NautDB: Toward a Hybrid Runtime for Processing Compiled Queries

Student: Samuel Grayson (University of Texas, Dallas)
Supervisor: Kyle Hale (Illinois Institute of Technology)

Abstract: General purpose operating and database system suffer under the load of their generality which makes achieving optimal performance extremely hard, especially on modern hardware. The goal of this research is to integrate, for the first time, specialization techniques from the OS community (hybrid runtimes) and DB community (compiled queries) for high-performance query processing on modern hardware. We envision a system called NautDB, a hybrid dataflow runtime for executing compiled queries. As a first step toward our goal, we evaluate the performance of compiled queries on Linux and run as a Nautilus hybrid runtime using a simple prototype. Our results demonstrate that combining these specialization techniques has transformative potential for building the next generation (distributed) high-performance query processing systems and big data platforms.

ACM-SRC Semi-Finalist: no

Poster: PDF
Poster Summary: pdf
Reproducibility Description Appendix: PDF


Back to Poster Archive Listing