What is a Graph Database?
A graph is a data structure describing the real-world relationships among entities. A graph database management system (henceforth, a graph database) is then an online database management system in support of semantic operations based on the graph structure. It uses the vertex and edge to represent and store data, and supports CRUD operations.
Different from the traditional relational database management system (RDBMS, henceforth, a relational database), a graph database directly stores the relationships among entities, which in relational databases resort to expensive and time-consuming JOINs by way of foreign keys. When compared with the traditional tabular storage model, the graph data model provides a natural and hi-fi representation of complex real-word relationships.
The following figure describes relationships among entities like director, movie, actor/actress, movie genre, etc. using the graph database Galaxybase.
Galaxybase Native Distributed Graph Database
Regarding the storage model, traditional relational databases store and manage data in a tabular form consisting of rows and columns. Graph databases, on the other hand, are optimized and designed for efficient graph data store and queries, using vertices and edges as basic storage units. Typically accessed directly from an application, a graph database exposes a graph data model through CRUD operations. It is the equivalent of online transactional processing (OLTP) databases in the relational world.
In terms of the underlying storage, the graph database landscape can be divided into native and non-native. A native graph database stores data directly in the graph structure, while a non-native graph database serializes graph data into a relational database or some other type of database. Native graph databases are specially optimized and designed for storing and querying graph data, and therefore have better performance in deep-link queries.
Galaxybase graph database developed by CreateLink Tech. adopts a native distributed parallel framework. As its underlying storage and compute engine are both independently developed，and together with its remarkable scalability, Galaxybase is deeply optimized for efficient graph data store and processing at scale in the era of big data.
Why Graph Database?
The inevitable correlations among people, events, objects, and places result to the complex relationships hidden in data that describe those entities, and the relationships are becoming a critical role in making business decisions. As enterprises undergo digital transformation, data is now an important strategic asset for organizations. Great at processing connected data, graph databases enable organizations to extract actionable insights from interconnected data which might otherwise be overlooked when using traditional technologies. Currently, 70% of the world’s top 100 companies are using graph databases to obtain business values. For instance, google builds a large-scale knowledge graph, based on which semantic searches are enabled; Ebay constructs its complex logistics network to improve users’ shopping experience; Facebook utilizes graph database to construct a social network for people to connect with friends, work colleagues or people they don’t know, online. The status of graph has increased such that Gartner, a research and advisory firm, forecasts that 80% of data and analytics innovations will be made using graph technology by 2025.
Graph databases boast the following advantages compared with relational databases:
Data model is more intuitive
Data model is more intuitive
A graph data model directly reflects business scenarios in such an intuitive way that it greatly improves the communication efficiency between business staff and engineers.
Query Language is more concise
Query Language is more concise
Take OpenCypher, one commonly-used graph query language as an example. The lines of codes written in OpenCypher for data-relationship questions are sharply decreased compared to those in SQL, which helps a lot in speeding up the development process.
Link analysis is more efficient
Link analysis is more efficient
Take Galaxybase for example. In a test against large network dataset ‘who-trust-whom’, Galaxybase outperforms MySQL in relationship questions that spans more than 3 hops by an exponential order of magnitude: 80 times faster in 3-hop queries; over 33,000 times faster in 5-hop queries.
Get Started for Free with Galaxybase Single Server
Get Started for Free with Galaxybase Single Server
Start Free
CreateLink is a leading graph DBMS vendor based in China. It provides services ranging from relationship mining, deep-link analytics, visualized analysis to graph intelligent computing on massive heterogeneous data from various sources.
 Buisness Cooperation：400-882-6897
 Pre-Sales Consulting：0571-88013575、0571-88016275
 partner@chuanglintech.com
 L12, Building E2, Zhi Hui Zhong Chuang Center, No.3 Xiyuan
No.8 Road, Xihu District, Hangzhou, China