Column-Oriented Graph Data Storage Engine
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Graph databases face limitations in scalability and computing performance when storing and retrieving graph data, making it desirable to store graph data in a column-oriented data store to improve these aspects.
Innovation Solution
A column-oriented data store population engine is used, comprising a row assignment engine, a column assignment engine, and an ID assignment engine, which assigns and populates vertex and edge rows with specific column families and IDs, allowing efficient storage and retrieval of graph data by separating vertex and edge data into distinct columns and sub-rows.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If graph data is stored in a graph database, then graph relationships can be represented using vertices and edges, but scalability and computing performance deteriorate
Solution Approach 1:
The patent introduces a column-oriented data store as an intermediary system between graph data representation needs and computing performance requirements. This intermediary converts graph database structures into column-oriented formats, enabling efficient storage and retrieval while maintaining graph relationship semantics through structured column families and row assignments.
2Stability of the object's composition
If graph data is stored in a graph database, then graph structures can be maintained, but scalability deteriorates
Solution Approach 1:
The patent applies segmentation by dividing graph data into distinct column families (vertex column families, edge column families) and further segmenting them into specific columns (e.g., vertex ID, edge ID, properties). This segmentation enables independent scaling of different data types while preserving graph structure integrity through systematic row assignments and foreign key relationships.
3Adaptability or versatility
If graph data is stored in a column-oriented data store, then scalability improves, but data structure complexity increases
Solution Approach 1:
The patent implements universality by designing column families that can serve multiple functions. For example, vertex column families store both vertex identification data and edge reference data, while edge column families maintain both directional and property information. This multi-functionality reduces the number of separate structures needed while preserving scalability benefits.
Data Source
AI summary
Methods and apparatuses are provided for storing graph data within a column-oriented data store. Graph data including vertex data describing one or more vertices in the graph and edge data describing one or more edges within the graph may be received. One or more vertex rows within the column-oriented data store may be assigned, whereby each vertex row of the one or more vertex rows is assigned to one vertex of the one or more vertices. One or more edge rows within the column-oriented data store may also be assigned, whereby each edge row of the one or more edge rows is assigned to one edge of the one or more edges. At least one vertex row and at least one edge row may be populated based on the graph data.