Predlog arhitekture sistema visokih performansi za generalnu obradu podataka na klasterima za podatke velikog obima
In recent years, the application and widespread adoptionof Big Data, Internet of Things (IoT), Cloud technologieshave increased the use of large-scale data processing systems.These technologies increased significantly and exponentiallywith the heterogeneous data generated (structured,unstructured, and semi-structured). The processing andanalysis of a tremendous amount of data is cumbersome andis gradually moving from the classic "batch" processing -extraction, transformation, loading (ETL) techniques to realtimeprocessing. For example, in the domain of theautomobile industry, healthcare, but also in other disciplines.Tracking, data processing, environmental management, timeseriesdata, and historical data set are crucial to forecastingmodels not only in these domains.This doctoral dissertation is about the design of ageneral architecture for processing a large amount of data.The architecture as such enables efficient acquisition of data,their optimal placement, processing of large amounts of data,use of various algorithms for drawing conclusions as well asfor displaying data. The doctoral dissertation shows thecomplete process of modeling and designing architecture, theselection of appropriate software components for itsrealization. The presented platform met very demandingparameters for meeting the system's performance, includingthe standard for decision support of the TransactionProcessing Council (TPC-H) by following the EuropeanUnion (EU) legislation and the Czech Republic. Currently, thepresented proof of concept (PoC) that has been upgraded tothe production environment has united isolated parts of theCzech Republic's healthcare. The reported PoC Big DataAnalytics platform, artefacts and concepts can be transferredto health systems in other countries interested in developingor upgrading their national health infrastructure in a costeffective,secure, scalable, and high-performance way.
Bibliografija: listovi 95-100. Datum odbrane: 10.07.2023. Distributed (cluster) data processing systems
This work is licensed under a
CC BY-NC-ND 3.0 AT - Creative Commons Attribution-NonCommercial-NoDerivs 3.0 Austria License.