
Diplo Cloud: Efficient and Scalable Management of RDF Data in the Cloud
Abstract
System Configuration
Platform : cloud computing
Conclusion
DiploCloud is an efficient and scalable system for managing RDF data in the cloud. From our perspective, it strikes an optimal balance between intra-operator parallelism and data co-location by considering recurring, fine-grained physiological RDF partitions and distributed data allocation schemes, leading however to potentially bigger data (redundancy introduced by higher scopes or adaptive molecules) and to more complex inserts and updates. DiploCloud is particularly suited to clusters of commodity machines and cloud environments where network latencies can be high, since it systematically tries to avoid all complex and distributed operations for query execution. Our experimental evaluation showed that it very favorably compares to state-of-the-art systems in such environments. We plan to continue developing DiploCloud in several directions: First, we plan to include some further compression mechanisms (e.g., HDT ). We plan to work on an automatic templates discovery based on frequent patterns and untyped elements. Also, we plan to work on integrating an inference engine into DiploCloud to support a larger set of semantic constraints and queriesnatively.Finally,wearecurrently testing and extending our system with several partners in order to manage extremely-large scale, distributed RDF datasets in the context of bioinformatics applications.