|
|
|
|
| Index | Recent Threads | Unanswered Threads | Who's Online | Help |
DataNucleus Forum
DataNucleus News
Articles
Thread: JPOX, Hibernate, and JPA Comparison Article
|
| No member browsing this thread |
|
Thread Status: Active Total posts in this thread: 6
|
|
| Author |
|
o_swas
Member Joined: Aug 8, 2007 Post Count: 55 Status: Offline |
Hello, First of all, congratulations on the JPOX 1.2.0 release!!!! It's a very impressive release. I was researching persistence mechanisms the other day and came across an article in an IEEE trade magazine that compared JPOX, Hibernate, and JPA (apparently using TopLink as the JPA persistence provider). I was surprised by the results. I know it's subjective, but they rated JPOX as being harder to learn than the other two -- a measure with which I disagree. The strangest result, though, was the performance results, shown in Figure 7. It shows JPOX as being significantly slower than either JPA or Hibernate. Judging by the Pole Position performance metrics I find these results hard to believe, especially by the margins shown in the article. Any idea how they could have come up with such strange results? Were there any performance enhancements between JPOX 1.1 and 1.2? I only ask because the article is dated February 2007, so maybe they were using JPOX 1.1 (the article doesn't state the version used). Also, I think JDO/JPOX is at least as easy to use as JPA or Hibernate. I don't know why they think it's harder. Maybe they were using a JDO version and using queries that used import statements? Just looking for some insight into how they could have arrived at such strange conclusions.... I haven't performance benchmarked JPOX myself but I haven't seen any problems with it myself, and I think it's as easy to use as other solutions. Again, congratulations on the 1.2.0 release and keep up the good work!! -Ryan |
||
|
|
andy
Expert UK Joined: Mar 13, 2004 Post Count: 5686 Status: Offline |
Just looking for some insight into how they could have arrived at such strange conclusions.... I haven't performance benchmarked JPOX myself but I haven't seen any problems with it myself, and I think it's as easy to use as other solutions. Insights ? read this one. The "highly qualified persistence expert" who performed this comparison didn't even bother turning on database connection pooling on JPOX hence giving it 60% performance deficit ... and even with that it still came out comparable to Hibernate. Why did this "expert" do this ? well because Hibernate had it turned on out of the box and JPOX didn't - clearly not up to reading docs, or asking on the Forum of the software he was using. Yes 1.2 is faster than 1.1. Do you have an email address for this magazine by any chance, so we can address our concerns over their capability at performing comparisons ? ---------------------------------------- -Andy   |
||
|
|
o_swas
Member Joined: Aug 8, 2007 Post Count: 55 Status: Offline |
Thanks Andy. I just saw that other article by an "expert" who apparently didn't bother with connection pooling for JPOX. He should be embarrassed. Do you have an email address for this magazine by any chance, so we can address our concerns over their capability at performing comparisons ? Unfortunately I don't have an email address. The best that I can do (and this is how I obtained the article myself) is provide a link: http://ieeexplore.ieee.org/Xplore/login.jsp?u...7568.pdf?arnumber=4297568 -Ryan |
||
|
|
o_swas
Member Joined: Aug 8, 2007 Post Count: 55 Status: Offline |
The "highly qualified persistence expert" who performed this comparison didn't even bother turning on database connection pooling on JPOX hence giving it 60% performance deficit Aha, I wonder if that's not the problem in this article too. The source code is not available (that I could find) but they do show snippets of the Spring configuration for each one. For JPA and Hibernate a DataSource is specified, but there isn't one in the JPOX configuration. So I wouldn't doubt if connection pooling isn't the problem. In fact, I don't know how they're getting any connection information, considering I don't see it in the Spring configuration. |
||
|
|
andy
Expert UK Joined: Mar 13, 2004 Post Count: 5686 Status: Offline |
Having now seen the report I posted an official reply on JPOX Performance Page, and I also added the reply to the DevX article. As per our promise on that page, if there is any comparison of JPOX with other persistence solutions we are prepared to assist whoever is making the study in judging best JPOX performance. For those groups that don't take us up on this offer, they could at least follow the performance and best practice docs on the JPOX website. Failure to do so is judged as unprofessional and we will post a reply highlighting any inadequacies in your comparisons ---------------------------------------- -Andy ---------------------------------------- [Edit 1 times, last edit by andy at Mar 3, 2008 9:00:15 PM] |
||
|
|
menno
Newbie Joined: Mar 4, 2008 Post Count: 2 Status: Offline |
Maybe you should change the default to turn database connection pooling on, to avoid these comparison problems. |
||
|
|
|
|
Current timezone is GMT May 17, 2012 11:42:33 AM |
