JanusGraph

JanusGraph is an open-source distributed graph database optimized for storing and querying massive-scale graphs containing billions of vertices and edges distributed across a multi-machine cluster. Originally developed by LinkedIn and later contributed to the Linux Foundation in 2017, JanusGraph was designed to address the scalability limitations of single-server graph databases by separating storage, indexing, and compute layers, allowing each component to scale independently. The database supports the Gremlin graph query language from Apache TinkerPop, enabling complex graph traversals, pattern matching, and path queries through a standardized API. JanusGraph uses a pluggable storage backend architecture, supporting Apache Cassandra, Apache HBase, Google Cloud Bigtable, and ScyllaDB as distributed storage engines, while offering Elasticsearch, Apache Solr, and Apache Lucene as indexing backends for efficient vertex and edge lookups by property values. This modular design allows organizations to choose the best combination of storage and indexing technologies for their specific latency, throughput, and consistency requirements. Key features include ACID-compatible transactions with optimistic concurrency control, multi-datacenter replication for high availability, vertex-centric indices for efficient local traversals, and support for both OLTP and analytical graph workloads. JanusGraph is used in production by major technology companies for fraud detection, identity resolution, recommendation systems, network topology management, and knowledge graph applications. The project is governed under the Linux Foundation's LF AI and Data Foundation and is available under the Apache 2.0 license.