RDF mapping engines enable access to existing heterogeneous data sources as RDF Knowledge Graph (KG). However, these map- ping engines have two challenges: i) processing streaming data sources with changing velocity efficiently, ii) and providing a rich variety in the format of the generated KG output. To tackle these challenges, I carry out my research in 3 steps. I will first design a highly scalable data stream mapping solution to handle dynamic velocity of streaming data sources. Preliminary results indicate that our stream mapping solution outperforms state of the art engines with lower latency, constant memory usage, and higher throughput. I will then refine this architecture in a task-based fashion, aiming to be a common architecture for any kind of mapping. Finally, I will utilize the common modular mapping architecture and extend it with a component to derive an intermediate representation of the data mapping process, enabling heterogeneous to heterogeneous data mapping. The combined solution will provide a highly scalable heterogeneous to heterogeneous data stream mapping engine, enabling us to have multiple views of the underlying KG.

Stream-reasoning query languages such as CQELS and C-SPARQL enable query answering over RDF streams. Unfortunately, there currently is a lack of efficient RDF stream generators to feed RDF stream reasoners. State-of-the-art RDF stream generators are limited with regard to the velocity and volume of streaming data they can handle. To efficiently generate RDF streams in a scalable way, we extended the RMLStreamer to also generate RDF streams from dynamic heterogeneous data streams. This paper introduces a scalable solution that relies on a dynamic window approach to generate RDF streams with low latency and high throughput from multiple heterogeneous data streams. Our evaluation shows that our solution outperforms the state-of-the-art by achieving millisecond latency (compared to seconds that state-of-the-art solutions need), constant memory usage for all workloads, and sustainable throughput of around 70,000 records/s (compared to 10,000 records/s that state-of-the-art solutions take). This opens up the access to numerous data streams for integration with the semantic web.