121 |
Information Centric Development of Component-Based Embedded Real-Time SystemsHjertström, Andreas January 2009 (has links)
This thesis presents new techniques for data management of run-time data objectsin component-based embedded real-time systems. These techniques enabledata to be modeled, analyzed and structured to achieve data managementduring development, maintenance and execution.The evolution of real-time embedded systems has resulted in an increasedsystem complexity beyond what was thought possible just a few years ago.Over the years, new techniques and tools have been developed to manage softwareand communication complexity. However, as this thesis show, currenttechniques and tools for data management are not sufficient. Today, developmentof real-time embedded systems focuses on the function aspects of thesystem, in most cases disregarding data management.The lack of proper design-time data management often results in ineffectivedocumentation routines and poor overall system knowledge. Contemporarytechniques to manage run-time data do not satisfy demands on flexibility,maintainability and extensibility. Based on an industrial case-study that identifiesa number of problems within current data management techniques, bothduring design-time and run-time, it is clear that data management needs to beincorporated as an integral part of the development of the entire system architecture.As a remedy to the identified problems, we propose a design-time data entityapproach, where the importance of data in the system is elevated to beincluded in the entire design phase with proper documentation, properties, dependenciesand analysis methods to increase the overall system knowledge.Furthermore, to efficiently manage data during run-time, we introduce databaseproxies to enable the fusion between two existing techniques; ComponentBased Software Engineering (CBSE) and Real-Time Database ManagementSystems (RTDBMS). A database proxy allows components to be decoupledfrom the underlying data management strategy without violating the componentencapsulation and communication interface. / INCENSE
|
122 |
Energy-Aware Data Management on NUMA ArchitecturesKissinger, Thomas 29 May 2017 (has links) (PDF)
The ever-increasing need for more computing and data processing power demands for a continuous and rapid growth of power-hungry data center capacities all over the world. As a first study in 2008 revealed, energy consumption of such data centers is becoming a critical problem, since their power consumption is about to double every 5 years. However, a recently (2016) released follow-up study points out that this threatening trend was dramatically throttled within the past years, due to the increased energy efficiency actions taken by data center operators. Furthermore, the authors of the study emphasize that making and keeping data centers energy-efficient is a continuous task, because more and more computing power is demanded from the same or an even lower energy budget, and that this threatening energy consumption trend will resume as soon as energy efficiency research efforts and its market adoption are reduced. An important class of applications running in data centers are data management systems, which are a fundamental component of nearly every application stack. While those systems were traditionally designed as disk-based databases that are optimized for keeping disk accesses as low a possible, modern state-of-the-art database systems are main memory-centric and store the entire data pool in the main memory, which replaces the disk as main bottleneck. To scale up such in-memory database systems, non-uniform memory access (NUMA) hardware architectures are employed that face a decreased bandwidth and an increased latency when accessing remote memory compared to the local memory.
In this thesis, we investigate energy awareness aspects of large scale-up NUMA systems in the context of in-memory data management systems. To do so, we pick up the idea of a fine-grained data-oriented architecture and improve the concept in a way that it keeps pace with increased absolute performance numbers of a pure in-memory DBMS and scales up on NUMA systems in the large scale. To achieve this goal, we design and build ERIS, the first scale-up in-memory data management system that is designed from scratch to implement a data-oriented architecture. With the help of the ERIS platform, we explore our novel core concept for energy awareness, which is Energy Awareness by Adaptivity. The concept describes that software and especially database systems have to quickly respond to environmental changes (i.e., workload changes) by adapting themselves to enter a state of low energy consumption. We present the hierarchically organized Energy-Control Loop (ECL), which is a reactive control loop and provides two concrete implementations of our Energy Awareness by Adaptivity concept, namely the hardware-centric Resource Adaptivity and the software-centric Storage Adaptivity. Finally, we will give an exhaustive evaluation regarding the scalability of ERIS as well as our adaptivity facilities.
|
123 |
Adopting research data management (RDM) practices at the University of Namibia (UNAM): a view from researchersSamupwa, Astridah Njala 14 February 2020 (has links)
This study investigated the extent of Research Data Management (RDM) adoption at the University of Namibia (UNAM), viewing it from the researcher’s perspective. The objectives of the study were to investigate the extent to which RDM has been adopted as part of the research process at UNAM, to identify challenges encountered by researchers attempting to practice RDM and to provide solutions to some of the challenges identified. Rogers’ Diffusion of Innovation (DOI) theory was adopted for the study to place UNAM within an innovation-decision process stage. The study took a quantitative approach of which a survey was used. A stratified sample was drawn from a list of all 948 faculty members (the number of academics taken from the UNAM annual report of 2016). The Raosoft sample size calculator (Raosoft, 2004) states that 274 is the minimum recommended sample size necessary for a 5% margin of error and a 95% confidence level from a population of 948, and this was the intended sample size. A questionnaire administered via an online web-based software tool, SurveyMonkey, was used. A series of questions was asked to individuals to obtain statistically useful information on the topic under study. The paid version of SurveyMonkey was used for analysis while graphics and tables were created in Microsoft Excel. The results of the study showed that for the group that responded to the survey, the extent to which they have adopted RDM practices is still very low. Although individuals were found to be managing their research data, this was done out of their own free will; this is to say that there was no policy mandating and guiding their practices. The researcher placed most of the groups that responded to the survey at the first stage of the innovation-decision process, which is the information stage. However, librarians who responded to the survey were found to be more advanced as they were seen to be aware of and engaged in knowledge acquisition regarding RDM practices. Thus, the researcher placed them at the second stage in the innovation-decision process (Persuasion). Recommendations for the study are based on the analysed data. It is recommended, among others, that UNAM should give directives in the form of policies to enhance the adoption of RDM practices and this should be communicated to the entire UNAM community to create awareness regarding the concept of RDM.
|
124 |
Investigating the relevance of quality measurement indicators for South African higher education librariesNtshuntshe-Matshaya, Pateka Patricia January 2021 (has links)
Philosophiae Doctor - PhD / This study investigates the relevance of quality measurement indicators at higher education libraries for faculty academics, librarians, and students. The study followed a mixed-method design with a mixture of quantitative and qualitative data collection. Faculty academics, librarians and students ranked the existing quality measurement indicators for South African higher education libraries.
The findings revealed that for library quality measures to meet the needs of faculty academics, librarians, and students, the resources must be accessible both physically and virtually, and staff should be accountable and willing to offer services responsive to the users' needs and expectations of a safe, secure, and comfortable library space, be it physical or virtual. The qualitative data highlighted the importance of adequate resources and the adoption of new developments as measures for quality.
Quality measurement indicators must include elements such as adequate funding; relevant resources aligned with teaching and learning programmes; programmes that are integrated into teaching plans; effective supplier collaboration with respect to the process of acquiring relevant learning materials; effective student training; communication of the value of library services and alignment with the student learning outcomes; research support in a digital environment with e-tools and website navigability; research data management; and open access, which is a prominent role of the library. Based on the data, there was a quality measure (process) that was commendable even though it did not form part of the existing quality measures nor a service whose relevance was assessed. The separation of undergraduate and postgraduate learning spaces was amongst those services that ranked quite high from the students' responses (qualitative data). Even though there were differences emphasized on each indicator by either faculty academics or students, there were also discrepancies in the interpretation of what each quality indicator means to each study population group. As the study of this nature has recommendations and gaps identified in terms of research findings, it is quite important to record that there was a series of gaps that were identified in terms of library expectations and perceptions. These gaps were suggested as part of further research that must be conducted to fill the void in terms of library users’ voices in the development of higher education library measurement indicators.
|
125 |
Automatický nástroj k získávání metadat komponent pro úlohy průběžné integrace / Automatic Component Metadata Extractor and Consolidator for Continuous IntegrationKulda, Jiří January 2017 (has links)
Tato diplomová práce popisuje úpravu průběžné integrace pro Platform tým ve společnosti Red Hat. Výsledkem práce je nástroj Metamorph, který umožní sjednocení ostatních nástrojů průběžné integrace pod týmem Platform. Teoretická část popisuje vznik, popis a přidané hodnoty průběžné integrace. Následně jsou blíže přiblíženy existující nástroje na trhu. Dále je zde popsáno použití průběžné integrace v nástroji Jenkins. V práci jsou také dopodrobna popsány existující řešení průběžné integrace ve společnosti Red Hat. Dále je zde popsán návrh a implementace výše zmíněného nástroje. V závěru jsou výsledky práce otestovány týmem z firmy Red Hat a nastíněny možnosti rozšíření.
|
126 |
Developing an implementation plan for research data management (RDM) at the University of GhanaAvuglah, Bright Kwaku January 2016 (has links)
The current global and data intensive outlook of research provides new opportunities and challenges for HEIs including effective and sustainable RDM. As a growing area of interest in the global research arena, experiences from developed countries have dominated the body of literature on RDM. This study is in part, to fill this gap by assessing the state of the art of RDM and institutional preparedness at the University of Ghana (through existing data management activities and capabilities) in order to develop a plan for implementation.
The study used a qualitative case study method and gathered data using semi-structured interviews and document analysis. Thematic analysis method was used to analyse the data collected. A total of seven respondents (five service providers and two senior researchers) were selected purposively using two sampling techniques ("priori criteria sampling" and snowball sampling). Criteria were set for their inclusion and each respondent provided information about institutional support, capabilities, policies and expectations on RDM.
The findings of the study revealed a number of RDM related activities, these include support for collaborative research, support for data analysis and computational science, guidance on RDM and grant applications as well as support for storage and high-speed connectivity to facility the research enterprise at UG. In terms of capabilities, no specific RDM policy was identified, existing infrastructure identified include an HPC cluster, a private cloud facility (HP Cloud Matrix), an Institutional repository (UGSpace), an institutional Google Drive platform, data analysis packages (NVivo and SPSS) and a robust network and security infrastructure. These were not necessarily provisioned for RDM purposes. Also, the findings show that staff do not possess the necessary skills or adequate knowledge to fully support RDM at UG.
In terms of the specific objectives of the study, the results of the semi-structured interviews and document analysis provided an understanding of the current situation (i.e. requirements, current activities and capabilities at the UG) which is the first objective of the study. These findings were then benchmarked against the EPSRC policy framework following the outline of the DCC CARDIO Matrix and using the optimal desirable expectation or level of development as the standard for comparison. This was useful in identifying gaps in RDM awareness, support and capabilities at UG which is the second objective of the study. To achieve the third objective, which was identifying priority areas for RDM development, the researcher examined both initial findings (i.e. findings on requirements, current activities and capabilities identified under the first objective as well as the gaps identified in the second objective) and proposed six broad areas where UG must focus its RDM development agenda. Finally, the six broad areas proposed in objective three were further cascaded into a number of specific initiatives and tasks to be implemented. This was done taking cognisance of the potential of current infrastructure, gaps identified in institutional awareness and capabilities as well as essentials for a cultural changed. The study concluded that RDM at the University of Ghana is currently underdeveloped but with immense potential for growth. While a few RDM related activities were identified, existing capabilities were generally found to be inchoate, uncoordinated and not formally instituted.
The study recommended six main areas where the UG should focus RDM development, these include: constituting a steering group to spearhead and coordinate RDM development at the UG, developing a coordinated policy framework for RDM at UG, streamlining existing technical infrastructure to support data management requirements, creating opportunities for RDM training and capacity development for professional staff, researchers and students, developing services to support requirements, and exploring internal funding strategies to facilitate RDM development and support at the UG. The study also recommends that the academic community at the UG should be actively engaged throughout the RDM development process as this is critical to ensure that the eventual solutions are fit for purpose and acceptable. / Mini Dissertation (MIT)--University of Pretoria, 2016. / Information Science / MIT / Unrestricted
|
127 |
Scalable Dynamic Big Data Geovisualization With Spatial Data StructureSiqi Gu (8779961) 29 April 2020 (has links)
Comparing to traditional cartography, big data geographic information processing is not a simple task at all, it requires special methods and methods. When existing geovisualization systems face millions of data, the zoom function and the dynamical data adding function usually cannot be satisfied at the same time. This research classify the existing methods of geovisualization, then analyze its functions and bottlenecks, analyze its applicability in the big data environment, and proposes a method that combines spatial data structure and iterative calculation on demand. It also proves that this method can effectively balance the performance of scaling and new data, and it is significantly better than the existing library in the time consumption of new data and scaling<br>
|
128 |
Energy-Aware Data Management on NUMA ArchitecturesKissinger, Thomas 23 March 2017 (has links)
The ever-increasing need for more computing and data processing power demands for a continuous and rapid growth of power-hungry data center capacities all over the world. As a first study in 2008 revealed, energy consumption of such data centers is becoming a critical problem, since their power consumption is about to double every 5 years. However, a recently (2016) released follow-up study points out that this threatening trend was dramatically throttled within the past years, due to the increased energy efficiency actions taken by data center operators. Furthermore, the authors of the study emphasize that making and keeping data centers energy-efficient is a continuous task, because more and more computing power is demanded from the same or an even lower energy budget, and that this threatening energy consumption trend will resume as soon as energy efficiency research efforts and its market adoption are reduced. An important class of applications running in data centers are data management systems, which are a fundamental component of nearly every application stack. While those systems were traditionally designed as disk-based databases that are optimized for keeping disk accesses as low a possible, modern state-of-the-art database systems are main memory-centric and store the entire data pool in the main memory, which replaces the disk as main bottleneck. To scale up such in-memory database systems, non-uniform memory access (NUMA) hardware architectures are employed that face a decreased bandwidth and an increased latency when accessing remote memory compared to the local memory.
In this thesis, we investigate energy awareness aspects of large scale-up NUMA systems in the context of in-memory data management systems. To do so, we pick up the idea of a fine-grained data-oriented architecture and improve the concept in a way that it keeps pace with increased absolute performance numbers of a pure in-memory DBMS and scales up on NUMA systems in the large scale. To achieve this goal, we design and build ERIS, the first scale-up in-memory data management system that is designed from scratch to implement a data-oriented architecture. With the help of the ERIS platform, we explore our novel core concept for energy awareness, which is Energy Awareness by Adaptivity. The concept describes that software and especially database systems have to quickly respond to environmental changes (i.e., workload changes) by adapting themselves to enter a state of low energy consumption. We present the hierarchically organized Energy-Control Loop (ECL), which is a reactive control loop and provides two concrete implementations of our Energy Awareness by Adaptivity concept, namely the hardware-centric Resource Adaptivity and the software-centric Storage Adaptivity. Finally, we will give an exhaustive evaluation regarding the scalability of ERIS as well as our adaptivity facilities.
|
129 |
Indexallokation in Parallelen DatenbanksystemenStöhr, Thomas 23 October 2018 (has links)
Die effiziente Nutzung von Zugriffsstrukturen ist eine wichtige Voraussetzung für die performante Durchführung von Datenbankanfragen. Die in Parallelen Datenbanksystemen vom Typ Shared-Nothing übliche, durch die Allokationsstrategie für Relationen weitgehend vorgegebene Indexallokation führt oftmals zu unnötigen I/O-, Verarbeitungs- und Kommunikationskosten. Parallele Shared-Disk Datenbanksysteme bieten durch ihren gemeinsamen Plattenzugriff ein hohes Potential zur flexiblen Allokation von Indexstrukturen. Wir präsentieren eine Klassifikation und eine qualitative Bewertung von Indexallokations-Strategien für diese Architekturklasse, die zeigt, daß sich durch die flexible Wahl von Größen wie Verteilattribut und Verteilgrad die Performanz der parallelen Indexverarbeitung steigern läßt.
|
130 |
Multi-User Evaluation of XML Data Management Systems with XMach-1Böhme, Timo, Rahm, Erhard 09 November 2018 (has links)
XMach-1 was the first XML data management benchmark designed for general applicability [1]. It is still the only benchmark supporting a multiuser performance evaluation of XML database systems. After a brief review of XMach-1 we summarize three additionally proposed benchmarks (XMark, XOO7, Mbench) and provide a comparison between these benchmarks. We then present experiences and performance results from evaluating XML database systems with XMach-1.
|
Page generated in 0.0758 seconds