RE: speed question

2003-03-12 Thread Mahler Thomas
Hi Don,

OJB does nothing special if you are running it in a managed environment.
So IMO the overhead you see must be J2EE overhead.

I expect that you will see a drastic improvement if you just use a
(stateless) SessionBean and not BMP entity beans.
EntityBeans require *a lot* of management overhead.
Proper usage of transaction demarcation with entity beans also has an
enourmous impact!
So in OJB.properties you should switch to container managed connections!

cheers,
Thomas

 -Original Message-
 From: Don Walling [mailto:[EMAIL PROTECTED]
 Sent: Tuesday, March 11, 2003 11:58 PM
 To: OJB Users List
 Subject: Re: speed question
 
 
 Its the average time. The first time takes close to 2 seconds.
 I'm not so worried about the absolute value of the time. Its the 
 relative difference when we put the code inside an EJB that is 
 bothersome. Just wondering if there is something we're 
 missing regarding 
 configuration, or if we can reasonably expect the timings to 
 be similar.
 Thanks for your help,
 Don
 ---
 
 [EMAIL PROTECTED] wrote:
 
 Is 450ms the average time it takes or is it the first time 
 the ejb container 
 takes a request?
 
 Jeff
 
 
 On Tuesday 11 March 2003 09:44 am, you wrote:
   
 
 We're experiencing a speed difference we can't explain. 
 We've tried the
 following code in two circumstances:
 
 Date start = new Date();
 PersistenceManager manager = factory.getPersistenceManager();
 PersistenceBrokerFactory.defaultPersistenceBroker().clearCache();
 manager.currentTransaction().begin();
 String filter = registrationNumber == 2;
 Query query = manager.newQuery(AircraftValue.class,filter);
 Collection all = (Collection)query.execute();
 Interator result = all.iterator();
 Date end = new Date();
 
 When packaged in a .war and invoked by a .jsp, the code executes in
 about 15 ms. When packaged in an ejbLoad() method on a BMP, it takes
 about 450 ms. The OJB.properties and registry.xml files are 
 identical in
 both cases (copied from OJB 0.9.9), and we're using a 
 datasource to SQL
 Server. Is this speed difference resulting from being 
 executed inside a
 BMP transaction? Or is something else going on?
 Thanks in advance,
 Don
 ---
 
 
 
 -
 To unsubscribe, e-mail: [EMAIL PROTECTED]
 For additional commands, e-mail: [EMAIL PROTECTED]
 
 
 
 -
 To unsubscribe, e-mail: [EMAIL PROTECTED]
 For additional commands, e-mail: [EMAIL PROTECTED]
 
   
 
 
 -- 
 Your favorite stores, helpful shopping tools and great gift ideas. 
 Experience the convenience of buying online with [EMAIL PROTECTED] 
 http://shopnow.netscape.com/
 
 


-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]



speed question

2003-03-11 Thread Don Walling
We're experiencing a speed difference we can't explain. We've tried the 
following code in two circumstances:

Date start = new Date();
PersistenceManager manager = factory.getPersistenceManager();
PersistenceBrokerFactory.defaultPersistenceBroker().clearCache();
manager.currentTransaction().begin();
String filter = registrationNumber == 2;
Query query = manager.newQuery(AircraftValue.class,filter);
Collection all = (Collection)query.execute();
Interator result = all.iterator();
Date end = new Date();
When packaged in a .war and invoked by a .jsp, the code executes in 
about 15 ms. When packaged in an ejbLoad() method on a BMP, it takes 
about 450 ms. The OJB.properties and registry.xml files are identical in 
both cases (copied from OJB 0.9.9), and we're using a datasource to SQL 
Server. Is this speed difference resulting from being executed inside a 
BMP transaction? Or is something else going on?
Thanks in advance,
Don
---

-
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]