No description available.

Typescript (photocopy). / Digitized by Kansas State University Libraries

Call number: LD2668 .T4 1979 W37 / Master of Landscape Architecture

Continued dimensional scaling of CMOS processes is approaching fundamental limits and therefore, alternate new devices and microarchitectures are explored to address the growing need of area scaling and performance gain. New nanotechnologies, such as memristors, emerge. Memristors can be used to perform stateful logic with nanowire crossbars, which allows for implementation of very large binary networks that can be easily reconfigured. This research involves the design of a memristor-based massively parallel datapath for various applications, specifically SIMD (Single Instruction Multiple Data) like architecture, and parallel pipelines. The dissertation develops a new model of massively parallel memristor-CMOS hybrid datapath architectures at a systems level, as well as a complete methodology to design them. One innovation of the proposed approach is that the datapath design is based on space-time diagrams that use stateful IMPLY gates built from binary memristors. This notation aids in the circuit minimization in logic design, calculations of delay and memristor costs, and sneak-path avoidance. Another innovation of the proposed methodology is a general, new, architecture model, MsFSMD (Memristive stateful Finite State Machine with Datapath) that has two interacting sub-systems: 1) a controller composed of a memristive RAM, MsRAM, to act as a pulse generator, along with a finite state machine realized in CMOS, a CMOS counter, CMOS multiplexers and CMOS decoders, 2) massively parallel, pipelined, datapath realized with a new variant of a CMOL-like nanowire crossbar array, MsCMOL (Memristive stateful CMOL), with binary stateful memristor-based IMPLY gates. Next contribution of the dissertation is the new type of FPGA. In contrast to the previous memristor-based FPGA (mrFPGA), the proposed MsFPGA (Memristive stateful logic Field Programmable Gate Array) uses memristors for memory, connections programming, and combinational logic implementation. With a regular structure of square abutting blocks of memristive nanowire crossbars and their short connections, proposed architecture is highly reconfigurable. As an example of using the proposed new FPGA to realize biologically inspired systems, the detailed design of a pipelined Euclidean Distance processor was presented and its various applications are mentioned. Euclidean Distance calculation is widely used by many neural network and associative memory based algorithms.

Calculating a sparse code for signals with high dimensionality, such as high-resolution images, takes substantial time to compute on a traditional computer architecture. Memristors present the opportunity to combine storage and computing elements into a single, compact device, drastically reducing the area required to perform these calculations. This work focused on the analysis of two existing sparse coding architectures, one of which utilizes memristors, as well as the design of a new, third architecture that employs a memristive crossbar. These architectures implement either a non-spiking or spiking variety of sparse coding based on the Locally Competitive Algorithm (LCA) introduced by Rozell et al. in 2008. Each architecture receives an arbitrary number of input lines and drives an arbitrary number of output lines. Training of the dictionary used for the sparse code was implemented through external control signals that approximate Oja's rule. The resulting designs were capable of representing input in real-time: no resets would be needed between frames of a video, for instance, though some settle time would be needed. The spiking architecture proposed is novel, emphasizing simplicity to achieve lower power than existing designs. The architectures presented were tested for their ability to encode and reconstruct 8 x 8 patches of natural images. The proposed network reconstructed patches with a normalized, root-mean-square error of 0.13, while a more complicated CMOS-only approach yielded 0.095, and a non-spiking approach yielded 0.074. Several outputs competing for representation of the input was shown to improve reconstruction quality and preserve more subtle components in the final encoding; the proposed algorithm lacks this feature. Steps to address this were proposed for future work by scaling input spikes according to the current expected residual, without adding much complexity. The architectures were also tested with the MNIST digit database, passing a sparse code onto a basic classifier. The proposed architecture scored 81% on this test, a CMOS-only spiking variant scored 76%, and the non-spiking algorithm scored 85%. Power calculations were made for each design and compared against other publications. The overall findings showed great promise for spiking memristor-based ASICs, consuming only 28% of the power used by non-spiking architectures and 6.6% as much power as a CMOS-only spiking architecture on this task. The spike-based nature of the novel design was also parameterized into several intuitive parameters that could be adjusted to prefer either performance or power efficiency. The design and analysis of architectures for sparse coding should greatly reduce the amount of future work needed to implement an end-to-end classification pipeline for images or other signal data. When lower power is a primary concern, the proposed architecture should be considered as it surpassed other published algorithms. These pipelines could be used to provide low-power visual assistance, highlighting objects within high-definition video frames in real-time. The technology could also be used to help self-driving cars identify hazards more quickly and efficiently.

A majority of software development today involves maintenance or evolution of legacy systems. Evolving these legacy systems, while maintaining good software design principles, is a significant challenge. Research has shown the benefits of using software architecture as an abstraction to analyze quality attributes of proposed designs. Unfortunately, for most legacy systems, a documented software architecture does not exist. Developing a good architectural description frequently requires extensive experience on the part of the developer trying to recover the legacy system's architecture. This work first describes a four-phase process that provides a framework within which architectural recovery activities can be automated. These phases consist of: extraction (obtaining a subset of information about the legacy system from a single source), classification (partitioning the information based upon its viewpoint), union (combining all the information in a particular viewpoint into a candidate view), and fusion (cross-checking all candidate views for consistency. The work then concentrates on the major problem facing automated architectural recovery---the concept assignment problem. To overcome this problem, a technique called semantic approximation is presented and validated via experimental results. Semantic approximation uses a combination of text data mining and a mathematical technique called concept analysis to build a lattice of similar concepts between higher-level domain information and low-level code concepts. The experimental data reveals that while semantic approximation does improve results over the more traditional lexical and topological approaches, it does not yet fully solve the concept assignment problem.

A software architecture is one of the first steps towards a software system. A software architecture can be designed in different ways. During the design phase, it is important to select the most suitable design of the architecture, in order to create a good foundation for the system. The selection process is performed by evaluating architecture alternatives against each other. We investigate the use of continuous simulation of a software architecture as a support tool for architecture evaluation. For this purpose, we study a software architecture of an existing software system in an experiment, where we create a model of it using a tool for continuous simulation, and simulate the model. Based on the results from the simulation, we conclude that the system is too complex to be modeled for continuous simulation. Problems we identify are that we need discrete functionality to be able to correctly simulate the system, and that it is very time-consuming to develop a model for evaluation purposes. Thus, we find that continuous simulation is not appropriate for evaluating a software architecture, but that the modeling process is a valuable tool for increasing knowledge and understanding about an architecture.

The purpose of this thesis is to characterize, compare and contrast four network-centric software architectures, namely Client-Server Architecture (CSA), Distributed Objects Architecture (DOA), Service-Oriented Architecture (SOA) and Peer-to-Peer Architecture (PPA) and seven associated frameworks consisting of .NET, Java EE, CORBA, DCOM, Web Services, Jini and JXTA with respect to a set of derived criteria. Network-centric systems are gaining in popularity as they have the potential to solve more complex problems than we have been able to in the past. However, with the rise of SOA, Web Services, a set of standards widely used for implementing service-oriented solutions, is being touted as the "silver bullet" to all problems afflicting the software engineering domain with the danger of making other architectures seem obsolete. Thus, there is an urgent need to study the various architectures and frameworks in comparison to each other and understand their relative merits and demerits for building network-centric systems. The architectures studied here were selected on the basis of their fundamentality and generality. The frameworks were chosen on the basis of their popularity and representativeness to build solutions in a particular architecture. The criteria used for comparative assessment are derived from a combination of two approaches — by a close examination of the unique characteristics and requirements of network-centric systems and then by an examination of the constraints and mechanisms present in the architectures and frameworks under consideration that may contribute towards realizing the requirements of network-centric systems. Not all of the criteria are equally relevant for the architectures and frameworks. Some, when relevant, are relevant in a different sense from one architecture (or framework) to another. One of the conclusions that can be drawn from this study is that the different architectures are not completely different from each other. In fact, CSA, DOA and SOA are a natural evolution in that order and share several characteristics. At the same time, significant differences do exist, so it is clearly possible to judge/differentiate one from the other. All three architectures can coexist in a single system or system of systems. However, the advantages of each architecture become apparent only when they are used in their proper scope. At the same time, a sharp difference can be perceived between these three architectures and the peer-to-peer architecture. This is because PPA aims to solve a totally different class of problems than the other three architectures and hence has certain unique characteristics not observed in the others. Further, all of the frameworks have certain unique architectural features and mechanisms not found in the others that contribute towards achieving network-centric quality characteristics. The two broad frameworks, .NET and Java EE offer almost equivalent capabilities and features; what can be achieved in one can be achieved in the other. This thesis deals with the study of all the four architectures and their related frameworks. The criteria used, while fairly comprehensive, are not exhaustive. Variants of the fundamental architectures are not considered. However, system/software architects seeking an understanding of the tradeoffs involved in using the various architectures and frameworks and their subtle nuances should benefit considerably from this work. / Master of Science

Over the last years, online learning has gained significant popularity, and this trend is expected to continue in the future. Online learning is a multi-dimensional activity where each dimension should be adequately supported by an online learning system to provide a fruitful learning experience to its users. Similar to online learning, a particular approach in the way systems are designed has gained increasing popularity over the last years as a result of the emergence of the computer network. Systems have started to be envisioned as integrated collections of components that are deployed over the network. This network-centric approach addresses some of the issues that are not directly addressed by the traditional architectural design approaches. We believe that such an approach can be employed for the architectural design of online learning environments (OLEs) where different dimensions of online learning are addressed by different components residing on the network. This thesis presents a network-centric system architecture for OLEs. The architecture is based on the Service-Oriented Architecture (SOA) paradigm and web services. The proposed architecture is described using a number of Department of Defense Architecture Framework (DoDAF) representations. We also developed a quality model for OLEs and evaluated the architecture using this quality model. We believe that this research is an important first step in reconciling those two emerging ideas, namely online learning and network-centric architecture. This reconciliation is important as the flexible nature of SOA can be very useful in creating OLEs that address a wide variety of user needs. / Master of Science

Architecting e-business applications is a complex design activity. This is mainly due to the numerous architectural decisions to be made, including the selection of alternative technologies, software components, design strategies, patterns, standards, protocols, platforms, etc. Further complexities arise due to the fact that these alternatives often vary considerably in their support for different quality attributes. Moreover, there are often different groups of stakeholders involved, with each having their own quality goals and criteria. Furthermore, different architectural decisions often include interdependent alternatives, where the selection of one alternative for one particular decision impacts the selections to be made for alternatives from other different decisions. There have been several research efforts aiming at providing sufficient mechanisms and tools for facilitating the architectural evaluation and design process. These approaches, however, address architectural decisions in isolation, where they focus on evaluating a limited set of alternatives belonging to one architectural decision. This has been the primary motivation behind the development of the Architectural DEcision-Making Support (ADEMS) framework, which basically aims at supporting stakeholders and architects during the architectural decision-making process by helping them determining a suitable combination of architectural alternatives. ADEMS framework is an iterative process that leverages rigorous quantitative decision-making techniques available in the literature of Management Science, particularly Multiple Attribute Decision-Making (MADM) methods and Integer Programming (IP). Furthermore, due to the number of architectural decisions involved as well as the variety of available alternatives, the architecture design space is expected to be huge. For this purpose, a query language has been developed, known as the Architecture Query Language (AQL), to aid architects in exploring and analyzing the design space in further depth, and also in examining different ???what-if??? architectural scenarios. In addition, in order to support leveraging ADEMS framework, a support tool has been implemented for carrying out the sophisticated set of mathematical computations and comparisons of the large number of architectural combinations, which might otherwise be hard to conduct using manual techniques. The primary contribution of the tool is in its help to identify, evaluate, and rank all potential combinations of alternatives based on their satisfaction to quality preferences provided by the different stakeholders. Finally, to assess the feasibility of ADEMS, three different case studies have been conducted relating to the architectural evaluation of different e-business and enterprise applications. Results obtained for the three case studies were quite positive as they showed an acceptable accuracy level for the decisions recommended by ADEMS, and at a reasonable time and effort costs for the different system stakeholders.