Is a bigger traversal/query on titan/gremlin possible without making too many network requests (if possible with just one!) to the storage backend (for example cassandra)? Because if there are let us say 100'000 vertices and edges that need to get loaded and it would make a separate request for each vertex and edge. That would result in a long loading time (multiple seconds). But if the whole traversal can be done very close to the database or in the database itself it would be faster.
0
votes
1 Answers
0
votes
With Titan 1.0, the query optimizer is much smarter about batching together requests to the backend system; the query you imagine won't need a separate round trip for every vertex and edge. You have to enable the batching explicitly (http://s3.thinkaurelius.com/docs/titan/1.0.0/titan-config-ref.html) and there are still a few rough edges that I would imagine will be sanded off in a future release of Titan.
