Pavlo Zhezhnych Personal Site and Blog

The thesis is dedicated to solving an important scientific problem of relational database quality increase during their long-term usage regardless of a data domain, software and hardware features using created methods and means of organizing a relational database as an information product taking into consideration time characteristics of data. To solve the problem a formal universal informational model of a data domain is proposed. A metamoment time relational data model based on the introduced notion of time-dependent data was built. It is the basis of new methods of relational databases representation by means of XML-technologies, and of a complete operations set in the form of SQL queries templates set intended for manipulating relational databases taking into account time. ISO-9126 standard-based methods of quality evaluation of relational databases with time characteristics were developed specifying directions of quality increase of relational databases considered as information products.
At an application layer a set of quality parameters of relational databases considered as information products was developed with the purpose of their long-term usage. Several samples of relational time-depended databases were created as well. Continuous experiments have proved correctness of the problem statement, trustworthiness of formal models built and efficiency of the proposed methods and means of relational time-dependent databases organizing.
In the first chapter the analysis of modern tendencies and development prospects of database systems in the information society was held. Peculiar features of organizing database as an information product taking into consideration time characteristic of data were described. In the second chapter the methods of data domain dynamics modeling by means of databases, which are the basis for organizing relational time-dependent databases, are developed. The notion of time-dependent data which represents the objective influence of time on database storage as information product was introduced. The structure of data domains according to the types of logical database organization was formalized. In the third chapter a metamoment time relational data model which is the basis for analyzing design and implementation of relational time-dependent databases was developed. The mathematical definition of time-dependent data relations was presented and their classification devised. The algebra of relations, the notions of functional dependencies and keys as well as normal forms of relational time-dependent databases were improved. The completeness of algebra of relations of time-dependent data relatively to relational algebra was proved. In the fourth chapter the methods of organizing relational time-dependent databases by means of specialized XTDQML XML-language were developed. Methods of SQL-scripts and XML-documents generation with XTDQML and XSLT transformations were worked out that are the basis of time-dependent data processing in service-oriented architectures. In the fifth chapter methods and means of relational time-dependent database organizing on conceptual and logical levels of their hierarchy were developed. New UTDER diagram means of relational databases analysis and modeling taking into consideration metamoment time were devised. The templates of conceptual and logical schemes on the basis of universal information data domain model were created. The general quality rate of a database as an information product was formalized. In the sixth chapter a set of database samples based on methods and means of relational time-dependent database organizing was described.