Thursday Jul 12, 2007

Fastest App Server in the West

We recently reported 2 SPECJappserver2004 results which are detailed in blogs here and here.  However, allow me to shed some light on these recent publications.  We published two results that highlight vastly different capabilities of the app server. The first was a speed record that shows GlassFish V2 is 10% faster than BEA on comparable hardware and the 2nd shows how fast GlassFish V1 runs with Postgres.  The first result was 883 SPECJAppServer2004 JOPS@Standard\* which was for the very latest, fastest version of GlassFish that we call V2 which is the enterprise edition that includes lightweight fail-over, clustering and administration features - you know those things you need to run the app server in an enterprise environment.  Now the interesting point is that this result was 10% faster than BEA running on the exact same hardware / DB setup!  This is a wonderful result for open source as it shows that an open source (and open community) can not only deliver a product but the fastest one on the market!

The 2nd test was done to show that Sun (and you) can create an open source stack that runs as well as an equivalent proprietary stack.  In this case we paired GlassFish V1 (now this is the original developer edition) with Postgres and delivered a result of 778 SPECJAppServer2004 JOPS@Standard\*\*.  This should answer the question of whether open source is good enough to run in production (yes) and should help everyone understand that Postgres rocks!  Why did we use V1 rather than V2 in this test? What were we hiding? Well, nothing, this was simply a matter of timing.  Because the product you test must be released and the PostGres folks wanted to announce immediately, they had to use GlassFish V1 as V2 will not be released until September. If you want to know how fast GlassFish V2 is with other open source data bases - just wait a bit.

Legal stuff: 

Sun Fire T2000 (1 chips, 8 cores) 883.66 SPECjAppServer2004
JOPS@Standard. SPEC, SPECjAppServer reg tm of Standard Performance Evaluation
Corporation. Results from as of 7/10/07.

\*\*Sun Fire X4200 (6 chips, 12 cores) 778.14 SPECjAppServer2004
JOPS@Standard. HP rx2660 (2 chips, 4 cores) 874.61 SPECjAppServer2004 JOPS@Standard.  SPEC, SPECjAppServer reg tm of Standard Performance Evaluation Corporation.  Results from as of 7/10/07

Wednesday Jun 27, 2007

4th and Final Winner of GlassFish Beta Contest

We have finally choosen our 4th winner - it took so long because the 1st few winners never responded to our messages.  Let this be a lesson to all you who enter contests!  Enter a valid email address and check your email every week or so as you never know when you might be a winner!
Wes from the USA is our 4th and final winner. He writes "I'm new to application servers, but the performance alone is worth
considering a migration from Apache! However, simple tasks in Apache, such as protecting a directory from public access, are reasonably difficult for new users of GlassFish. I especially like the administration interface look and design, but unfortunately, there isn't much to change in many sections unless you are intimately familiar with the innards of this server as many of the areas to tweak are cryptic and don't have a listing of available parameters to change and the available values. In this case, Webmins Apache module is much easier for novices to learn from. Server restart should be able to be done completely from the web GUI...not just a shutdown, go to CLI to

Congratulations to our winners and thank you to all who entered the contest!

Tuesday Apr 03, 2007

Update to GlassFish Beta Contest - Blog About it!

Someone asked "if they blogged about the GlassFish V2 Beta would that count as a contest entry"? Well yes, but you still HAVE to fill out the official form to enter. For the entry to be valid either enter a permalink to the blog entry (so we can find it) or copy and paste the blog contents into the form. Contest details are here

Sunday Apr 01, 2007

And we have the 1st winner in the GlassFish V2 Beta contest!

The GlassFish V2 Beta contest has produced its first winner, but before I talk about them, let me cover some of the interesting feedback we have been given. We had roughly 50 entries to the contest from folks all over the world - Russian Federation, Germany, Thailand and the USA to name a few - that is pretty exciting! In general, the comments are mostly very positive and some are just...interesting. Remember the engineers get all your comments and use them to A)cheer themselves up B)gripe about why you don't love the code they have sweated over for years and C)improve GlassFish!

On the Interface
"The UI looks much better than it did in previous versions" "At first glance, I really like the interface. It looks nice, clean and simple. I like a simple interface". "It is great to see a JSR admin seems more robust that the previous version" and, my favorite: "The administration console which has more configuration options than the JBoss equivalent without being overy complex like the Weblogic equivalent"

"GlassFish is very robust and stable", "Stable product, takes little resource to to work", "Installs like a charm,...stable and reliable", "My Wicket Web app ran the first time".

"Need more Japanese informations about GlassFish to increase users in Japan..same for most Asian countries)" ( I think this is true in general so please join the community and help translate documents - Ken)

Interesting Comments
"(GlassFish) can be deployed on obsolete hardware" - (Yes, and all of us are forced to use obsolete hardware at sometime! - Ken) "Really it is good to use and it was Faster than the Microsoft Product" - (I am really happy to hear this. I haven't used the MSFT Java EE 5 implementation, but look forward to the opportunity ;) - Ken)

Mac Users
"On my MacBook, the application server seems to start faster than V1". "The installation process worked well in console mode. It, however, did not work when using X11 on MacOS 10.4 for the display"

So the 1st winner (remember we have 3 more to go) is Tim Davidson out of the UK. To read his full comment just click here.

Wednesday Nov 22, 2006

Sun's App Server is #2 in latest Evan's Survey AHEAD of IBM

There is alot of press about the IBM interpretation of the latest Evans data on application server use by developers. IBM spins the data as they are growing faster than JBoss. This may be true and when you are growing from a small base, its always easier to "outgrow" someone who has a large base.

What I find interesting is the REAL data from the Evans Fall OSS/ Linux Developer Survey - the survey that asks Linux developers what they use, you find that that the application servers they develop on are: #1 is JBoss, #2 is Sun and #3 is IBM. (In this case it was NOT IBM's community edition but their commercial version.) This data is very revealing. The open source community has quite a few application servers to choose from these days and they are tending to use the supported, commercial based ones. Early on, developers used open source because it was open source, now, it seems, they are getting a bit more pragmatic and are choosing the more robust application servers. I think the Sun application server has a 4 things going for it 1) its Java EE 5 compatible, 2) its free to download and deploy (as is the GlassFish open source app server), 3) it is based on the robust commercial code that Sun has been supporting for years and 4) there is a transparent migration path up to the enterprise grade, high availabilty version.




« February 2017