A query sampling method for estimating local cost parameters in amultidatabase system

摘要

In a multidatabase system (MDBS), some query optimizationninformation related to local database systems may not be available atnthe global level because of local autonomy. To perform global querynoptimization, a method is required to derive the necessary localninformation. This paper presents a new method that employs a querynsampling technique to estimate the cost parameters of an autonomousnlocal database system. We introduce a classification for grouping localnqueries and suggest a cost estimation formula for the queries in eachnclass. We present a procedure to draw a sample of queries from eachnclass and use the observed costs of sample queries to determine the costnparameters by multiple regression. Experimental results indicate thatnthe method is quite promising for estimating the cost of local queriesnin an MDBS