Search Mailing List Archives


Limit search to: Subject & Body Subject Author
Sort by: Reverse Sort
Limit to: All This Week Last Week This Month Last Month
Select Date Range     through    

[protege-owl] Query: Upper-bound of Protege 3.4 individual capacity

william fitzgerald wfitzgerald at tssg.org
Wed Oct 29 07:59:38 PDT 2008


Dear All,

What is the upper-bound on the amount of individuals in an OWL-DL 
ontology before Protege 3.4 crashes?

Personally, I have a latitude d510 running Ubuntu Hardy with 1GB RAM, 
latest *Full* version of Protege 3.4 build 513 and the java 1.6 VM.

I modified the heap space from 209715200 bytes to:

lax.nl.java.option.java.heap.size.initial 409715200
lax.nl.java.option.java.heap.size.max  809715200

Example:
I have used DataMaster to import the "Vulnerabilities" Table from the 
Open Source Vulnerabilities Database (OSVDB) stored in a MySQL database 
also running on my laptop.

It contains 48,0000 instances when imported into the ontology.

I cannot import using DataMaster the other tables without a crash. I 
estimate if the entire OSVDB database was to be imported into an 
ontology there would be around 100,000 instances. I also intend on using 
instances from other sources, so my ontology has the potential to be large.

The current security ontology I have now populated with just the OSVDB 
Vulnerabilities Table is very slow to load when opening it with Protege IDE.

I then tried to work with what knowledge (instances) I had and to 
reclassify various individuals of the Vulnerability class into more 
specialised classed (for example, BufferOverflowVul). To do this I 
decided to write some SWRL rules to look at each individual and its 
data-properties (DataMaster creates data-properties and doesn't 
currently include an object properties option) to make a decision as to 
where that individual can be re-classified.

However, SWRL-Jess runs out of memory when I execute a single SWRL rule 
or query. Presumably this is due to the fact that JESS uses its own 
memory space (I could be wrong!). And the fact that protege uses a lot 
of the limited memory already.


How do people deal with large knowledge bases? Is it a matter of 
installing more RAM?


Kind regards,
Will.


-- 
  ______________________________________________________________________
|                   William M. Fitzgerald (MSc, BSc)                   |
|______________________________________________________________________|
|Researcher                                  | PhD Student             |
|Telecommunications Software & Systems Group,| Computer Science Dept., |
|ArcLabs Research and Innovation Centre,     | University College Cork,|
|Waterford Institute of Technology,          | Cork.                   |
|WIT West Campus,                            |                         |
|Carriganore,                                |                         |
|Waterford.                                  |                         |
|Office Ph: +353 51 302937                   |                         |
|Mobile Ph: +353 87 9527083                  |                         |
|____________________________________________|_________________________|
|Web: www.williamfitzgerald.org                                        |
|     www.linkedin.com/in/williamfitzgerald                            |
|     www.ryze.com/go/wfitzgerald                                      |
|     www.tssg.org/people/wfitzgerald/                                 |
|______________________________________________________________________|



More information about the protege-owl mailing list