I recently gained access to computing cluster (using a Linux OS). Has anyone had any experience running pellet (https://github.com/stardog-union/pellet) on a cluster? I have some large ABox datasets I would like to run using pellet.
Asked
Active
Viewed 132 times
0
-
Pellet is an open-source OWL reasoner, there is no cluster support. Distributed OWL reasoning is research topic, progress is stuck in the last decade as it's far away from being trivial. In addition, Pellet is not maintained anymore, expertise is nowadays contained inside the Stardog triple store, which in fact supports to setup the triple store on a cluster. – UninformedUser Feb 07 '18 at 14:26
-
Indeed, you can run Pellet on a single machine and increase the Java heap space. Anything beyond isn't supported. – UninformedUser Feb 07 '18 at 14:27
-
Thanks. I'm not sure how much heap space I'm allowed to use on the cluster ... I'll have to experiment :) – Bill Feb 07 '18 at 17:35
-
https://github.com/Galigator/openllet is very simmilar to Pellet. But a cluster mode would be a little tricky as reasoning have a bad complexity. A better approch is to cluster data more that cluster reasoning. – Galigator Feb 19 '18 at 13:21
-
@Galigator Thanks for the link to openllet. Not sure I understand your comment about clustering data, though. – Bill Feb 20 '18 at 20:50