Bridging the gap between expressivity and efficiency in stream reasoning: a structural caching approach for IoT...

Area of Science:

• Computer Science
• Artificial Intelligence
• Data Science

Background:

• Data streams from domains like IoT, Smart Industries, and Social Media require real-time processing for insights.
• Integrating these streams with static data and domain knowledge is challenging.
• Ontologies and expressive reasoning (e.g., OWL2 DL) are crucial for data integration but struggle with high-velocity data.

Purpose of the Study:

• To bridge the gap between expressive reasoning capabilities and the real-time efficiency demands of IoT data streams.
• To enable efficient event enrichment over high-velocity IoT data.

Main Methods:

• Development of a structural caching technique tailored for IoT data streams.
• Exploitation of the structural and size similarity of events in typical IoT streams.
• Application of the technique with OWL2 DL, OWL2 RL, and OWL2 EL reasoners.

Main Results:

• The caching technique achieves significant speedups in reasoning time.
• Speedups reach up to thousands of times for OWL2 DL reasoners.
• Tens to hundreds of times speedup is observed for OWL2 RL and OWL2 EL reasoners.

Conclusions:

• The proposed caching method effectively enhances the efficiency of expressive reasoning over IoT data streams.
• This approach facilitates real-time data integration and event enrichment in data-intensive IoT applications.
• It represents a significant step towards practical, efficient semantic reasoning in dynamic data environments.