The standard well-established answer to this kind of question is that no benchmark can provide a reliable basis for predicting the behaviour of any particular application.

It's easy to benchmark the simple insertion of facts (= events) where there is no or just a single rule; this will give you an upper threshold for #/sec. But then it depends on the number and complexity of your rules; the (average) number of activations that have to be created; how long firings take; what has to be done in the WM; etc., etc.

Event processing is not different from plain fact processing - it's the very same engine. Temporal operators are bound to be a little slower than plain comparison operators.

-W

2011/6/28 Weiss, Wolfgang <Wolfgang.Weiss@joanneum.at>

Hi all!

 

Can anybody point me to a performance evaluation of Drools? I found some papers [1], [2] which focus on the rules part but I’m particularly interested in the event processing part of Drools. I want to know how many events can be processed per second with which latency, something that is similar to the performance evaluation of Esper [3].

 

cheers,

Wolfgang

 

[1] Open Rule Bench: http://www2009.org/proceedings/pdf/p601.pdf

[2] Focuses on rules: http://www.biomedcentral.com/content/pdf/1472-6947-10-3.pdf

[3] Performance Evaluation of Esper: http://docs.codehaus.org/display/ESPER/Esper+performance

 

 

--

Wolfgang Weiss

DIGITAL – Institute of Information and Communication Technologies

 

JOANNEUM RESEARCH Forschungsgesellschaft mbH
Steyrergasse 17, 8010 Graz, AUSTRIA

phone:  +43-316-876-1209

general fax: +43-316-876-1191

web:    http://www.joanneum.at/digital
e-mail: wolfgang.weiss@joanneum.at

 


_______________________________________________
rules-users mailing list
rules-users@lists.jboss.org
https://lists.jboss.org/mailman/listinfo/rules-users