FESTIval: A versatile framework for conducting experimental evaluations of spatial indices

Area of Science:

• Computer Science
• Database Systems
• Geographic Information Systems

Background:

• Spatial indices enhance spatial query performance in database and GIS systems.
• Evaluating spatial index performance is challenging due to complex data, varied configurations, and diverse storage systems (disk, flash memory).

Purpose of the Study:

• To introduce FESTIval, a versatile framework for the experimental evaluation of spatial indices.
• To provide a quantitative method for analyzing and comparing the performance of different spatial index strategies.

Main Methods:

• FESTIval supports various disk-based and flash-aware spatial indices.
• The framework allows specification and execution of user-defined workloads.
• A data schema stores index configurations and workload statistics for reproducibility.

Main Results:

• FESTIval enables reproducible experimental evaluations of spatial indices.
• The framework handles diverse spatial datasets and index configurations.
• FESTIval has been utilized to validate novel flash-aware spatial indices, such as eFIND-based indices.

Conclusions:

• FESTIval offers an extensible and robust environment for spatial index performance analysis.
• The framework addresses the complexities of evaluating spatial indices across different storage media and configurations.