Date
1 - 2 of 2
Back-end storage selection for better performance
Prahlad Negi <mecoolpa...@...>
New to janus graph, I've query about selection of backend storage(Cassandra, HBase, Bigtable, BerkeleyDB), which backend is best for small datasets and which one is best for large datasets? Is anyone done performance bench-marking before for small and large datasets separately?
HadoopMarc <bi...@...>
Hi Prahlad,
My opiniated response (without much arguments):
Cassandra or cql: best choice, this is where JanusGraph/Titan was built for.
HBase: only consider if you already have an HBase cluster running
BigTable: only consider if you want to run on GCS
BerkeleyDB: single machine, not a serious option
If you just want to play, anything goes. If you want something serious, you should be able to choose for yourself!
Best wishes,
Marc
Op dinsdag 8 september 2020 om 15:39:10 UTC+2 schreef Prahlad Negi:
New to janus graph, I've query about selection of backend storage(Cassandra, HBase, Bigtable, BerkeleyDB), which backend is best for small datasets and which one is best for large datasets? Is anyone done performance bench-marking before for small and large datasets separately?