Tree-Structured Graph Storage for Fewer I/O Operations
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing data storage methods for graph structures in memories and storage devices lack associativity between vertex and edge data, leading to inefficient data access and storage, requiring numerous I/O operations due to chaotic storage states and separate storage of vertices and edges.
Innovation Solution
A tree-structure storage method is employed, where data of a graph structure is organized into a tree structure in memory and disk blocks, associating index and attribute information to reduce I/O operations by storing data of vertices and edges in associated positions, thereby reflecting the local properties of the graph structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If graph structure data is stored using key-value pairs with vertices and edges stored separately, then storage simplicity is improved, but data access efficiency deteriorates due to lack of associativity and numerous I/O operations required
Solution Approach 1:
The patent combines vertex data and edge data into a unified tree-structure storage format. Instead of storing vertices and edges separately as key-value pairs, the invention merges them into a single structured representation where vertices and their incident edges are stored together in a hierarchical format, eliminating the need for separate storage operations and improving data access efficiency through associative storage
Solution Approach 2:
The patent transforms the flat key-value storage model into a hierarchical tree structure with multiple levels. The storage format organizes data into a hierarchical representation with vertices at one level and their incident edges at another level, adding a structural dimension to the storage model that enables more efficient data access patterns while maintaining storage simplicity
2Adaptability or versatility
If vertices and edges are stored in chaotic separate positions, then storage flexibility is improved, but storage efficiency deteriorates due to increased I/O operations
Solution Approach 1:
The patent merges vertex and edge storage into a single unified structure where related data is stored together. This consolidation reduces the number of separate I/O operations required to access graph data, as vertices and their incident edges are retrieved in a single operation rather than requiring multiple separate reads from different storage locations
Solution Approach 2:
The patent performs preliminary organization of graph data into a tree structure during the storage phase. By pre-organizing vertices and their incident edges into a hierarchical format before actual storage, the system eliminates the need for chaotic separate positioning and reduces subsequent I/O operations, as the data is already arranged for efficient access patterns
3Ease of operation
If graph data is stored without reflecting local properties, then storage uniformity is improved, but query efficiency deteriorates due to lack of associative storage
Solution Approach 1:
The patent applies local quality by organizing graph data to reflect local properties of the graph structure. The tree-structure storage format preserves the local relationships between vertices and their incident edges, allowing queries to efficiently access locally-related data without requiring uniform treatment of all data. This local organization enables faster query execution while maintaining overall storage uniformity through the consistent hierarchical format
Data Source
AI summary
Methods, apparatus, computer devices, products, and storage medium are described for generating tree-structure storage data for a graph structure, applicable to the fields of cloud computing or the Vehicle of Internet. One method includes: obtaining data of a graph structure comprising vertex data of at least one vertex of the graph structure and edge data of at least one edge of the graph structure; constructing memory data of a tree structure for the data of the graph structure in a memory of the computer device; constructing a plurality of disk blocks in a storage device based on the memory data of the tree structure; and storing the data of the root node, the data of each first-subnode, and data of a subtree using each second-subnode as a root node into corresponding disk blocks respectively, to obtain the storage data of the tree structure for the graph structure.