PROVIDER Β· Native multi-model graph and document database
ArangoDB
ArangoDB is a native multi-model database sharing one storage and transaction across documents, graphs, and key-value with unified AQL, removing dual-store sync for knowledge graphs and businesses needing both relationships and document attributes.
Pros
Cons
Product Information
Detailed Notes
Overview
ArangoDB was released in 2011 by German company triAGENS, now maintained by ArangoDB GmbH, built around native multi-model design where the same data is modelled and queried as documents, graphs, and key-value through one language, AQL.
For Backend & Database work the value is one system fewer, since entity relationships plus document attributes no longer need MongoDB and Neo4j side by side, and a single engine handles multi-hop traversal with document aggregation while AQL supports join-like linking and recursion, fitting knowledge graphs, permission graphs, and recommendation paths.
The limits are scale and ecosystem, because deep or highly concurrent traversals are bounded by shard locality with very large graphs still favouring dedicated engines, ecosystem and talent are smaller than MongoDB and Neo4j, and enterprise capabilities such as smart graphs need licences.
Core Strengths
- Native multi-model design shares one storage and transaction across documents, graphs, and key-value
- AQL unifies querying with join-style linking, recursion, and traversal
- Multi-hop knowledge graph queries run inside one database, removing sync pipelines
- Clustering and sharding distribute graph data to scale out
Product Ecosystem
AQL and multi-model
One language spans document filters, traversal, and aggregation with LET and RETURN composing complex pipelines.
Clustering and sharding
Collections shard across nodes with inter-shard hops adding cost for traversals, and enterprise SmartGraphs optimise very large graphs.
Ecosystem and deployment
ArangoGraph managed service, a Kubernetes operator, and drivers for Python, Java, Go, and JavaScript.
Drawbacks
- Very large graphs with billions of edges trail dedicated graph engines
- Traversals pay inter-shard cost after sharding, demanding careful shard key design
- Smaller ecosystem and talent than MongoDB or Neo4j with scarce Chinese material
- Enterprise features such as SmartGraphs need licences
Use Cases
- Knowledge graphs and relationship modelling (β β β β β ) - Traversal plus document attributes
- Permission and recommendation paths (β β β β ) - AQL recursion is convenient
- Businesses mixing documents and graphs (β β β β β ) - Avoids dual-store sync
- Billion-edge graph analytics (β β ) - Prefer a dedicated graph engine
Pricing
| Plan | Price | Scope |
|---|---|---|
| Community | $0 | Open source with multi-model and basic clustering |
| Enterprise | Annual per node | SmartGraphs, backup, and support |
| ArangoGraph managed | Per size and usage | Hosted with monitoring and upgrades |
The community edition is free with basic clustering for small to mid-scale mixed graph and document needs, while enterprise fees buy SmartGraphs and support, and graph-heavy workloads may compare dedicated engines instead of upgrading.
FAQ
-
ArangoDB or Neo4j? Neo4j for deep traversal and graph algorithms, ArangoDB when the business needs document attributes, aggregation, and relationships together in one store.
-
Is AQL easy to learn? Its declarative pipeline resembles JavaScript, so developers knowing SQL and JSON pick it up in days, with traversal semantics and sharding effects the hard part.
-
Can it replace MongoDB? Document capability covers most cases with graphs on top, while ecosystem, managed options, and community trail MongoDB, so weigh operations and talent before migrating.