Kravets I. Distribution of load balancing in large volume databases using horizontal fragmentation

The thesis describes developing methodologies, techniques and algorithms for distribution of load in a large volume database. A new heuristic method iSmartRoute optimization for large-scale graphs is proposed. Improved way of optimization for functions with multiple extremums, the results of which confirmed the practical and scientific value of evolutionary algorithms. De-veloped a method of decomposing SQL-query in a distributed database to generate aggregate data about the frequency of traffic to tuples. An experimental study using permanent connections to the database in the traditional cluster system and segmented has been performed. Developed a method of horizontal fragmentation of large volume databases. Performed an organization of load balancing in database data-processing system using horizontal fragmentation and describes its main characteristics, including its performance and data protection.