Does anyone know any good RDF store based on Hadoop ?
This might be related in your context:
- not Hadoop, but HBase (to be concrete ;) ): "Storing RDF data into HBase?"
- "Best practices for processing RDF data using MapReduce (via Hadoop)?"
- "Native Triple Stores Performance"
I guess, that SHARD (HDFS based) is a hot potato.
I think @castagna may have made some progress since posting http://www.semanticoverflow.com/questions/716/storing-rdf-data-into-hbase
See: http://mail-archives.apache.org/mod_mbox/incubator-jena-users/201102.mbox/browser
I'm curious why you are interested in HBase? I speculate that range queries in a map-reduce implementation isn't going to perform very well. I suspect other more mature stores that implement their own federation, such as Allegrograph and Virtuoso, may fair better if performance alone is your concern.
Incomplete, unsupported and experimental... but it's very close to a first functioning prototype which would allow us to run some tests: https://github.com/castagna/hbase-rdf.
Todo:
- have a first prototype working
- run BSBM
- pluggable layouts
- experiment with different layouts
Paolo