Abstract
Nowadays, the increased amount and complexity of connected data stimulated by the appearance of social networks has shed a new light on the importance of managing such data, especially handling information about the connections. The most natural way of representing connected data is to represent them as nodes connected with relationships forming a graph. The idea of storing data as a set of nodes and edges comprising a graph was implemented in various forms in data models used in the past. The network data model, developed in late 1960s, can be considered as the first data model, which most accurately incorporated this idea. However, it was not long before the relational data model appeared, and took over the entire database market for years, which it dominates even nowadays. Therefore, the objective of this article is to give a timeline overview of developed graph data storage solutions in order to gain insight into past, present and future trends of GDBMSs. Additionally, the most influential factors and reasons for changes in trends in GDBMSs' usage will be analyzed.
TopBackground
In general, there are various definitions of a graph database; De Virgilio et al. defined graph database as a “multigraph g=(N,E), where every node n∈N is associated with a set of pairs <key, value>, and every edge e∈E is associated with a label” (De Virgilio, Maccioni, & Torlone, 2013), whereas He and Singh defined graph database as a set of graphs D={G1, G2, …, Gm}, where graph G is denoted by (V, E), V being a set of all vertices, and E being a set of all edges (He & Singh, 2006).
Graph database model can be defined as a data model, in which “data structures for the schema and instances are modeled as graphs or their generalizations, and data manipulation is expressed by graph-oriented operations and type constructors” (Angles & Gutierrez, 2008). Graph database model consists of three components (Angles & Gutierrez, 2008):
- -
structural component (graph data structures),
- -
operational component (graph-oriented operators), and
- -
integrity component (integrity constraints).
Key Terms in this Chapter
Graph Pattern Matching: The process of finding all subgraphs matching a given graph query pattern in the original graph.
Graph Data Warehouse: A system used for data analysis, in which a data source is a graph database.
“Pure” Graph Database Management System: A system, which supports CRUD (Create, Read, Update, Delete) operations on graph data represented in the graph database model.
Graph Query Processing: The process of transforming a query written using a given graph query language syntax into a low-level query execution plan.
Graph Data Partitioning: The process of logically partitioning data stored in a graph database into segments, which can then be stored in different locations.
Graph Mining: The process of discovering patterns in data stored in a graph database through complex data analysis.
Graph Database Model: A conceptual representation of graph-like data in the form of graph nodes and edges.