Tiled Data Tree Structure for Efficient Query Processing

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional data processing methods become inefficient and costly as data volume increases, particularly due to delays and high memory and processing loads when handling large datasets, and existing techniques for tile-based distribution and parallel processing do not adequately address these issues.

Innovation Solution

A computing system that uses a tree structure to subdivide data into tiles, allowing for parallel processing and caching precomputed information, which enables efficient computation and presentation of data by recursively determining information from sub-tiles and merging precomputed data with other information to quickly respond to user requests.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If data is aggregated for computation and presentation, then computation can be performed on complete datasets, but processing time and memory load become prohibitively expensive as data volume increases

Engineering Contradiction:
Improvecomputation accuracyVSAvoidprocessing time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent divides large datasets into hierarchical tile structures, where data is organized into parent tiles and child tiles at multiple levels. This segmentation allows the system to process only relevant tiles rather than entire datasets, dramatically reducing processing time while maintaining computation accuracy through selective aggregation of precomputed tile data.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements precomputation of data aggregates at the tile level before queries are executed. By precomputing and storing aggregate information for each tile in the hierarchical structure, the system avoids performing expensive computations on large datasets during query execution, thus reducing processing time while preserving accuracy through the use of precomputed values.

Inventive Principle:
Principle #10Preliminary action

2Quantity of substance

If data is sharded into multiple data shards, then data storage capacity increases, but delays and latencies are introduced when data from multiple shards are required for computation

Engineering Contradiction:
Improvedata storage capacityVSAvoidprocessing speed
Core Design Contradiction:
Quantity of substanceVSSpeed

Solution Approach 1:

The patent merges multiple sharded data tiles into a hierarchical structure where parent tiles contain aggregate information from their child tiles. This merging allows the system to retrieve precomputed aggregate data from parent tiles without needing to access and combine data from multiple child shards, thus maintaining storage capacity while eliminating the latency associated with cross-shard data aggregation.

Inventive Principle:
Principle #5Merging (Combining)

Solution Approach 2:

The patent performs preliminary aggregation of data from multiple shards at the tile level during data ingestion and preprocessing. By precomputing aggregates across shard boundaries and storing them in the hierarchical tile structure, the system eliminates the need for expensive cross-shard joins during query execution, thus improving processing speed while maintaining storage capacity.

Inventive Principle:
Principle #10Preliminary action

3Measurement precision

If data is stored in detailed granular form, then query accuracy is improved, but memory load and processing requirements increase significantly

Engineering Contradiction:
Improvequery accuracyVSAvoidmemory load
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The patent segments data into a hierarchical tile structure where detailed granular data is stored at the leaf tile level while aggregate summaries are stored at parent tile levels. This segmentation allows the system to load only the necessary level of detail into memory based on query requirements, reducing memory load while maintaining query accuracy through selective access to detailed data when needed.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements a progressive refinement approach where queries first access aggregate data from parent tiles to determine if detailed data is needed. Only when necessary does the system load detailed granular data from child tiles into memory, thus minimizing memory load while ensuring query accuracy is maintained when detailed information is required.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentEP3410315B1Systems and methods for using tiled data
Publication Date: 2022.12.07 PALANTIR TECHNOLOGIES INC
  • EP3410315B1 patent drawingFigure 1
  • EP3410315B1 patent drawingFigure 2A~2B
  • EP3410315B1 patent drawingFigure 3

AI summary

Systems, methods, and computer readable media are provided for using tiled data. An information request for an area may be received. Information for the area may be accessed. The information for the area may be stored in a tree structure such that the information is subdivided into one or more levels of tiles. Individual tiles may correspond to a portion of the area. One or more tiles may include precomputed information for the corresponding portion(s) of the area. A response to the information request may be determined based on the precomputed information.