We are connecting to janusgraph using java using an API. The API executes the query and returns the data.Facing memory issues frequently while using gremlin driver(org.apache.tinkerpop.gremlin.driver). 

The size occupied by the graph in db/cassandra/data directory is 690 MB. While we load the entire graph using API and gremlin driver, its takes approximately 2GB. 

Backend: Cassandra

Janusgraph: 0.5.2

Request you to suggest an alternative efficient way to perform the desired operation successfully.

Thanks & Regards,

Vinayak