Starting from the new SPARQL extension of the previous chapter, we aim exploiting IBM InfoSphere Streams as execution framework for data stream processing. In order to do that, we need to provide a translator that from DubExtentions queries is able of producing streaming applications, by exploiting the SPL program language of InfoSphere Streams. For this reason the translator has
to parse and to interpret DubExtentions queries by their query string representations, providing an internal representation (a query object) that can be better manage by the system;
to transform the query object that represents the DubExtentions continuous queries in the effective SPL streaming application, so that we can exploit InfoSphere Streams as query engine to execute the streaming applications.
We implement the translator based on the ARQ architecture, by exploiting the parser and analyser components, and by extending the ARQ with new component to provide the complete translation. In this way, our DubExtentions execution framework uses a “white box” approach as CQELS, because we generate streaming applications exploiting native operators of InfoSphere Streams and analysing data stream on the fly, without any intermediate translation.
