On 10/9/11 5:56 PM, Mohamed Morsey wrote:
Dear all,

the AKSW [1] group is pleased to announce the development of a benchmark
called DBpedia SPARQL Benchmark (DBPSB) for evaluating and comparing the
performance of Virtuoso, Sesame, Jena-TDB, and BigOWLIM.

We have also written a paper about it. It indicates our experiment and
the results we obtained. It is available at
“http://iswc2011.semanticweb.org/fileadmin/iswc/Papers/Research_Paper/03/70310448.pdf”.

This page, “http://aksw.org/Projects/DBPSB”, indicates the results in
more detail.

Any feedback and/or suggestions are highly appreciated.

Mohamed,

Since Virtuoso exists in Single-Server and Cluster Editions, it's really important that you don't use "Virtuoso" generically in benchmark. Thus, please use "Virtuoso Single-Server" and "Virtuoso Cluster Edition" where appropriate.

Simple effects: the cluster edition offers full parallelization which even applies to a single machine hosting multiple Virtuoso instances working in parallel. Immediate effect is that DB load times are much faster. Also note that the live instance of DBpedia is based on a 4-node cluster hosted on a single machine.

I haven't delved into the results, yet, but the qualification immediately caught my attention when I saw the dataset load times.

Note: on the lod.openlinksw.com instance which has an 8-node cluster spread over 8 machines, we have been able to load the entire DBpedia dataset in under 20 mins., courtesy of 8 virtuoso instances working in parallel across 8 machine hosts.

--

Regards,

Kingsley Idehen 
President&  CEO
OpenLink Software
Web: http://www.openlinksw.com
Weblog: http://www.openlinksw.com/blog/~kidehen
Twitter/Identi.ca: kidehen






Attachment: smime.p7s
Description: S/MIME Cryptographic Signature

Reply via email to