Parallel implementation of the join operation database

The paper deals with the solution of the task of parallelizing join operations in multiprocessor computing systems, as applied to databases. Experimental data on the implementation of the operation of the connection of lax orderly files on the basis of indexed sequential files. The impact analysis of the implementation on the resources of the multiprocessor computing system. Theoretical calculations and experimental data for the goal of a merge table based on cloud technologies. Defined the relationship of the number of concurrently running processes with the number allocated to the implementation of the operation cores on a multiprocessor system or cluster. It is shown that the use of the investigated operations on the metadata, allows you to effectively parallelize the join operation and to reduce the amount of data processed in the framework of this operation. The graphic interpretation of experimental data confirming the theoretical calculations.