Zpracování velkých objemů nestrukturovaných dat na platformě Hadoop

Abstract

This thesis is concerned with processing unstructured text on Hadoop platform. First part focuses on the reasons of creation Big Data concept. I explain data issue of these days and show, why common database systems are inappropriate for working with huge amounts of unstructured data. The next part focuses on theory about Big Data concept and processing on Hadoop platform. I introduce Hadoop architecture and how it differs from common warehouse database. I also explain parallel data processing theory and how parallel processing is solved on Hadoop platform. The last part focuses on practical part to solve processing huge amounts of unstructured server logs on Hadoop platform. How can we analyze these data and get some valuable information from them. The result will be presented in reporting program QlikView. Result was also processed on classic SQL database, to compare Hadoop platform contribution in processing unstructured data.