In early days of my career, Microsystems was the company we all looked towards. No other company innovated as much in hardware and then in software as Microsystems did. In fact, Apache Tomcat started as a project at Sun. I would have guessed in the early 90s that Sun would buy – oh well, how times change.

Oracle has a daunting task ahead of integrating some amazing technologies they acquired from Sun Microsystems. Several important and critical technology pieces such as MySQL, NetBeans, and the whole community need careful attention. It is only natural then that some other products might not get an equal amount of TLC from Oracle executives.

Oracle unveiled their plans for several of Sun’s products, and it’s no surprise that did not get as much attention as some of Sun’s core products. From Oracle’s FAQs, they state:

Oracle plans to continue evolving GlassFish Enterprise Server, delivering it as the open source reference implementation (RI) of the Java Enterprise Edition (Java EE) specifications, and actively supporting the large GlassFish community. Additionally, Oracle plans to invest in aligning common infrastructure components and innovations from Oracle WebLogic Server and GlassFish Enterprise Server to benefit both Oracle WebLogic Server and GlassFish Enterprise Server customers.

The highlighted statement should raise alarm bells in the minds of GlassFish users. The reason several users chose GlassFish over WebLogic is because GlassFish was a less complex application server than WebLogic, which has become increasingly complex and even cumbersome to use. Aligning GlassFish with WebLogic? That sounds like taking a complex component and stuffing it into GlassFish, step one towards moving GlassFish customers onto WebLogic. While I like to believe that some of the GlassFish features would make it into WebLogic, I am doubtful. Having worked in big corporations for quite sometime, I do know this – the architectural and product decisions are heavily biased towards who is running the division, and as of now, no indications that GlassFish executives will be given charge of Oracle application server business. This makes business sense for Oracle, as they make more money selling complex solutions and convincing customers to buy expensive consulting and services on top of their complex solutions. And guess which division GlassFish is going to? To the same team that sells the ultra complicated, heavy weight Fusion middleware.

GlassFish users can expect to hear the following from Oracle sales guys in the near future:

“Well, you really should consider moving to WebLogic application server, as Fusion middleware is the future, not GlassFish”

“Its okay for you to use GlassFish on departmental servers, but you really should move away from putting any serious applications on GlassFish.”

So what are the real alternatives? If your primary goal is to be efficient, get outstanding ROI, and keep your runtime infrastructure, you should evaluate moving to Apache Tomcat immediately. Given that most web applications do not actually require full Java EE application server, you would also be cutting the weight of your runtime. You will find that this gives you  at a minimum, leverage to negotiate a better deal with Oracle. Here is a whitepaper that discusses migration from WebLogic to Tomcat, but the basic concepts apply to GlassFish users as well.

Innovation in Apache Tomcat continues with the latest release of 6.0.24 and upcoming 7.0 release. There is a large and growing community around Tomcat, wider adoption, and several companies that are standing behind Tomcat and provide worldclass support. Apache Tomcat means you have a choice of vendors, and thats a good thing.  Given the uncertainty around Oracle’s plans for GlassFish, there’s never been a better time to move to Tomcat.

No related posts.


15 Responses to “GlassFish Users: Where to go next?”

Jacek February 2nd, 2010, 12:35 pm

Wrong.

You should evaluate a move to Jetty. More lightweight, perfect Maven integration. Use Spring instead of Java EE and you’re free of Oracle worries.

Eduardo Pelegri-Llopart February 2nd, 2010, 5:12 pm

Hi Sateesh.

We (old GlassFish and old WebLogic Server are sibling groups reporting to the same SVP) folks are working on the combined road map for WLS+GlassFish. Today is CiC + 4d; give us some time to deliver that roadmap and then you can use that to evaluate the situation for your readers.

– eduard/o

Sateesh Narahari February 3rd, 2010, 1:30 pm

Eduardo,

Thank you for your comment. I am painfully aware of how long it takes to make decisions in large companies and especially road maps take few months and quite a few meetings to arrive at.

Would be watching announcements and if there is an interesting item, I will discuss it on the blog.

Best Regards,
Sateesh

Reza Rahman February 3rd, 2010, 5:31 pm

Wow! This has to be one of the most blatantly self-serving sales pitches I have ever seen – considering Oracle considers GlassFish extremely important and MuleSoft has been pushing Tomcat for a while now…

Ryan de Laplante February 3rd, 2010, 5:50 pm

> If your primary goal is to be efficient, get outstanding ROI, and keep your runtime infrastructure,

yes, Yes, YES!!! You are going to recommend GlassFish V3 and Java EE 6 right?…

> Given that most web applications do not actually require full Java EE application server

I have to disagree. Developers want JPA 2.0, JTA, JAX-WS, JAX-RS, Java Mail, Beans Validation, JAXB, JAXP, Servlets 3.0, JSF 2.0, CDI, EJB 3.1, etc. Ok, there will always be some masochists that choose not to use JSF 2.0/EJB 3.1/CDI, but prefer to integrate a half dozen other frameworks and libraries with XML. The rest of the Java EE portfolio is common and useful to most developers.

Developers and operations also want full featured easy to use management tools, like what comes free in GlassFish. The lightweight development experience of GlassFish V3 and Java EE 6 really make the Tomcat + Spring + dozen other frameworks and libraries integrated with XML look so 2007. Java EE 6 and GlassFish V3 is more efficient and gives outstanding ROI as you described. That is the reason why GlassFish users (paying and non-paying) choose GlassFish as their strategic application server instead of Tomcat.

I also watched the Oracle videos and got a very different impression about GlassFish’s bright future. If they don’t prevent the GlassFish engineering team from re-integrating clustering & high availability into V3.1, then I am happy to keep supporting GlassFish.

Dimitris Menounos February 4th, 2010, 12:09 am

@Ryan

Tomcat is great because it is *focused* at doing *one* job well. That is serving your application code over HTTP, and this is the way most prominent servers work. All the other technologies you mentioned are not in the core business of an HTTP java server. They are extras that integrated together play the role of an application framework. Spring does that integration.

Karsten Silz February 4th, 2010, 6:29 am

To put some necessary distance between the free Glassfish and the “$10,000 per CPU” Weblogic, Glassfish “will be geared for departmental use” (http://www.sfgate.com/cgi-bin/article.cgi?f=/g/a/2010/01/27/urnidgns852573C400693880002576B900002824.DTL), like Apache Geronima and Websphere at IBM. So I expect high-end, enterprise features to either wither away from Glassfish or not being added at all, making Glassfish less attractive for some developers.

Ryan de Laplante February 4th, 2010, 9:26 am

@Dimitris: That is great, and Tomcat is built-into GlassFish as the servlet container. The servlet container is a small but important part of the big picture. With a Java EE 6 application server, I can focus on developing my application instead of having to select and integrate a dozen frameworks and libraries to make up for what Tomcat is missing. I don’t need to argue with every other developer because we each have our own favorite frameworks and libraries. Plus, I get sophisticated management and monitoring capabilities for free. There isn’t any reason for GlassFish users to migrate to Tomcat, but there are reasons for Tomcat users to migrate to GlassFish. I remember when Oracle bought WebLogic there was a similar article on theserverside.com warning everyone that WebLogic will be neglected, and recommending they switch to something else (I forget to which vendor). Look at WebLogic now.

@Karsten: It is too soon to make those kinds of assumptions or to write a blog entry about GlassFish’s alleged impending doom. I will quote Eduardo:

“We (old GlassFish and old WebLogic Server are sibling groups reporting to the same SVP) folks are working on the combined road map for WLS+GlassFish. Today is CiC + 4d; give us some time to deliver that roadmap and then you can use that to evaluate the situation for your readers.”

I will wait until then to decide whether or not to continue working with GlassFish. If I migrate, it will *not* be to Tomcat.

Geir Hedemark February 4th, 2010, 11:05 am

@Ryan:

>Developers and operations also want full featured easy to use management tools, like what comes free in GlassFish

Actually, the “operations” part of your statement is incorrect. Glassfish tries to encompass too much hosting-related functionality, and therefore is fairly costly to integrate into an automated deployment and change management regime. When it comes to operations, Jetty, Resin or Tomcat has a much lower cost both up front and maintenance-wise.

You won’t notice this in operations which use pointy-clicky interfaces to manage servers, though. I guess there still are some of them out there.

Alexis MP February 4th, 2010, 12:16 pm

@Geir :
GlassFish v3 is much closer to Tomcat and the likes than you seem to think (far from pointy-clicky in fact).
With admin CLI, RESTful admin API, standard “java -jar glassfish.jar foo.war”, a maven 2 plugin and overall fast startup time I think you need to try it!

Geir Hedemark February 8th, 2010, 10:14 am

@Alexis:
>GlassFish v3 is much closer to Tomcat and the likes than you seem to think (far from pointy-clicky in fact).

I never said it was.

I want an appserver with an XML configuration file that is easy enough to easily teach generic hosting people and have them assemble using standard hosting tools.

One-off CLIs or REST APIs are going to cost me more money to have integrated and maintained.

Alexis MP February 8th, 2010, 2:14 pm

GlassFish does have a *single* XML config and that’s precisely what the GlassFish built-in CLI and RESTful admin API modify and what seemed to have pleased a number of hosting companies recently. Granted this came with v3.

Erik Weibust February 20th, 2010, 8:19 pm

Sateesh,

The link to the whitepaper appears to be broken… Can you fix or send me the whitepaper?

Thanks…
Erik

Community February 22nd, 2010, 4:07 pm

Thanks, Erik. The link to the whitepaper has been fixed.

Nazrul March 2nd, 2011, 7:23 pm

GlassFish 3.1 has been released with clustering, centralized administration and high availability. Refer to http://blogs.sun.com/theaquarium/entry/glassfish_3_1_is_here


Nazrul

Leave a Comment