Polygon Simplification for In-Memory Spatial Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current column-oriented data management architectures in databases, such as SAP HANA, face inefficiencies in handling geographic data due to the use of tree-based index structures that are not optimized for in-memory systems, leading to performance overhead and inability to handle continuous scans effectively.
Innovation Solution
The approach involves simplifying polygons by dissecting them into line segments, omitting points that occupy the same pixel, and increasing local resolution where necessary, while maintaining validation through inverse mapping to ensure accurate representation and error control, allowing for efficient storage and transmission of spatial data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If tree-based index structures (R-Tree, K-d Tree, Quadtree) are used for spatial data management, then efficient access to secondary storage structures is achieved, but performance overhead increases and continuous scans cannot be performed efficiently in in-memory databases
Solution Approach 1:
The patent extracts the indexing function from traditional tree-based structures and replaces it with a bitmap index that directly maps spatial regions to data records. This eliminates the complex node-link structure of R-Trees and K-d Trees while maintaining efficient access capabilities through bitmap operations that are optimized for in-memory processing and continuous scans.
Solution Approach 2:
The patent replaces the mechanical tree-based indexing mechanism with a bitmap-based indexing system. Instead of traversing tree nodes and following links, the system uses bitmap operations (bitwise AND, OR, NOT) to efficiently query and scan spatial data, leveraging hardware-optimized bitwise operations for superior performance in in-memory environments.
2Measurement precision
If high-resolution polygon data is stored and transmitted, then accurate spatial representation is achieved, but storage requirements and bandwidth consumption increase
Solution Approach 1:
The patent applies local quality by storing polygon data at full resolution only where spatially relevant (intersecting or adjacent to query regions), while using lower-resolution or aggregated representations for other areas. The bitmap index enables selective access to high-resolution data only when needed, reducing overall storage and transmission requirements while maintaining accuracy where it matters.
Solution Approach 2:
The patent implements partial action by loading and processing only the portion of polygon data that is relevant to the current query or operation, rather than loading all high-resolution data. The bitmap index identifies which specific polygon segments need to be accessed, enabling partial processing that reduces memory usage and transmission bandwidth while maintaining full precision for the relevant portions.
3Adaptability or versatility
If traditional row-store architectures are used for spatial data, then transactional processing is supported, but analytical processing performance and compression efficiency are reduced
Solution Approach 1:
The patent implements a universal data architecture that supports both transactional and analytical workloads. The column-oriented storage with bitmap indexes provides efficient analytical processing through vectorized operations and compression, while the underlying data model maintains ACID transactional capabilities. This multi-functional system eliminates the need to choose between row-store and column-store architectures.
Solution Approach 2:
The patent segments spatial data into column-oriented structures that can be independently processed and compressed. By organizing data by columns rather than rows, the system enables efficient analytical queries on specific spatial attributes while maintaining transactional integrity through the database's ACID compliance. Each column can be compressed and processed separately, improving analytical performance without sacrificing transactional capabilities.
Data Source
AI summary
Polygons can be simplified from an original, higher resolution to a simplified, lower resolution such that the simplified versions of the polygons do not introduce errors and also do not render boundaries shared with other polygons invalid.