W3C home > Mailing lists > Public > semantic-web@w3.org > September 2006

RE: Performance issues with OWL Reasoners

From: Kashyap, Vipul <VKASHYAP1@PARTNERS.ORG>
Date: Mon, 18 Sep 2006 14:24:52 -0400
Message-ID: <2BF18EC866AF0448816CDB62ADF6538104C16577@PHSXMB11.partners.org>
To: "Phillip Lord" <phillip.lord@newcastle.ac.uk>
Cc: <public-semweb-lifesci@w3.org>, <semantic-web@w3.org>


> The hypothesis is only going to be true IF the mapping is scalable.
> Otherwise, it doesn't work.

[VK] This is a valid point! The mapping has to be scalable. In the RDBMS
context, what this means is that this should be mappable into a specialized
schema so that the following "scalability features" of RDBMS become easily
available:
- The ability to specify and fine tune indices on specific columns (this is
possible as the schema is now specialized)
- The ability to leverage the query optimization engine by so that optimized SQL
query plans for implementing DL (or atleast OWL Lite) operations. 
- Of course the above would require expressing OWL Lite operations as
SQL/relational algebra queries (to the extent possible, as it is clear that the
relational model cannot fully accommodate OWL-DL).  This part is crucial as the
same operation can be expressed inefficiently  --- making the mapping unscalable
as suggested above.
- Caching

> There are two different things in the technologies you mentioned;
> relational to X mapping tools, and metaschema approaches. They are
> quite different. For the instance store, the relational database is
> really an implementation detail. It's basically a reasoner with
> somewhat limtied expressivity which is persistent and (hopefully)
> scalable.

[VK] Is this true only for TBox reasoning, or ABox and TBox reasoning? I had the
impression (possibly mistaken) that ABox reasoning takes advantage of the
relational backend. At least this is implied by the following snipper from an
earlier e-mail:

"PL> Flip side, is that if you put everything into the TBox, then you get"
"PL> nothing from the relational backend of the instancestore."

> Because it's using a metaschema approach, you can't do things like use
> RDBMS security, for example (beyond yes/no).

[VK] I guess I need to get out of my laziness and read the paper, but how
different is the metaschema approach from the X to relational mapping approach? 

Even if it is a very different approach, is it able to leverage the "scalability
features" of an RDBMS enumerated above? I would be interested in your responses.

Cheers,

---Vipul
Received on Monday, 18 September 2006 18:25:07 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 1 March 2016 07:41:53 UTC