RayBooster: A Ray Tracing Engine to Accelerate SedonaDB

Abstract

Hardware acceleration is becoming increasingly critical for spatial databases, as their workloads are geometrically complex, dataintensive, and subject to growing real-time requirements. Building on our prior strong evidence that Ray Tracing (RT) cores can significantly accelerate spatial queries through dedicated hardware support, this paper presents RayBooster, the first solution that incorporates RT-core acceleration into a production-grade geospatial database system, Apache SedonaDB. Our approach not only delivers substantial performance improvements but also does so costeffectively. We focus on spatial joins, which dominate execution time and computational cost in spatial databases. To enable this integration, we bridge the mismatch between spatial query engines and RT hardware through three key system innovations. First, to overcome the lack of random access in the standard Well-Known Binary format, we design a GPU-optimized Structure of Arrays storage layout. Second, we eliminate indexing scalability bottlenecks by constructing a monolithic Bounding Volume Hierarchy tree that encodes geometry IDs into the Z-axis of the geometric scene, bypassing the overhead of managing millions of micro-indexes. Third, to manage the combinatorial complexity of diverse geometry types and spatial predicates, we develop a unified RelateEngine. This engine utilizes RT cores to compute the Dimensionally Extended 9-Intersection Model, serving as a universal topological descriptor. Furthermore, we implement a memory-aware execution strategy to mitigate out-of-memory failures through robust, best-effort resource management. Seamlessly integrated as an extension to SedonaDB, RayBooster delivers up to a 5.93× performance speedup on SpatialBench and provides a 59.02% reduction in operational costs, effectively transforming these idle RT units into a highly efficient engine for spatial analytics.