InMemory Storage Backend

JanusGraph ships with an in-memory storage backend which can be used through the following configuration:

1
storage.backend=inmemory

Alternatively, an in-memory JanusGraph graph can be opened directly in the Gremlin Console:

1
graph = JanusGraphFactory.build().set('storage.backend', 'inmemory').open()

There are no additional configuration options for the in-memory storage backend. As the name suggests, this backend holds all data in memory. Shutting down the graph or terminating the process that hosts the JanusGraph graph will irrevocably delete all data from the graph. This backend is local to a particular JanusGraph graph instance and cannot be shared across multiple JanusGraph graphs.

Ideal Use Case

The in-memory storage backend was primarily developed to simplify testing (for those tests that do not require persistence) and graph exploration. The in-memory storage backend is NOT meant for production use, large graphs, or high performance use cases. The in-memory storage backend is not performance or memory optimized. All data is stored in the heap space allocated to the Java virtual machine.