Exploring Nebula Graph: Building Powerful Graph Database Applications

01 Introduction

Graphs are among the most flexible and high‑performance data structures. A graph database stores data as vertices and edges, enabling efficient retrieval of massive information networks. Early graph databases like Neo4j, JanusGraph, and Amazon Neptune gained traction, and the newer open‑source distributed graph database NebulaGraph has emerged to meet growing market demand.

02 What is NebulaGraph?

Nebula Graph is an open‑source distributed graph database capable of handling datasets with billions of vertices and trillions of edges. It provides high‑throughput, low‑latency read/write operations, built‑in ACL mechanisms, and user authentication for secure access.

It offers linear scalability and snapshot‑based data recovery. The query language nGQL is fully developed in‑house, and future compatibility with the OpenCypher interface will allow Neo4j users to transition seamlessly.

03 What Can Nebula Graph Do?

NebulaGraph can be applied to a variety of graph‑based business scenarios, reducing the time needed to convert data to relational databases and avoiding complex queries.

Fraud detection : Financial institutions can model large volumes of transaction data as a graph to uncover hidden relationships between fraudulent activities and devices, enabling detection of fraud rings.

Real‑time recommendation : NebulaGraph processes real‑time visitor information to deliver precise article, video, product, or service recommendations.

Knowledge graph : Natural language can be transformed into a knowledge graph stored in NebulaGraph; semantic parsers interpret questions and retrieve answers from the graph.

Social network : Human relationship data, a classic graph use case, can be handled at billions of nodes and trillions of edges, providing fast friend recommendations and job queries under massive concurrency.

04 Nebula Graph Architecture

NebulaGraph consists of three services: Graph Service, Meta Service, and Storage Service, following a storage‑compute separation design.

Each service runs as an executable binary (nebula-graphd, nebula-metad, nebula-storaged) on one or more machines to form a cluster. Meta Service manages schema, cluster configuration, and user permissions; Graph Service handles computation requests; Storage Service stores the actual graph data.

05 How to Use Nebula Graph

A Nebula Graph instance contains one or more graph spaces, each physically isolated, allowing different datasets to be stored within the same instance.

To insert data, a schema must be defined, consisting of vertex types (tags) and edge types.

1. Install Nebula Graph

On Ubuntu 20.04, download the package with wget from https://osscdn.nebulagraph.com.cn/package/2.6.1/nebula-graph-2.6.1.ubuntu2004.amd64.deb.

2. Start Nebula Graph services

Management is done via the nebula.service script.

sudo nebula.service start all

sudo nebula.service stop all

3. Use nGQL statements

Create a graph space: CREATE SPACE [IF NOT EXISTS](parameter) Define edges: CREATE EDGE [IF NOT EXISTS] (parameter) Define vertices (tags): CREATE TAG [IF NOT EXISTS] (parameter) Insert vertex data: INSERT VERTEX VALUES (data) Insert edge data: INSERT EDGE VALUES (data) After defining the schema, data can be imported into the graph space, and various graph queries and analyses—such as social network analysis, path analysis, clustering, recommendation systems, and custom knowledge‑graph construction—can be performed.

06 Related Resources

Nebula Graph Database Manual