Bryan Garber da Silva wrote:
Hi!
I'm wanting to use Jena with a persistent storage for the ontology. On the
site, it's said that it supports various SQL databases, such as PostgreSQL,
MySQL, and so on. I'm curious to know if there is some kind of support for
NoSQL databases, such as Apache's Cassandra, Google's BigTable, and so on.
Any help here?
Hi Bryan,
no, there is no support for NoSQL databases, even if TDB is actually not
using SQL (but it's not distributed, it's a scale vertically only solution
at the moment).
You might be interested in looking at the sources of some experiments and
not supported prototypes:
- https://github.com/afs/TDB-V
- https://github.com/afs/TDB-BDB
- https://github.com/castagna/hbase-rdf
TDB-V shows you how you could use a key-value store (such as Project Voldemort)
as a networked hard disk to store the blocks of your B+Tree indexes.
TDB-BDB shows you how you can plug into TDB a different indexing subsystem.
HBase-RDF is a (not currently working) playground to see how RDF could be
stored and queries over HBase. The design is taken from TDB (same approach
at implementing Jena Graph SPIs). You have questions/comments or some time
to spare, I'd love to see it running (and run a few BSBM tests... I don't
expect it to be "fast", but I am curious).
I don't think (after having read the book) Cassandra supports fast/efficient
range queries (and you need those for BGP for Graph SPI .find(...)). But, I
could be wrong and an experiment would be interesting.
So, not a great help... but someone to discuss with, if you want and, perhaps,
collaborate with.
Last but not least, I would like to ask you: are you really sure you need
a NoSQL (massively scalable) solution for your data? Have a look at TDB
(it's NoSQL in the sense that's a native RDF storage and indexing systems
and do not uses a relational database. It can have very good performances
(in particular if you have a machine with a lot of RAM!)):
- http://openjena.org/wiki/TDB
Have a look at HBase-RDF and let me know.
Nice to meet you.
Paolo
Thank you very much!
