Also be aware of this PR https://github.com/JanusGraph/janusgraph/pull/392 which will allow you to create new graphs dynamically (i.e. post server start) and store/manage configurations for these graphs through the ConfigurationManagementGraph API, where the configurations for your graphs are persisted rather than stored on disk in .properties files.

The PR also introduces the notion of a `Template_Configuration`, which should be useful in your case. If are _always_ connecting to one Cassandra cluster, then you can define that appropriate configuration (sans the keyspace parameter), and then call `ConfiguredGraphFactory.create("<graphName>")` to dynamically create and connect to a new graph whose keyspace would be equivalent to "<graphName>". For more specifics, about how to use the `ConfigurationManagementGraph` and `ConfiguredGraphFactory` you can look at the documentation I've written for it in that link.