Spatiotemporal Storage Layout for Fewer Range Scans and Hotspot Balancing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for handling spatial-temporal data, such as HBase and HDFS, are inefficient due to the need for encoding mechanisms to translate spatial and temporal information into unique keys, leading to increased resource consumption and poor performance in handling geometry queries and load balancing against moving hotspots.
Innovation Solution
A spatial-temporal storage system that includes a geometric translation circuit to split spatial-temporal information into row keys, a multi-scan optimization circuit to optimize range scans, and a block grouping circuit to preserve data locality during splits, using algorithms like Moore encoding and group-based replica placement.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If encoding mechanisms are used to translate spatial and temporal information into unique keys, then spatial-temporal data can be stored in HBase, but the number of range scans increases and query performance deteriorates
Solution Approach 1:
The patent segments spatial-temporal data into spatial components and temporal components, storing them in separate data structures. Spatial data is organized using a spatial index (e.g., R-tree or grid-based indexing) while temporal data is stored in time-series structures. This segmentation allows independent optimization of each component and enables efficient range queries without requiring encoding of the entire spatial-temporal key.
Solution Approach 2:
The patent introduces additional indexing dimensions beyond the traditional single primary key. Instead of encoding spatial-temporal data into a one-dimensional key, the system creates multi-dimensional indexes where spatial coordinates form one dimension and temporal timestamps form another dimension. This allows the database to perform range scans on individual dimensions independently, dramatically reducing the number of scans required for geometry queries.
2Reliability
If HBase is used as a general purpose solution for spatial-temporal data, then data can be stored, but resource consumption increases and load balancing against moving hotspots deteriorates
Solution Approach 1:
The patent applies local quality by optimizing storage and access patterns for different types of data access. Frequently accessed spatial regions are cached in memory with higher priority, while less accessed data remains on disk. The system dynamically adjusts caching strategies based on query patterns, ensuring that hotspots are served from fast memory while cold data remains in slower storage, thereby reducing overall resource consumption.
Solution Approach 2:
The patent implements dynamic load balancing that adapts to moving hotspots in real-time. When certain spatial-temporal regions experience increased query loads (hotspots), the system automatically redistributes data across region servers or adjusts replication factors dynamically. This dynamic adaptation prevents resource exhaustion at any single node and ensures balanced resource utilization across the distributed system.
Data Source
AI summary
A spatial-temporal storage method, system, and non-transitory computer readable medium include dynamically managing a plurality of region servers for querying spatiotemporal data in noSQL databases.