Nazrul's Weblog

GlassFish V2, the Fastest Application Server on the Planet

Thursday Jul 12, 2007

Based on the recently published benchmark, GlassFish V2 now has the best (883.66 JOPS@Standard) SPEC jAppServer 2004 score on Sun Fire T2000. Read this blog for details. On the same comparable hardware, GlassFish V2 is #1! Here are the relevant SPEC jAppServer 2004 scores...

System Title JOPS Disclosures
IBM WebSphere Application Server 6.1 with DB2 8.2 on Sun Fire T2000 using Solaris 10 616.22 Html Link
Oracle Application Server 10g Release 10.1.3.2 - Java Edition on Sun Fire T2000 733.22 Html Link
BEA WebLogic Server 9.2 on Sun SPARC Enterprise T2000 801.70 Html Link
GlassFish V2 (Sun Java Application Server 9.1) on Sun Fire T2000 883.66 Html Link

Note: I did not see any submissions from JBoss. Refer to the official SPEC jAppServer2004 Results page for more details.

Disclaimers: SPEC and the benchmark name SPECjAppServer 2004 are registered trademarks of the Standard Performance Evaluation Corporation. Competitive benchmark results stated above reflect results published on www.spec.org as of 07/10/06. For the latest SPECjAppServer 2004 benchmark results, visit http://www.spec.org/.

[9] Comments
Like this post? del.icio.us | furl | slashdot | technorati | digg
Posted at 07:23PM Jul 12, 2007 by nazrul in GlassFish
Tags:
Comments:

WebSphere       616.22       J2SE1.5.0_06 (IBM build)
Oracle AS          733.22        J2SE1.5.0_06
WebLogic          801.70        J2SE1.5.0_06 
GlassFish V2   883.66        J2SE6u2
I think the reason is "JDK6 out perform JDK5", not the app server itself.

Posted by 192.18.43.225 on July 12, 2007 at 08:27 PM PDT #

Latest GlassFish V2 result is almost 70% higher than previous score of 521.42 JOPS@Standard on a Sun Fire T2000 (was running at 1.2ghz). GF V2 is not 70% faster than GF V1, but it is quite substantially faster. I am very pleased to see the highest ever score on the Sun Fire T2000 with GlassFish V2.

Posted by Nazrul on July 12, 2007 at 11:05 PM PDT #

Also, see this thread on TSS about this topic and the usage of JDK 6.

Here is Scot's comment: If you want JDK6, you have to use glassfish
Henrik makes a good point that JDK 6 contributes to our performance (though probably not as much as he assumes). Which is yet another reason why Glassfish is the superior appserver: we've supported JDK 6 for more than 6 months. Weblogic doesn't support it today, and as far as I've heard they have no plans on supporting it for months to come. So it's not apples-to-apples, but then again, neither is the marketplace.

I am glad to see that Java SE (JDK) team at Sun also worked on performace. Great!

Posted by Nazrul on July 17, 2007 at 07:00 PM PDT #

I do a considerable amount of work for an enterprise software company that uses BEA Weblogic. We have considered switching application servers serveral times but BEA we have been told is the fastest on 2003 Server. Do you have any documentation on Application servers that run on 2003 Server with a MS SQL DB?

Jimmy
www.tracetechnologies.us

Posted by J Dixon on November 06, 2007 at 04:58 PM PST #

Hi Jimmy,

Did you mean <a href="http://www.spec.org/benchmarks.html#java">jAppServer2002</a>? It is retired now.

You are welcome to try GlassFish V2 <a href="https://glassfish.dev.java.net/downloads/v2-b58g.html">(download) </a> with your application and see how it does. That is probably the best way to find out. Please feel free to contact GlassFish community at <a href="mailto:dev@glassfish.dev.java.net">dev@glassfish.dev.java.net</a> if you have any questions.

Nazrul

Posted by Nazrul on November 22, 2007 at 11:53 AM PST #

While I don't actually doubt that GlassfishV2 is fast, the testing may be flawed. I see in there that a Sun SPARC Enterprise T2000 was used while the others were Sun Fire T2000. Then there's the Java version like the first poster mentions. The fact that the hardware is roughly the same is only a part of the equation. The operating system, the amount of memory, and the pipe between the database and the application server all make huge differences.

I clicked through to a couple of the linked results pages and you can see that there are a ton of differences between the deployments. The tests involved different database technologies and different JDBC drivers. So the tests are less comparative than they need to be.

I think the only thing really proven is that Glassfish can give all the other application servers a run for their money in real world deployments. To me that statement means I keep an eye on Glassfish and wait for some truly solid benchmarks in homogeneous system setups, especially Windows server setups with SQL Server databases.

Posted by Chris Arthur on March 30, 2009 at 05:23 AM PDT #

Hi Chris:

Thanks for expressing your interest in GlassFish.

Both Sun SPARC Enterprise T2000 and Sun Fire T2000 use the same UltraSPARC T1 processor. Refer to: http://www.sun.com/servers/index.jsp?tab=2

SPEC jAppServer 2004 is a industry standard benchmark that all vendors use to publish the results. AFAIK, this is a leading benchmark in this area.

--Nazrul

Posted by Nazrul on April 15, 2009 at 12:58 PM PDT #

TrustLeap G-WAN seems to be even faster than GlassFish.

With ANSI C scripts (executed on-the-fly) that are 5x faster than IIS 7.0 ASP.Net C#, G-WAN (in user-space) manages to be faster than IIS 7.0 (in the Windows kernel) for both static and dynamic contents.

The Sun web server claims to be 8x faster than Apache, but TrustLeap G-WAN is up to 38x faster than Apache.

And a Solaris version is in the works!!

Posted by Pierre on September 30, 2009 at 03:06 AM PDT #

Hi Pierre,

Is TrustLeap G-WAN a Web server? I could not tell from the web site. If there are any published standard benchmarks that shows TrustLeap G-WAN is faster than GlassFish, please send me a pointer.

Thanks.

Nazrul

Posted by Akm Islam on October 22, 2009 at 01:38 PM PDT #

Post a Comment:
  • HTML Syntax: NOT allowed