Hi @seanbarzilay :
Is there a plan to support in-memory gremlin providers like redis?
I have done some performance testing on ES cluster( 5 powerful machines), with about 10,000,000 vertices and 200,000,000 edges. There are about 20 to 30 edges between adjacent vertices. When do traversal from vertex to vertex with 3 hops (v-v-v-v), it spent around 10 seconds.
The graph traversal is a exponential increase with time-complexity. Do you have any good practice on this? Maybe using the in-memory db like redis could enhance the performance?
Currently I'm not planing on adding any more providers to Unipop, but I'm hoping to finish my developers guide soon so people can implement more providers as they need.
Hi @seanbarzilay : Is there a plan to support in-memory gremlin providers like redis? I have done some performance testing on ES cluster( 5 powerful machines), with about 10,000,000 vertices and 200,000,000 edges. There are about 20 to 30 edges between adjacent vertices. When do traversal from vertex to vertex with 3 hops (v-v-v-v), it spent around 10 seconds. The graph traversal is a exponential increase with time-complexity. Do you have any good practice on this? Maybe using the in-memory db like redis could enhance the performance?