PLATFORM, SYSTEM, PROCESS FOR DISTRIBUTED GRAPH DATABASES AND COMPUTING

• US 20170364534A1
• Filed: 06/15/2017
• Published: 12/21/2017
• Est. Priority Date: 06/15/2016
• Status: Active Grant

• Current Assignee: Zhejiang Chuaglin Technology Co., Ltd.
• Original Assignee: Chen Zhang, Jing Wu
• Inventors: ZHANG, Chen, WU, Jing
• Granted Patent: US 10,409,782 B2
• Time in Patent Office: Days
• CPC Class Codes: G06F 16/182 Distributed file systems
• CPC Class Codes: G06F 16/256 in federated or virtual databases
• CPC Class Codes: G06F 16/26 Visual data mining; Browsing structured data
• CPC Class Codes: G06F 16/284 Relational databases
• CPC Class Codes: G06F 16/9024 Graphs; Linked lists G06F16/9027 takes precedence
• CPC Class Codes: G06N 20/00 Machine learning
• CPC Class Codes: G06N 5/04 Inference or reasoning models

Abstract

Systems for distributed graph data management that use graph structures to model data and respond to different types of queries. Systems that provide analytic tools and machine learning based on graph structures. Systems for distributed graph data management that provide packaged solutions for different application areas.

Claims

1. A system for generating a dynamic visualization of a graph structure, the system comprising a processor configured to:

load the graph structure into distributed memory of a plurality of nodes based the graph topology, the graph structure having vertices and edges and a graph topology;

generate visual elements of an interface by traversing the graph structure over the plurality of nodes from a starting vertex node to an end node, the interface comprising a form field to receive a graph-based search as input;

determine a set of vertices and edges of the graph structure satisfying the graph-based search;

update the visual elements of the interface to indicate the set of vertices and edges; and

activate a selected visual element corresponding to a vertex of the set of vertices and edges to expand its depth neighbors as additional visual elements of the interface.

2. The system of claim 1, wherein a distributed in-memory computation layer optimizes graph data placement on the plurality of nodes using the graph topology so that adjacent graph data are loaded onto a node or proximate nodes.

3. The system of claim 2, wherein each vertex has a degree or total number of edges connected to the vertex, wherein the distributed in-memory computation layer optimizes the graph data placement on the plurality of nodes using the degree.

4. The system of claim 2, wherein the distributed in-memory computation layer optimizes the graph data placement on the plurality of nodes based on physical node information such as location and workload.

5. The system of claim 1, wherein a distributed in-memory computation layer optimizes graph data placement on the nodes using a plurality of daemons and a central daemon for load balancing, each daemon connecting to a node, the plurality of daemons connecting to the central daemon, the central daemon interfacing with a node status table and a data placement table for the load balancing.

6. The system of claim 1, wherein the processor is configured to:

store and update a vertex table structure for the vertices of the graph structure and an edge table for the edges of the graph structure;

assign unique vertex identifiers for the vertices, the vertex table linking each vertex identifier to a set of vertex data values for the corresponding vertex;

assign unique edge identifiers for the edges, the edge table linking each edge identifier to a set of edge data values for the corresponding edge; and

determine the set of vertices and edges using a vertex identifier or an edge identifier.

7. The system of claim 6, wherein the processor is configured to assign a timestamp to a data cell of the vertex table or the edge table;

and update the visual elements with a time dimension based on the timestamp.

8. The system of claim 1, wherein the processor is configured to:

store the graph structure by storing a graph table with vertex data for the vertices of the graph structure and edge data for the edges of the graph structure;

assign unique identifiers for the vertices and the edges, the graph table linking each identifier to a set of data values for a corresponding vertex or edge;

determine the set of vertices and edges using an identifier.

9. The system of claim 8, wherein the processor is configured to assign a timestamp to a data cell of the graph table;

and update the visual elements with a time dimension based on the timestamp.

10. The system of claim 1, wherein the processor is configured to assign a unique graph identifier for the graph structure, and unique vertex identifiers for the vertices, and unique edge identifiers the edges, wherein an edge identifier for an edge is a combination of a set of vertex identifiers for vertices that the edge connects.

11. The system of claim 1, wherein the visual elements of the interface implement snapshot isolation on the graph structure data, and exploit snapshot isolation to track graph evolution over a plurality of time periods.

12. The system of claim 1, wherein the graph structure is implemented on top of a column store to maintain original structures of data, while adding additional graph structure to its data, wherein the processor is configured to model property graphs and persist property graphs with index free adjacency, wherein the processor allows multiple query methods on top of the same data.

13. The system of claim 1, further comprising a graph data management platform and a graph analytic layer, the graph data management platform for storing and process input data from data sources for the graph structures, the graph analytic layer for providing a set of tools for queries for the input to compute the visual elements from the distributed graph structures.

14. The system of claim 1, wherein the processor is configured to to receive data from data transfer tools to load onto the distributed memory, a graph management system to store, update, query and compute the graph structure from the input, a graph analytic engine to determine graph operations for computations of the set of vertices and the edges in respond to the graph-based search, a machine learning engine to optimize graph parameters for the graph structure, a client application programming interface for data exchange between the system and the interface, and a visualization dashboard interface to generate the visual elements from the computations.

15. A method for generating a dynamic visualization of a graph structure comprising:

loading the graph structure into distributed memory of a plurality of nodes based the graph topology, the graph structure having vertices and edges and a graph topology;

generating visual elements of an interface by traversing the graph structure over the plurality of nodes from a starting vertex node to an end node, the interface comprising a form field to receive a graph-based search as input;

determining, by a processor, a set of vertices and edges of the graph structure satisfying the graph-based search;

updating, by the processor, the visual elements of the interface to indicate the set of vertices and edges;

assigning a timestamp to a data cell of the vertex table or the edge table; and

updating the visual elements with a time dimension based on the timestamp; and

activating, by the processor, a selected visual element corresponding to a vertex of the set of vertices and edges to expand its depth neighbors as additional visual elements of the interface.

16. The method of claim 15, wherein each vertex has a degree or total number of edges connected to the vertex, the method further comprising optimizing graph data placement on the plurality of nodes using the graph topology and the degree so that adjacent graph data are loaded onto a node or proximate nodes.

17. The method of claim 15, further comprising optimizing graph data placement on the plurality of nodes using a plurality of daemons and a central daemon for load balancing, each daemon connecting to a node, the plurality of daemons connecting to the central daemon, the central daemon interfacing with a node status table and a data placement table for the load balancing.

18. The method of claim 15, further comprising:

storing and updating a vertex table structure for the vertices of the graph structure and an edge table for the edges of the graph structure;

assigning unique vertex identifiers for the vertices, the vertex table linking each vertex identifier to a set of vertex data values for the corresponding vertex;

assigning unique edge identifiers for the edges, the edge table linking each edge identifier to a set of edge data values for the corresponding edge; and

determining the set of vertices and edges using a vertex identifier or an edge identifier.

19. The method of claim 15, further comprising:

storing the graph structure by storing a graph table with vertex data for the vertices of the graph structure and edge data for the edges of the graph structure;

assigning unique identifiers for the vertices and the edges, the graph table linking each identifier to a set of data values for a corresponding vertex or edge;

determining the set of vertices and edges using an identifier.

20. The method of claim 15, further comprising:

assigning a unique graph identifier for the graph structure, and unique vertex identifiers for the vertices, and unique edge identifiers the edges, wherein an edge identifier for an edge is a combination of a set of vertex identifiers for vertices that the edge connects.