Towards Enabling Semantic Reasoning with Timely Dataflow and Differential Dataflow

In Big Data applications Data Volume, Data Variety and Data Velocity constitute three important challenges. In this thesis, we focus on the problem of Data Velocity and Data Variety. The Semantic Web offers solutions to Data Variety. Stream Processing, on the other hand, focuses on Data Velocity. Semantic Reasoning (SR) is the area that attempts to tackle the two problems at the same time. It aims at performing reasoning over Streaming Data from complex domain in a timely fashion. In this work, we aim at exploring a new approach in the SR area. We argue that problems in this area can be compared and ultimately solved as problems in Graph Stream Processing. So in this direction, we select two frameworks for efficient Graph Stream Processing: Timely Dataflow and Differential Dataflow and use them to perform SR tasks. After describing the main basic components needed for a better understanding of the topic, we formalize the problem. Based on this problem, we design the flow of the computation that will lead the implementation of the code. At the end, we show some performance values that will delineate a meaningful starting point on top of which we define further future work.