CAESAR: Context-Aware Event Stream Analytics for Urban Transportation Services
- Olga Poppe ,
- Chuan Lei ,
- Elke A. Rundensteiner ,
- Daniel J. Dougherty ,
- Goutham Deva ,
- Nicholas Fajardo ,
- James Owens ,
- Thomas Schweich ,
- MaryAnn Van Valkenburg ,
- Sarun Paisarnsrisomsuk ,
- Pitchaya Wiratchotisatian ,
- George Gettel ,
- Robert Hollinger ,
- Devin Roberts ,
- Daniel Tocco
EDBT |
We demonstrate the first full-fledged context-aware event processing solution, called CAESAR, that supports application contexts as first class citizens. CAESAR offers human readable specification of context-aware application semantics composed of context derivation and context processing. Both classes of queries are only relevant during their respective contexts. They are suspended otherwise to save resources and to speed up the system responsiveness to the current situation. Furthermore, we demonstrate the context-driven optimization techniques including context window push-down and query workload sharing among overlapping context windows. We illustrate the usability and performance gain of our CAESAR system by a use case scenario for urban transportation services using real data sets.