Research On Adaptive RDF Stream Processing Architecture

In the processing of RDF stream data,most of the existing RDF stream processing languages are based on the continuous query language processing model.They implement the query of RDF stream data by extending the function of SPARQL,but they lack the definition of semantic layer.The semantics of RDF stream processing actually lacks a common formal representation.Secondly,the existing RDF stream processing engine has poor universality,which can only meet some of the requirements of stream processing,and it is difficult to support mixed query of heterogeneous data.This paper proposes a new continuous query language CT-SPARQL,which defines a unified formal semantics for continuous queries.Most RDF stream query languages can compare their different syntax semantics in this formal semantic framework and convert to CT-SPARQL queries.Based on this semantics,this paper proposes an adaptive RDF stream processing framework SRSP,which can evaluate the current mainstream RDF stream processing language under a unified framework.Further,this paper extends the SRSP and proposes the SRSPG framework.SRSPG uses GPU for collaborative processing when querying SPARQL engine to improve the query efficiency of RDF stream processing.In addition,this paper also proposes a hybrid data query framework FedQL from the perspective of application requirements,which is used to deal with the combination of RDF stream data and relational data query.Finally,this paper designs a large number of comparative experiments on the standard test dataset and real dataset for the proposed three frameworks.The experimental results show that SRSP and SRSPG can not only support the current mainstream query language in a unified framework,but also utilize the computing power of the GPU to improve the efficiency of the query.FedQL can support the query processing of mixed data to meet the needs of the practical application level.In summary,the formal semantic framework defined in this paper guarantees the correctness of RDF stream query.The proposed stream processing framework has strong universality and supports query processing of mixed data.