The Research On RDF Data Management System

Along with the application of RDF becoming more and more popular, the requirement of RDF storage and query has increased greatly.This paper researches on RDF data management system and proposes a prototype system called RDFBase to store and query RDF data. Firstly, the storage mechanism is introduced. We adopt the approach of storing all triples in a single, potentially huge triples table, with our own storage implementation underneath. We build six separate three-values indexes, six two-values indexes and three one-value indexes to enhance the search efficiency. Besides, we compress all the indexes to reduce the space. Secondly, we discuss the query processing mechanism. We estimate and compare the costs of executing a query using different execution strategies and choose the strategy with the lowest cost estimation. We use a method of more accurate join selectivity estimation for join-order optimization. We also implement light-weight methods for sideways information passing which can greatly accelerate index scans to pass information across operators in a way that cuts through the tree structure of the execution plan.Experimental studies demonstrate that our compression scheme has a good compression rate. Furthermore, our methods for sideways information passing and join selectivity estimation allow for quick and scalable query processing. Both of them enhance the query execution efficiently.