Hi Chris,

Thank you very much for your reply. I am trying to join in the development of janusgraph, so I tried to understand the code architecture of janus. But I am confused about the what janusgraph did based on Tinkerpop and gremlin. It seems that janusgraph do with the storage backend, but the tinkerpop knows nothing about the storage, so how did janusgraph execute a query when the core engine(tinkerpop) knows nothing about the storage? I cannot understand the query process, so I hope the developers can help me on this, thanks.

在 2018年9月11日星期二 UTC+8上午9:36:38，Chris Hupman写道：

Hi Raul,

This is more of a question for the JanusGraph users google group. That being said, here are some links that help explain the relationship. 

Here's the architecture overview from the docs

Here's a blog I wrote that covers the relationship in the intro

Also here's a free book by Kelvin Lawrence on gremlin that also covers JanusGraph

On Monday, September 10, 2018 at 2:19:42 AM UTC-7, Raul Guo wrote:

I am a little confused about the relationship between JanusGraph and gremlin, the storage backend(HBase, Cassandra, .etc) is configured by JanusGraph, but and the core traversal strategy(GraphTraversal, GraphTraversalSource, .etc) are defined in gremlin, so how did janusgraph interact with the storage backend now that its core calculate engine(which is gremlin) knows nothing about the storage at all? Can you help me to know about what gremlin did in janusgraph, and what janusgraph did based on gremlin? Thanks in advance.