Search code examples
janusgraphgremlin-server

How does JanusGraph.open() work and how to scale?

I am evaluating different Graph databases and libraries etc.. and JanusGraph seems to be providing most of what I need. I do have a couple of questions:

  1. I would like to connect to it via Gremlin Server with Cluster option, however I don't seem to see any Java examples to handle transaction rollbacks etc at all.
  2. And if I was to use JanusGraphFactory.open("...") option, how exactly does this work? Would it mean the entire Graph is loaded into memory in JVM?
  3. If the entire graph is loaded into memory, how would one scale up and different JVMs are keeping up to date with each other?

Thanks & regards Tin

Solution

  • I would like to connect to it via Gremlin Server with Cluster option, however I don't seem to see any Java examples to handle transaction rollbacks etc at all.

    Connecting to Gremlin Server involves sessionless communication, meaning each request equals one transaction. You can connect with a session but it is not typically encouraged for most use cases.

    And if I was to use JanusGraphFactory.open("...") option, how exactly does this work? Would it mean the entire Graph is loaded into memory in JVM?

    It just creates a reference to the data and provides a Graph instance from which you can create a GraphTraversalSource to interact with for spawning traversals. It doesn't load any of that data into memory just by virtue of calling it.