Dynamic solution for multiple question optimization

Abstract

The problem of multiple query optimization is considered as a complete optimization problem, the input parameters for optimization are determined: queries, tasks, plans. Proposed and substantiated architecture of the system in which the DBMS with the query optimizer is used - a computer system in which one or more computers (information processing system, one or more client computers, control computer, one or more the number of database servers) are interconnected by the client network and the server network. The data processing system receives the primary request made by one of the client computers, creates one or more secondary requests and transmits them to the database servers, in the case, if necessary, performs links or data updates, as defined in the primary request, and then returns the resulting data to the computer of the client from which the initial request was made. The I / O processing module, query analyzer, query optimizer, query processing module, optimization controller, external memory device are the elements that make up the information processing module. The I / O processing module receives the request from the client computers and the control request from the control computer and responds to these requests. The query optimizer optimizes the query using the query tree generated by the query analyzer, and develops a procedure for a series of operations (query execution plan) to obtain the results of query processing. The query processing module executes the query execution plan developed by the query optimizer. The order of execution of elementary requests which provide the minimum time for one, two and three processors is developed. A comparison of multi-request execution time with joint and incompatible processing depending on the number of processors is presented. The minimum execution time of the query can be achieved by executing elementary queries in the appropriate order, determined by the condition of order.