SPEAR: Dynamic Spatio-Temporal Query Processing over High Velocity Data Streams

Area of Science:

• Data Science
• Computer Science
• Geospatial Analysis

Background:

• The Internet of Things (IoT) and 5G generate massive real-time spatio-temporal data streams.
• Existing systems struggle with dynamic query adjustments, leading to system unresponsiveness and outdated results.
• Optimization is needed for objects with changing motion states in distributed spatio-temporal data management.

Purpose of the Study:

• To propose a dynamic spatio-temporal query processing strategy for high-velocity big data streams.
• To enhance the responsiveness and accuracy of spatio-temporal data stream management systems.
• To address the limitations of the specified-once-applied-continuously query model.

Main Methods:

• Adaptation of principles from streaming databases, spatial data management, and distributed computing.
• Formulation of spatio-temporal data types and functions for dynamic query state handling.
• Development of streaming spatio-temporal querying methods and geohash-based dynamic spatial partitioning.
• Implementation of a prototype on Apache Flink for in-memory stream processing.

Main Results:

• The proposed strategy effectively handles dynamic spatio-temporal query processing.
• Maintained high and consistent processing rates for both stationary and moving queries.
• Demonstrated effectiveness over high-velocity spatio-temporal big data streams.

Conclusions:

• The developed approach significantly improves real-time analysis of dynamic spatio-temporal data.
• The geohash-based partitioning enables effective parallel processing for big data streams.
• This work provides a robust solution for modern data stream management challenges.