Thursday 14 November 2013

Why I think GlassFish isn't dead

At this week's Perth Insync conference I attended a great session comparing Oracle Glassfish Server (OGS) with Web Logic Server (WLS), only to find out that Oracle will discontinue future commercial support of the OGS product as of 4th November 2013.

This instigated a great flurry of discussion as one of our current clients is on the way to deploying a server dedicated to APEX using one of the Glassfish products.


Oracle recommends WLS as the web tier of choice for JDeveloper ADF, and the only choice for legacy Forms applications currently running on Apache that want to keep up with upgrading OS platforms.
OGS is typically a cheaper, lightweight alternative that can also support ADF - but is a more obvious choice for APEX deployments, in conjunction with the APEX Listener.

My head is rarely in the application server space, and I tend to not get too caught up with licencing discussions - no doubt others I refer to in this post know more about the topic than I, but I'm more than happy to provide an opinion from my perspective.

I try to catch up on my RSS feeds before the conference to keep an eye out for major announcements, this one obviously got through.
My first impression is that GlassFish has returned to what it was prior to the Oracle-Sun acquisition, so I found this made reading people's passionate opinions on the topic even more interesting.

The announcements

After finding this news story from ZDNet, this first post from familiar sources I encountered on the topic was from Kris Rice regarding APEX Listener support for application servers.

He confirmed the APEX Listener was still supported for WLS and Glassfish (3+), and in future support will be available for Apache Tomcat - thanks to public demand.

He also linked to the Aquarium's official annoucement regarding the Java EE and GlassFish Server Roadmap Update.

The discussion

I felt I needed more input so I started poking around to find a very vivid discussion on the topic, from Markus Eisele suggesting OGS is essentially dead and "thanks for all the fish".

I continued to read articles and they did seem to lean to one direction (anti-Oracle). Some looked back at Oracle's history with with open source projects, considering MySQL's fork into MariaDB.
The general concern is that since MySQL wouldn't be a revenue earner for Oracle, it would go neglected - stating that Oracle is even removing features.

To me this sounded like a slanted discussion without asking the other side. I'm not familiar with how MySQL works, but I'm sure Oracle limits the amount of hard parsing of SQL statements similar to this feature, just in a different component of the architecture.

I remember listening to a podcast describing what's new in MySQL 5.6 thinking that Oracle is vastly improving the MySQL architecture with solutions to match how the Oracle DB solves typical database issues. I didn't get the impression of neglect after reading these new features.

A number of people seem cynical of Larry Ellison's commitment to open source projects, but the information out there doesn't seem to reflect that.

Oracle's opinion

Then I started finding other contributions from Oracle's direction that started to provide some balance - and facts over fancy. Bruno Borges offered 6 Facts about GlassFish Announcement.

Someone else (I can't find the reference) noted the number of free/open source projects Oracle still delivers  - some of them rather high profile, so I think GlassFish shares some great company. To note just a few:
  • VirtualBox
  • SQL Developer
  • Oracle Express Edition
  • Hudson CI
  • Various Linux projects
Many of them supporting arms for Oracle's revenue engines - as per what the Aquarium originally stated:
The primary role of GlassFish Server Open Source Edition has been, and continues to be, driving adoption of the latest release of the Java Platform, Enterprise Edition
Oracle can't resource every project, so others such as OpenOffice have been handed over to the Apache Software Foundation.


At the end of the day we can always refer to Oracle support document. Premier support ends for OGS 3.1.x in March 2016, extended support until 2019 - but sustaining support is indefinite.

GlassFish continues to feed & support important components of Oracle's infrastructure and I would have confidence it will continue to thrive in the forseeable future.

The following points help me come to this conclusion
  1. APEX Listener continues to support OGS, in addition to WLS (and maybe Tomcat) - Oracle is pandering to customers needs, they don't need to 
  2. OGS and WLS share code and are closely coupled with the Oracle JDK - oracle is committed to Java EE
  3. OGS fills a niche and already shares a decent portion of the market - it has critical mass
I welcome further discussion on this and welcome any corrections to my statements.

No comments: