OrientDB — Multi-Model Database Combining Documents, Graphs & Key-Value

Introduction

OrientDB is a multi-model database that stores documents, graphs, key-value pairs, and objects in a single engine without requiring separate systems. Its extended SQL dialect lets developers query documents and traverse graph relationships in the same statement, eliminating the need for JOINs.

What OrientDB Does

• Stores and queries data as documents, graphs, key-value pairs, or objects using a single API
• Provides an extended SQL syntax that adds graph traversal operators like OUT(), IN(), and TRAVERSE
• Supports ACID transactions across both document and graph operations
• Scales horizontally with built-in multi-master replication and automatic sharding
• Includes OrientDB Studio, a web-based management interface for schema design, queries, and monitoring

Architecture Overview

OrientDB uses a record-based storage engine where each record has a unique Record ID (RID). Graph edges are stored as direct physical links between records, making traversals an O(1) pointer follow rather than an index lookup. The database supports schema-full, schema-less, and schema-mixed modes. It runs on the JVM and embeds a distributed coordination layer based on the Hazelcast protocol for cluster management.

Self-Hosting & Configuration

• Deploy via Docker or download the binary distribution and run server.sh / server.bat
• Set the root password via the ORIENTDB_ROOT_PASSWORD environment variable or during first startup
• Configure storage, networking, and memory in config/orientdb-server-config.xml
• For clusters, set up distributed mode by editing config/default-distributed-db-config.json
• Enable automatic backups with the built-in backup scheduler or use EXPORT DATABASE for manual dumps

Key Features

• True multi-model engine avoids the overhead of synchronizing separate document and graph databases
• Direct record links make graph traversals faster than traditional relational JOINs
• Extended SQL with graph syntax lowers the learning curve compared to purely graph-specific query languages
• ACID-compliant transactions work across document and graph operations in the same commit
• Built-in HTTP and binary protocol APIs with official drivers for Java, Node.js, Python, and more

Comparison with Similar Tools

• Neo4j — purpose-built graph database with the Cypher query language; OrientDB adds document and key-value models in the same engine
• ArangoDB — another multi-model database supporting documents, graphs, and search; OrientDB uses an extended SQL dialect while ArangoDB uses AQL
• MongoDB — leading document database; OrientDB adds native graph capabilities that MongoDB does not offer out of the box
• Dgraph — distributed GraphQL-native database; OrientDB provides a broader set of data models beyond graphs

FAQ

Q: Is OrientDB still actively maintained? A: OrientDB is maintained by SAP, which acquired the original company. The Community Edition remains open source under the Apache 2.0 license.

Q: Can I use standard SQL with OrientDB? A: OrientDB supports an extended SQL dialect that covers most standard SQL operations plus graph-specific commands like TRAVERSE and MATCH.

Q: How does OrientDB handle relationships compared to a relational database? A: Instead of foreign keys and JOINs, OrientDB stores direct physical pointers (edges) between records, making relationship traversal significantly faster.

Q: What is the maximum dataset size OrientDB can handle? A: OrientDB can handle billions of records across a sharded cluster. Individual nodes can manage hundreds of gigabytes with proper JVM tuning.

Sources

• https://github.com/orientechnologies/orientdb
• https://orientdb.org