This thesis describes the application of parallel simulation techniques to represent the structured functional parallelism present within the Space Shuttle Operations Flow using the Synchronous Parallel Environment for Emulation and Discrete-Event Simulation (SPEEDES), an object-oriented multi-computing architecture. SPEEDES is a unified parallel simulation environment, which allocates events over multiple processors to get simulation speed up. Its optimistic processing capability minimizes simulation lag time behind wall clock time, or multiples of real-time. SPEEDES accommodates an increase in process complexity with additional parallel computing nodes to allow sharing of processing loads. This thesis focuses on the process of translating a model of Space Shuttle Operations from a procedural oriented and single processor approach to one represented in a process-driven, object-oriented, and distributed processor approach. The processes are depicted by several classes created to represent the operations at the space center. The reference model used is the existing Space Shuttle Model created in ARENA by NASA and UCF in the year 2001. A systematic approach was used for this translation. A reduced version of the ARENA model was created, and then used as the SPEEDES prototype using C++. The prototype was systematically augmented to reflect the entire Space Shuttle Operations Flow. It was then verified, validated, and implemented.

Distributed generation (defined as the production of power in small quantities at the point of use) has recently gained significant interest due to its benefits over a centralized approach. This thesis investigates the integration of a natural gas fed solid-oxide fuel cell (SOFC) and compressed air energy storage (CAES) technologies for distributed generation at the building-level scale. The SOFC/CAES system is also integrated with multiple vital sub-systems (including on-site solar panels) for the building to provide the heat, through an in-floor heating system, hot water, and power demanded by the building. This thesis investigates the models for the SOFC/CAES system, and implements them in a generic analysis tool providing a means for rapid analysis of a wide variety of case studies. The analysis tool determines the ability of the SOFC/CAES system to follow the power and heat loads demanded by the building, and evaluates its performance with an assortment of metrics, including efficiencies, CO2 emissions and grid-independence. The SOFC/CAES system was investigated for the new ExCEL building at McMaster University. It was found that the system was able to produce upwards 75% of the heat and hot water demand, and upwards of 94% of the power demand of the building. When compared to the current state-of-the-art natural gas based power producing technology and high efficiency furnace, the SOFC/CAES system reduces the CO2 emissions associated with the building by a minimum of 8.7% and a maximum of 26.95%. The cost of electricity for the system is significantly (21% to 150%) more costly than current market prices; however the SOFC/CAES system is the least costly of all other distributed generation technologies investigated for the case of the ExCEL building. / Thesis / Master of Applied Science (MASc)

Distributed Approximation / The classic vehicle routing problem (VRP) is generally concerned with the optimal design of routes by a fleet of vehicles to service a set of customers by minimizing the overall cost, usually the travel distance for the whole set of routes. Although the problem has been extensively studied in the context of operations research and optimization, there is little research on solving the VRP, where distributed vehicles need to compute their respective routes in a decentralized fashion. Our first contribution is a synchronous distributed approximation algorithm that solves the VRP. Using the duality theorem of linear programming, we show that the approximation ratio of our algorithm is $O(n . (\rho)^{1/n} .log(n+m))$, where $\rho$ is the maximum cost of travel or service in the input VRP instance, $n$ is the size of the graph, and $m$ is the number of vehicles. We report results of simulations comparing our algorithm results with ILP solutions and a greedy algorithm. / Thesis / Master of Science (MSc) / The Open Multi-Depot Vehicle Routing Problem(OMDVRP) problem is solved using an synchronous distributed algorithm and the approximation ratio is found and simulation results comparing the performance of ILP , greedy and the designed algorithm is done.

No description available.

No description available.

No description available.

Supercomputing application simulations on hundreds of thousands of cores produce vast amounts of data that need to be analyzed on smaller-scale clusters to glean insights. The process is referred to as an end-to-end workflow. Extant workflow systems are stymied by the storage wall, resulting from both the disk-based parallel file system (PFS) failing to keep pace with the compute and memory subsystems as well as the inefficiencies in end-to-end workflow processing. In the post-petaflop era, supercomputers are provisioned with flash devices, as an intermediary between compute nodes and the PFS, enabling novel paradigms not just for expediting I/O, but also for the in-situ analysis of the simulation output data on the flash device. An array of such active flash elements allows us to fundamentally rethink the way data analysis workflows interact with storage systems. By blending the flash storage array and data analysis together in a seamless fashion, we create an analysis workflow-aware storage system, AnalyzeThis. Our guiding principle is that analysis-awareness be deeply ingrained in each and every layer of the storage system—active flash fabric, analysis object abstraction layer, scheduling layer within the storage, and an easy-to-use file system interface—thereby elevating data analyses as first-class citizens. Together, these concepts transform AnalyzeThis into a potent analytics-aware appliance. / Master of Science

Cylindrical shells are used for a variety of engineering applications such as undersea vehicles and aircraft. The models currently used to determine the vibration characteristics of these shells are often approximated by assuming the shell is infinitely long or has shear-diaphragm boundary conditions. These models also ignore complex loading conditions such as plane waves in favor of point forces or free vibration models. This work expands on the capabilities of these models by examining the acoustic response of a finite length cylinder with flat plate endcaps to multiple types of distributed loading conditions. Starting with the Donnell equations of motion for thin cylinders and the classical plate theory equations of motion for the endcaps, spacial domain displacement field solutions for the shell and plates take an assumed form that includes unknown wave propagation coefficients. These solutions are substituted into stress boundary conditions and continuity equations evaluated at the intersections between the shell and plates. An infinite summation is contained within the boundary conditions and continuity equations which is decoupled, truncated, and compiled in matrix form to allow for the propagation coefficients to be found via a convergent sum of vectors. System responses due to a ring loading and multiple cases of plane waves are studied and validated using a finite element analysis of the system. It is shown that the analytical model matches the finite element model well. / Master of Science

Doctor of Philosophy / Department of Computing and Information Sciences / Gurdip Singh / Many frameworks have been proposed which provide distributed algorithms encapsulated as middleware services to simplify application design. The developers of such algorithms are faced with two opposing forces. One is to design generic algorithms that are reusable in a large number of applications. Efficiency considerations, on the other hand, force the algorithms to be customized to specific operational contexts. This problem is often attacked by simply re-implementing all or large portions of an algorithm. We propose InDiGO, an infrastructure which allows design of generic but customizable algorithms and provides tools to customize such algorithms for specific applications. InDiGO provides the following capabilities: (a) Tools to generate intermediate representations of an application which can be leveraged for analysis, (b) Mechanisms to allow developers to design customizable algorithms by exposing design knowledge in terms of configurable options, and (c) An optimization engine to analyze an application to derive the information necessary to optimize the algorithms. Specifically, we optimize algorithms by removing communication which is redundant in the context of a specific application. We perform three types of optimizations: static optimization, dynamic optimization and physical topology-based optimization. We present experimental results to demonstrate the advantages of our infrastructure.

Graph databases offer an efficient way to store and access inter-connected data. However, to query large graphs that no longer fit in memory, it becomes necessary to make multiple trips to the storage device to filter and gather data based on the query. But I/O accesses are expensive operations and immensely slow down query response time and prevent us from fully exploiting the graph specific benefits that graph databases offer. The storage models of most existing graph database systems view graphs as indivisible structures and hence do not allow a hierarchical layering of the graph. This adversely affects query performance for large graphs as there is no way to filter the graph on a higher level without actually accessing the entire information from the disk. Distributing the storage and processing is one way to extract better performance. But current distributed solutions to this problem are not entirely effective, again due to the indivisible representation of graphs adopted in the storage format. This causes unnecessary latency due to increased inter-processor communication. In this dissertation, we propose an optimized distributed graph storage system for scalable and faster querying of big graph data. We start with our unique physical storage model, in which the graph is decomposed into three different levels of abstraction, each with a different storage hierarchy. We use a hybrid storage model to store the most critical component and restrict the I/O trips to only when absolutely necessary. This lets us actively make use of multi-level filters while querying, without the need of comprehensive indexes. Our results show that our system outperforms established graph databases for several class of queries. We show that this separation also eases the difficulties in distributing graph data and go on propose a more efficient distributed model for querying general purpose graph data using the Spark framework.