Sean Gallagher has written a piece called The Geronimo Effect which delves into his opinion that: When the Apache Software Foundation's Geronimo Java application server arrives, it could set off a chain of events that change the application server market forever. He discusses the various open source licenses, and how he feels Geronimo brings a new twist to J2EE in an open source world.

Read: The Geronimo Effect

That is decission Geronimo team made by choosing ASF license, and if that's ok for them, it's ok for me as well.
I don't see anything wrong if someone use Geronimo the same way Tomcat is used by lots of vendors.

Between Geronimo, JSF, Java5, EJB3 and some of the great tools that are emerging in the Java and J2EE space, I can't see it as being a net loss for the major J2EE vendors (IBM and BEA) because Geronimo will only help the market to expand, and in the long run they'll do well with their higher-value products (portals, integration, etc.)

Like most open source solutions, Geronimo will primarily hurt the smaller-market-share players (e.g. Sun, Oracle, Caucho, Pramati, etc.) because they will have to work that much harder to stand apart, even if their products are good (Caucho rocks!)

Sun already realized and accepted this, and changed their licensing approach accordingly .. I think in the long run, they're probably ahead of the curve on realizing how the software industry has changed. Java is doing very well, Solaris 10 is going to do very well, their x86 (AMD) servers are doing well, and the only question left is whether Sparc will ever be compelling again from a price/performance perspective.

Peace,

Cameron Purdy
Tangosol, Inc.
Coherence: Shared Memories for J2EE Clusters

Geronimo will primarily hurt the smaller-market-share players (e.g. Sun, Oracle, Caucho, Pramati, etc.) because they will have to work that much harder to stand apart, even if their products are good (Caucho rocks!)

I don't think Geronimo will have much impact on Oracle Oracle has put a big focus on J2ee, but much of that focus has been to centeralize *their* product developement and deployment. Oracle Forms, Oracle Reports, Oracle Discoverer, Oracle Portal , and Single Sing-on now all run as Java applications inside 10g AS. This has been been a big help to Oracle from a mantience, development, and integration point of view. It is also allowing them to consolidate their development tools. It looks like JDeveloper will end up becoming THE ide for all Oracle products.

I know that many -Java only- people who are not interested in running Oracle's 10gAS, but companies vested in existing Oracle products are likely to web enable their products with 10gAS. Our company upgraded to 10gAS to web enable our existing Oracle Reports. At that point it only made sense to deploy our java applications to OC4J, and now we use JDeveloper for all of our inhouse development.

Geronimo will primarily hurt the smaller-market-share players (e.g. Sun, Oracle, Caucho, Pramati, etc.)

I think the the biggest player to be hurt by this is JBoss. It's currently the premiere open source J2EE server and with the recent JBoss 4.0 J2EE 1.4 certification has just made it that more attractive to businesses looking for a low cost alternative to WebSphere or Weblogic.

But just like Apache webserver and Tomcat have become the defacto webserver and JSP/Servlet container, I think the same will happen to Geronimo over time.

JBoss realized that last year when their lawyers contacted Apache, http://geronimo.apache.org/news.html.

... I think the the biggest player to be hurt by this is JBoss. ...

I not sure that JBoss will be that hurt. JBoss has survived this long without generating revenue from license fees. Geronimo is not going to take away the services that make JBoss money.

... I think the the biggest player to be hurt by this is JBoss. ...

I not sure that JBoss will be that hurt. JBoss has survived this long without generating revenue from license fees. Geronimo is not going to take away the services that make JBoss money.

You are absolutely right. Geronimo is not going t otake away the services that make money for Jboss Inc. Except that they might shift alitle and sell services aroung GERONIMO to survive :-))

You are absolutely right. Geronimo is not going t otake away the services that make money for Jboss Inc. Except that they might shift alitle and sell services aroung GERONIMO to survive :-))

You are absolutely right. Geronimo is not going to take away the services that make money for Jboss Inc. Except that they might shift alittle and sell services around GERONIMO to survive :-))

Please excuse my spelling.

as a member of apache community, I don't get all the hype around geronimo. there's a ton of open source EJB containers out there. It's great that geronimo will push competitors, but there's no need to hype it. middleware was already becoming commodity anyways, this is just another step in the gradual evolution.

as a member of apache community, I don't get all the hype around geronimo. there's a ton of open source EJB containers out there. It's great that geronimo will push competitors, but there's no need to hype it. middleware was already becoming commodity anyways, this is just another step in the gradual evolution.

Agree totally. The thing that impress me the most is NOT the open source or free use of Geronimo; it's the fact that their Kernel is absolute ignorant of the J2EE spec. They throw in some other opensource jars (e.g. OpenEJB) and some configuration to "running" Geronimo server and suddenly they have a J2EE application server. That way users not only can choose what services they want to use (and more importantly, what J2EE service NOT to use ... did I hear entity beans!!) but also they can choose best of the breed vendors! Moreover, they are well positioned to respond to trend/spec changes more quickly and more reliabllly as all they need to do is search a vendor who has that change implemented.

-Anshuman

I do not see any commercial vendors suffering from the release of Geronimo release. Since it is ASF license, they as any other developer or group, can use Geronimo as a core for their application server platform. Give credit to Apache and than hack away at it. It reduces cost of maintenance and implementation for the big vendors, allowing them to spend more money on marketing and service-based add-ons around the core. Vendors like IBM and BEA would not have a slightest problem adopting and refactoring their code to compete with a decent but still low-end Geronimo alternative by integrating and growing it to their liking because they can, due to ASF licensing.

This nervous hype is just irritating, all that is going to happen IMO is that people will have an alternative to a commercial J2EE server implementations. But people already have such alternatives for servlet containers, database servers, messaging servers and operating systems and it does not seem to rule any of the commercial vendors out of the market. Even more, in many cases, it let's commercial vendors use such alternatives as low-end examples to make their systems look better in comparison.

I am glad to have such alternatives though...

Just my 2 cents...

Artem D. Yegorov
http://www.activexml.org

Will Geronimo be any good in a production environment or will it be just a Reference Implementation? In case the latter is true then what difference it would make if i use geronimo or jboss or etc during development. Can i use geronimo in a production environment where BEA & IBM rule?

To expect Geronimo to match Weblogic or Websphere feature by feature or even in robustness (especially weblogic) would be little too much. What is really worth looking forward to is whether geronimo can fit somewhere between jboss and Weblogic/Websphere in terms of production quality.

Nitin Chaumal

I think u r undermining the strengths of JBoss & Geronimo as compared to Weblogic & Websphere. Geronimo is not RI.

Will Geronimo be any good in a production environment or will it be just a Reference Implementation? In case the latter is true then what difference it would make if i use geronimo or jboss or etc during development. Can i use geronimo in a production environment where BEA & IBM rule?To expect Geronimo to match Weblogic or Websphere feature by feature or even in robustness (especially weblogic) would be little too much. What is really worth looking forward to is whether geronimo can fit somewhere between jboss and Weblogic/Websphere in terms of production quality.Nitin Chaumal

expect to match Weblogic or Websphere feature by feature or even in robustness

Most Open Source implementations have more features, more robustness and higher quality than propriatory alternatives. Start with Java Logging vs Log4j. Or Tomcat, Eclipse, Apache, Mozilla, ...

There are web doucments that state that CodeDevelopers Network has planed and refered to JBoss code when they wrote the GlueCode/BlueCode, the parts of code that they did not wite. If this is so, then derived works are GPL. Also products that use GPL also become GPL.

That is the effect.
.V

<blockquoteMost Open Source implementations have more features, more robustness and higher quality than propriatory alternatives. [...]
Why is OSS not propriatory, and commercial software is? I guess there is no such thing as a "Log4J standard".

Product quality is not only the "ilities", but also the documentation. _Most_ OSS are very poorly documented, in contrast to _most_ commercial software.

PS: To make on thing clear: I'm not against OSS, rather I use OSS very often and think it is a very good concept.

Regards,
    Dirk

Why is OSS not propriatory, and commercial software is? I guess there is no such thing as a "Log4J standard".

Not sure I want to argue with your general point, but this is a badly chosen example: I would say that there is, and that it is more of a standard than JDK 1.4 logging.

_Most_ OSS are very poorly documented, in contrast to _most_ commercial software.

Strongly disagree. Most commercial software I have used had appalling documentation, and even good documentation all too often lets you down completely when you hit a problem or there's something nonstandard that you need to do: i.e., when you really need it. In such situations, OSS at least has the ultimate (if hard to read) documentation available: the source code.

 - Peter

Why is OSS not propriatory, and commercial software is? I guess there is no such thing as a "Log4J standard".

Not sure I want to argue with your general point, but this is a badly chosen example: I would say that there is, and that it is more of a standard than JDK 1.4 logging.

Ok, maybe a bad example, but what I was actually trying to say is that distributing software as open source does not make it a standard _per se_. I admit, that commercial software has a more propriatory character than OSS in that I don't have a chance to influence the development of commercial software (at least not in such a degree I have with OSS).

_Most_ OSS are very poorly documented, in contrast to _most_ commercial software.

Strongly disagree. Most commercial software I have used had appalling documentation, and even good documentation all too often lets you down completely when you hit a problem or there's something nonstandard that you need to do: i.e., when you really need it. In such situations, OSS at least has the ultimate (if hard to read) documentation available: the source code.&nbsp;- Peter

I think we don't get together on that point :-), maybe we made different experience with commercial and OSS products. But the argument "the code is the documentation" is not valid for me. When things go really bad in a project I don't have the time to study tons of source code in order to _maybe_ find what I am looking for. I'd rather call the (commercial) product support to help me out. I made very good experience with that one for the BEA application server, for example (which also provide outstanding documentation, by the way; never seen that for OSS).

Regards,
   Dirk

[...] maybe we made different experience with commercial and OSS products. [...] When things go really bad in a project I don't have the time to study tons of source code in order to _maybe_ find what I am looking for. I'd rather call the (commercial) product support to help me out.

Yes, I think our different backgrounds are reflected in the shape of this discussion. My experience with commercial support is pretty dismal compared to the support I tend to get from OSS communites, or even compared to the support I can give myself given the source code.

I made very good experience with that one for the BEA application server, for example (which also provide outstanding documentation, by the way; never seen that for OSS).

Are we talking about big-name commercial software only? That casts a slightly different light on things. I still don't think you can say that most commercial software has decent documentation, but most big-name software probably does. But the big names may not have what you need.

In any case, for a current example of OSS with excellent documentation, a very responsive team, and a codebase that you can actually easily find your way around in should you wish to do so, have a look at the Spring framework. In terms of overall documentation, quality and support it's easily the equal of any commercial product I've ever worked with (haven't had the pleasure to work with the boys from BEA though).

 - Peter

and i m all for it. JBOSS needs to face some good open source competition.

JBoss already has open source J2EE competition, but I agree that more competition could have a positive effect on all these implementations.

Everybody refers as if JBoss was the only player in the Open Source application server market and that Geronimo will change that.

But what about the excellent application server that JOnAS is (LGPL) or even now Resin has a GPL version.

Have a look at the JOnAS admin console to see how well finished it is.

Everybody refers as if JBoss was the only player in the Open Source application server market and that Geronimo will change that.But what about the excellent application server that JOnAS is (LGPL) [..] Have a look at the JOnAS admin console to see how well finished it is.

I've yet to see anyone use JOnAS .. I looked at the list of success stories for JOnAS and didn't recognize any names except for Lofi's OpenUSS .. so is JOnAS just poorly marketed, considering it doesn't cost anything to use?

Peace,

Cameron Purdy
Tangosol, Inc.
Coherence: Shared Memories for J2EE Clusters

Have a look at http://jonas.objectweb.org/partners.html too

The big names apart Red Hat are in Europe...but that does not mean the product is not worth a look :)

Benjamin

.. so is JOnAS just poorly marketed, considering it doesn't cost anything to use?Peace,Cameron Purdy

I would definitely say that plays a pretty big part in it. I've been many places in the last 2 years (not as many as you for sure). And things that don't cost usually don't show up on their radar because no one is there to sell them the solution/tool/etc.

Of course things that do cost don't always show up because they
-- Don't research beyond the basic tools (Unbelieveable how so many exerienced web developers have never heard of Coherence)
-- Don't have any money left after spending it on expensive hardware/software/"consulting". So they don't look.

Oddly, even after people are shown tools that can help (save money and/or solve their problem) they still try to make that dam hold with dirt, spit and sticks.

.. so is JOnAS just poorly marketed, considering it doesn't cost anything to use?

I would definitely say that plays a pretty big part in it. I've been many places in the last 2 years (not as many as you for sure). And things that don't cost usually don't show up on their radar because no one is there to sell them the solution/tool/etc.

True, but JBoss has done a pretty good self-promotion job .. even back when the code-base didn't even compile I had heard of it ;-)

Of course things that do cost don't always show up because they -- Don't research beyond the basic tools (Unbelieveable how so many exerienced web developers have never heard of Coherence)

Speaking from experience, you are absolutely correct -- even with magazine ad campaigns and boothing at conferences etc. there is still a majority that has never heard of our software.

Don't have any money left after spending it on expensive hardware/software/"consulting".

We see this as well, although less and less as the economy picks up and as the "dot com bust" recedes into the past.

Oddly, even after people are shown tools that can help (save money and/or solve their problem) they still try to make that dam hold with dirt, spit and sticks.

If I had a dollar for every company I've visited that has built their own complete ORM framework, web framework, and/or application server .. and no, I'm not kidding. ;-)

Peace,

Cameron Purdy
Tangosol, Inc.
Coherence: Shared Memories for J2EE Clusters

True, but JBoss has done a pretty good self-promotion job .. even back when the code-base didn't even compile I had heard of it ;-)

But then at least someone was promoting it. But they really were promoting their service. And that definitely is a path to take. One I am looking at. Only supporting more things and at a higher level.

Don't have any money left after spending it on expensive hardware/software/"consulting".

We see this as well, although less and less as the economy picks up and as the "dot com bust" recedes into the past.

Hopefully they will spend that money wisely. Not gonna put good money on that though.

Oddly, even after people are shown tools that can help (save money and/or solve their problem) they still try to make that dam hold with dirt, spit and sticks.

If I had a dollar for every company I've visited that has built their own complete ORM framework, web framework, and/or application server .. and no, I'm not kidding. ;-)

I know you are not. I've seen it too. To top it off some don't realize they have "rolled their own".

I've yet to see anyone use JOnAS .. I looked at the list of success stories for JOnAS and didn't recognize any names except for Lofi's OpenUSS .. so is JOnAS just poorly marketed, considering it doesn't cost anything to use?Peace,Cameron Purdy

Cameron: you can't be expected to know everything! There's a difference between adoption and hype. Some big accounts actually use JOnAS in production but don't want to talk about it. They have their reasons.

Cameron: you can't be expected to know everything! There's a difference between adoption and hype. Some big accounts actually use JOnAS in production but don't want to talk about it. They have their reasons.

My question reflects my honest naïveté .. I haven't personally seen JOnAS being used so I can't recommend if / when to use it, etc. I don't know I am "typical" in not having seen it used; it does appear to be more used in Europe, for example. I guess what I'm saying is that, if JOnAS is good, then it could use a little marketing hype ;-)

Also, are parts of JOnAS being used in Geronimo? I don't understand the extent of the ObjectWeb involvement in Apache Geronimo. There's some donations being made, right?

Peace,

Cameron Purdy
Tangosol, Inc.
Coherence: Shared Memories for J2EE Clusters

Maybe it works so well they don't need a tool to help what your product solves. :) It would be a first.

If pricing and openess is all that matters, then all Germans would drive Japanese cars because they are cheaper, have better maintenance reports and come with a lot extra functionality. Believe me, they don't.

As with cars there are many factors that play in to purchase decisions. I would be surprised to see this server becoming a killer application. The article looks like whishful thinking to me.

Frank

If pricing and openess is all that matters, then all Germans would drive Japanese cars because they are cheaper, have better maintenance reports and come with a lot extra functionality. Believe me, they don't. As with cars there are many factors that play in to purchase decisions. I would be surprised to see this server becoming a killer application. The article looks like whishful thinking to me.Frank

Apache as a "brand" is now well know in enterprises thanks to the Apache Webserver (~78% market share and growing), Apache Jakarta Tomcat JSP/Sevlet container (which is part of Geronimo), and numerous popular libraries, both C++ and Java (Jarkarta).

So if Apache Geronimo does turn out to be a solid J2EE server and becomes J2EE 1.4 certified (as is the plan), it's going to the top of the short "alternative" J2EE server list for enterprises who currently use BEA WebLogic or IBM's WebSphere. Currently JBoss leads that list. I've seen more than a few companies that use WebSphere or Weblogic as their main server to make customers happy, but also use JBoss as well to reduce costs. The same goes for Apache's Tomcat.

Going forward I could see enterprises rolling out Geronimo instead of Tomcat, since Geronimo already constains Tomcat but add the numerous features that make up J2EE.

More competition is good for everyone and hopefully this doesn't hurt JBoss as I find it to be a great server. If anything I think it will just drive BEA and IBM higher up on the food chain, away from the commodity J2EE servers.

And remember, Geronimo is under the Apache license, so commercial versions could be forth coming from start ups.

cite="thread.tss?thread_id=30080#147003">Geronimo already constains Tomcat

I thought Geronimo used Jetty.

cite="thread.tss?thread_id=30080#147003">Geronimo already constains Tomcat

Well, I see TSS doesn't handle attributes in the <blockquote> tag. Sorry for gibberish before.

Well, ASF is not a software development organization down the street. ASF has given HTTP server, Tomcat Container and frameworks like Struts, Avalon, Cocoon etc and has contributed lot to Java bull. it's just a first foot forward in J2EE market by ASF and i'm sure they w'll make Geronimo another HTTP Server and Tomcat container. Most of the J2EE servers now uses Tomcat as their servlet container.

I'll wish to see Geronimo as a rock solid J2EE server


cheers !!