JanusGraph server vs. embedded mode.


HadoopMarc <m.c.d...@...>
 


Hi Manoj,

It totally depends on your total system setup: how many end users, application servers, backend nodes, security demands, etc.

For just issuing a gremlin query and getting results back, the deployment mode does not matter. Gremlin server, however, gives you far more control regarding security, scaling, reliability, non-JVM client connections, etc.

Hope this, helps, otherwise sketch your system setup.

Cheers,   Marc


Op donderdag 23 februari 2017 07:22:58 UTC+1 schreef Manoj Waikar:

Also, is there any equivalent of the Frames component of the older Tinkerpop stack in the current stack?


Manoj Waikar <mmwa...@...>
 

Also, is there any equivalent of the Frames component of the older Tinkerpop stack in the current stack?


Manoj Waikar <mmwa...@...>
 

Hi,

What is the recommended approach of working with JanusGraph? Should one - 
 (a) Setup Gremlin (JanusGraph) server and submit Gremlin queries to the server? Or,
 (b) Embed JanusGraph inside the application and use the native Java API to execute queries against the graph within the same JVM?
 
So, what are the pros and cons of running a JanusGraph server and then connecting to it vs. directly connecting to a data store using JanusGraphFactory.open() method?

Also, in the Appendix A. API Documentation (JavaDoc), it is mentioned - 
We strongly encourage all users of JanusGraph to use the Gremlin query language for any queries executed on JanusGraph and to not use JanusGraph’s APIs outside of the management system.

Does it mean, that only for creating schema should we use the JanusGraph’s management APIs and for everything else, use Gremlin?

Thanks,
Manoj.