Post-relational database management system based on dictionary technology

The upcoming phase of transformations aimed at expanding the application of information technologies to the activities of companies in rocket and space industry involves a significant increase in the flow of processed data. To achieve reliable and safe storage, transmission, classification and subsequent analysis of various types of data, one needs to have technologies not only capable of supporting the internal activities of the company, but also providing the ability to predict the costs and time required for completion of research and development efforts based on stringent requirements set for rocket and space hardware and current manufacturing environment. The large amount of data, as well as their structural diversity have a significant impact on both the complexity of designing relational databases and the query response time. Electing not to use unambiguous fields in the relational database (wavering the requirement of the first normal form) results post-relational data model. In post-relational data model the number of tables is significantly reduced (due to their nesting) while preserving simplicity of structure and the option of object orientation. Constructing a post-relational database management system entails a multitude of technical problems related to data access and storage. The authors propose to solve them through constructing the post-relational database management system based on dictionary technology. The dictionary technology developed by the authors has functional completeness and high rate of handling the data presented in lexicographic form. The modular architecture of the database makes it possible to present data as multidimensional cubes, which is inherent in multidimensional databases. Data storage is organized in the form of dictionaries, the number of which is driven by the number of blocks, the number of attributes and the method of representing tables, taking into account the option to maintain nested tables. Managing a large number of dictionaries is implemented in the form of a dictionary manager. The dictionary-based manager organizes access to all the components of the database dictionaries. Constructing a dictionary with specialized structure solves the problem of augmenting the database without significant restructuring of the data storage dictionary, while obviating the need to use indexing procedure. The data structure is described in the reference dictionary, which makes it possible to quickly change the database structure without the need to update data storage. To execute queries, a dictionary query language was developed, which makes it possible to set multiple conditions for data selection. This approach seems promising in terms of development of industrial systems for database management.