W3C home > Mailing lists > Public > semantic-web@w3.org > April 2007

Re: resources for network-based/hierarchical RDF store

From: Andreas Langegger <andreas.langegger@gmx.at>
Date: Thu, 26 Apr 2007 23:06:55 +0200
Message-ID: <4631146F.70702@gmx.at>
To: semantic-web@w3.org

Hi Paul,

thanks for the hint. I know Mulgara and I've also read about these 
issues with Kowari and Tucana. However, at first I was looking for a 
SPARQL-implementation since we wanted to be as close as possible to 
evolving standards. I will take a closer look on Mulgara again.

We need support for aggregate functions and would have to extend SPARQL 
anyway. Are there agg functions except count, like sum/min/max/avg...?

And we also need iterators resp. cursors for SPARQL results.
Are there any publications about query processing/optimizing in Mulgara?

Does anybody now about clustered Kowari/Mulgara application scenarios. 
We are working on a distributed query processor for SPARQL. Any pointers 
are appreciated.

Thanks again and best regards,
Andy


Paul Gearon schrieb:
> Hi Andy,
> 
> While not network or hierarchical, the Mulgara database 
> (http://mulgara.org/) was written from the ground up to take RDF data 
> and nothing else.  The database used to be called Kowari, and before 
> that it was a commercial system called TKS (Tucana Knowledge Store).
> 
> To describe it briefly....
> 
> Mulgara does OK on 32 bit systems, but it scales really well on 64 bit 
> systems.  All joins are evaluated lazily, which removes memory limits, 
> and complex queries don't cause any problems.  We have been slow getting 
> SPARQL supported (though some simple SPARQL queries already work) but we 
> support our own query language that predates SPARQL by several years and 
> offers several more features.  These features include extra join 
> operations, and insert/delete commands.  There is also an SPI for 
> plugging any kind of external data source in as a supplemental back-end.
> 
> We went through a quiet period after the company that started the 
> original project shut down, but the project has been actively moving 
> ahead for the past year, with a few new features and a lot of bug fixes.
> 
> 
> Regards,
> Paul Gearon
> 
> On Apr 25, 2007, at 4:30 AM, Andreas Langegger wrote:
> 
>>
>> -----BEGIN PGP SIGNED MESSAGE-----
>> Hash: SHA1
>>
>> Dear group members,
>>
>> most of currently available RDF triple store implementations (Jena,
>> 3store, sesame, etc.) use relational database systems as the underlying
>> information system. Well, I guess this is because they are well known,
>> have proven to scale and perform, you get existing features for data
>> management, indexing, transactions, etc. and can use existing tools.
>>
>> However, are there any resources (papers/projects/links/...) where RDF
>> storage in network/hierarchical databases is investigated? Would there
>> be benefits?
>>
>> The bottlenecks with RDBMS are either huge joins over the same table =>
>> memory problem for large datasets or complex query processing which at
>> the moment has to be done in the higher app level. There should be
>> better support inside the DBMS in future like special join operations.
>>
>> If sb has pointers or comments, I'd appreciate!
>>
>> best regards,
>> Andy
>>
>> - --
>> - ----------------------------------------------------------------------
>> Dipl.-Ing.(FH) Andreas Langegger
>> Institute for Applied Knowledge Processing
>> Johannes Kepler University Linz
>> A-4040 Linz, Altenberger Straße 69
>>> http://www.faw.at
>>> http://www.langegger.at
>> -----BEGIN PGP SIGNATURE-----
>> Version: GnuPG v1.4.2.2 (GNU/Linux)
>> Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org
>>
>> iD8DBQFGLx+oKk9SuaNc5+IRAhuRAKDVaCd+Gur+rWZlRH6ZC2HbePBVCgCg2iJ5
>> eGHDozDvl7EauBy06TFnsZ8=
>> =H1Ia
>> -----END PGP SIGNATURE-----
>>
Received on Thursday, 26 April 2007 21:07:10 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 21:45:15 GMT