Design And Implementation Of Temporal RDF Storage System Based On TimeDB

With the rapid development of Semantic Web and temporal data processing,temporal RDF(Resource Description Framework)data has influxed into the network.However there are few researches about temporal RDF data model and related storage methods.In order to realize the mapping from data model to temporal database,a model of temporal RDF data model in the logical layer has been put forward based on the past work about temporal RDF data model in this paper,and the related mapping rules as well.Finally,a temporal RDF data storage system has been realized.The main work about this paper are as follows:1)A framework of the temporal RDF data storage system was proposed and implemented in this paper.We discussed the efficiency about the mapping method from traditional RDF storage to RDBMS in temporal RDF storage application,and then proposed a storage method which store ontology data and instant data separately.Then,we compared the complexity about BFS and DFS,and choosed the more efficient one.A mapping rule has defined between temporal RDF model and a temporal RDF storage system has been built according to this,we gave the detail about different part in this system in this paper.In order to deal with the consistency in RDF storage,we proposed a inference engine which dealt with RDF triple part and timestamp part separately.Finally,we realized a temporal RDF storage system,TRDF-Storage,based on a temporal database TimeDB,and the effectiveness of storage scheme were proved by experiment.2)The optimization of the temporal RDF storage system has been further optimized.In order to improve the query efficiency of proposed storage system,I set up an index structure based on MBR(Minimal Bounding Rectangle,the smallest boundary rectangle)--2T-tree on temporal tags of temporal RDF data.To solve the problem that traditional R-tree index can not handle with temporal data,a method of transforming time variables is proposed,and then two kinds of temporal RDF data types in 2T-tree are processed.The experimental results showed that the storage system after the optimization is more efficient than before.