Does Dgraph follows data locality principle so it allows to send requests and algorithms to machines where vertices and edges are stored to process them?
If so, is there a good documentation to read it about?
Each “subgraph” processing is performed on the Alpha (the data node) where data is available. But then, to merge the results, we pull it to the data node that is processing the query. I think this could be further optimized with the use of a query planner, something that we have on our roadmap.
We are working on more documentation, we already have some documents internally that will be out as soon as they are ready. You could look into the code starting here dgraph/query.go at master · dgraph-io/dgraph · GitHub for more details until then.