metrics, such as maximum throughput and minimum latency, leaving aside other aspects like guarantee to reduce the delay, blocking probability, and mean queue length, etc. Performance modelling and analysis has been and continues to be of great theoretical and practical importance in the design and development of communication networks. This thesis aims to investigate efficient and cost-effective QoS mechanisms for performance analysis and improvement of InfiniBand networks in cluster-based computing systems. Firstly, a rate-based source-response link-by-link admission and congestion control function with improved Explicit Congestion Notification (ECN) packet marking scheme is developed. This function adopts the rate control to reduce congestion of multiple-class traffic. Secondly, a credit-based flow control scheme is presented to reduce the mean queue length, throughput and response time of the system. In order to evaluate the performance of this scheme, a new queueing network model is developed. Theoretical analysis and simulation experiments show that these two schemes are quite effective and suitable for InfiniBand networks. Finally, to obtain a thorough and deep understanding of the performance attributes of InfiniBand Architecture network, two efficient threshold function flow control mechanisms are proposed to enhance the QoS of InfiniBand networks; one is Entry Threshold that sets the threshold for each entry in the arbitration table, and other is Arrival Job Threshold that sets the threshold based on the number of jobs in each Virtual Lane. Furthermore, the principle of Maximum Entropy is adopted to analyse these two new mechanisms with the Generalized Exponential (GE)-Type distribution for modelling the inter-arrival times and service times of the input traffic. Extensive simulation experiments are conducted to validate the accuracy of the analytical models.

Indiana University-Purdue University Indianapolis (IUPUI) / Biomechanical Big Data is of great significance to precision health applications, among which we take special interest in Physical Activity Detection (PAD). In this study, we have performed extensive research on deep learning-based PAD from biomechanical big data, focusing on the challenges raised by the need for real-time edge inference. First, considering there are many places we can place the motion sensors, we have thoroughly compared and analyzed the location difference in terms of deep learning-based PAD performance. We have further compared the difference among six sensor channels (3-axis accelerometer and 3-axis gyroscope). Second, we have selected the optimal sensor and the optimal sensor channel, which can not only provide sensor usage suggestions but also enable ultra-lowpower application on the edge. Third, we have investigated innovative methods to minimize the training effort of the deep learning model, leveraging the transfer learning strategy. More specifically, we propose to pre-train a transferable deep learning model using the data from other subjects and then fine-tune the model using limited data from the target-user. In such a way, we have found that, for single-channel case, the transfer learning can effectively increase the deep model performance even when the fine-tuning effort is very small. This research, demonstrated by comprehensive experimental evaluation, has shown the potential of ultra-low-power PAD with minimized sensor stream, and minimized training effort. / 2023-06-01