Information Technologies and Computer Engineering

The introductory part of the research justifies the relevance of the work. Because of the tendency to move from local to distributed databases, and the disadvantages of using "universal" accounting systems many companies use the set of special solutions instead. According to this, it is very important to distribute data of corporate database between its nodes correctly. The key factor that affects the reliability and availability of the database is the link localization. If the database is distributed optimally, the data hosted in this node is called exclusively by its user. 
The research aims to determine and build a mathematical model of the optimality criteria for the structure of a remote node of the distributed database. The statistics of user SQL-queries activity is taken into account. Statistics results is presented in the form of a multidimensional data model. It this model each dimension element is marked with a marker of data presentation on the remote node. Next, the statistics data is consolidated according to the aggregation function proposed by the author. As a result, the necessity level value of data cell representing is obtained. 
At the next stage, criteria the model effectiveness are formulated (independence from the central node of the database, the size of the local database, and an indicator of the level of need for data synchronization). Then mathematical dependencies between criteria values and the value of data representation level on the remote node are built. This allows, having solved multicriteria analysis problem, to determine the level of data representation for which the values of the criteria will correspond to the coordinates of the optimal point in the solution space

References