Hi,
I'm evaluating the create/insert performance of OrientDB (version 2.0.12
Linux) with my 64-bit Linux VM, 2 core processor @ 2.6Ghz. Out of 16GB RAM,
I allocated 12GB to the server -Dstorage.diskCache.bufferSize
Below is my Java statement to create vertex:
for (int i= 0; i < 10000; i++){
graph.addVertex("class:User", "id", i, "name", "n"+i);
}
I tried the above statement in different ways: plocal, remote and
distributed. I also tried SQL request with REST API.
plocal:
final OrientGraphFactory factory = new OrientGraphFactory(
"plocal:orientdb-community-2.0.12/databases/testDB", "admin", "admin");
factory.declareIntent(new OIntentMassiveInsert());
OrientGraphNoTx graph = factory.getNoTx();
This took 4s to create 10k vertex.
remote:
final OrientGraph graph = new OrientGraph("remote:192.168.x.x/testDB",
"admin", "admin");
graph.declareIntent(new OIntentMassiveInsert());
This took 9s.
Using 2 node distributed machines where the 2nd machine has 4GB to disk
cache, it took 16s (using hazelcast tcp-ip instead of multicast)
I also observed that using OrientGraphNoTx actually took 15s while the
distributed way took 80s...
Lastly, I tried with REST API with Apache Fluent POST request and this took
50s.
Any advice on how can the performance with plocal way be further improved
and the distributed way too? Is there anything wrong with the distributed
way as it seemed to performed worse.
Thanks,
Weili
--
---
You received this message because you are subscribed to the Google Groups
"OrientDB" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to [email protected].
For more options, visit https://groups.google.com/d/optout.
