The Spatiotemporal big data store is designed for massive datasets and optimized for high-speed visualization, time-enabled querying, and big data analytics. It is ideal for storing millions of point features such as sensors or oil rig locations, especially when performance and scalability are critical.

From ArcGIS documentation:

“The spatiotemporal big data store is designed for writing and querying large volumes of spatial and temporal data. It’s optimized for visualizing high-density datasets in web maps.”

Option B (Relational) works for moderate data volumes but may not scale well.

Option C supports cached tiles, not dynamic feature access.