XML has emerged as leading document format for exchanging data. Because of vast amounts of XML documents available and transfered, there is a strong need to store and query information in these documents. However, the most companies are still using a RDBMS for their data warehouses and it is often necessary to combine legacy data with the ones in XML format, so it might be useful to consider storage possibilities for XML documents in a relation database. In this thesis we focused on structured and semi-structured data-based XML documents, because they are the most common when exchanging data and they can be easily validated against an XML schema. We propose a slightly modified Hybrid algorithm to shred doc- uments into relations using an XSD scheme and we allowed redundancy to make queries faster. Our goal was not to provide an academic solution, but fully working system supporting latest standards, which will beat up native XML databases both by performance and vertical scalability.

In data-centric environments, for example, in the field of scientific computing, the transmission of large amount of structured data to Web services is required. In service-oriented environments (SOA), the Simple Object Access Protocol (SOAP) is commonly used as the main transport protocol. However, the resulting 'by value' data transmission approach is not efficiently applicable in data-centric environments. One challenging bottleneck of SOAP arises from the XML serialization and deserialization when processing large SOAP messages. In this paper, we present an extended Web service framework which explicitly considers the data aspects of functional Web services. Aside from the possibility to integrate specialized data transfer methods in SOA, this framework allows the efficient and scalable data handling and processing within Web services. In this case, we combine the advantages of the functional perspective (SOA) and the data perspective to efficiently support data-centric environments.