While war has not yet been declared among Java application server vendors, this week's JavaOne conference in San Francisco did showcase their lack of unity. A light analysis of the vendors of Java One (by the Meta Group) discusses current offerings and positionings of the J2EE vendors.
Read Java's widening gulf
Competition in the app server space is always good. At least Oracle has finally admit that its own app server core technology, after years of marketing hype, has been just that, and has gone for the buyout approach.
However, unlike BEA's buyout of Weblogic a few years ago, Orion is a open source server with a small market share, which then begs the following questions
1) Why should I pay $10,000 (or whatever the 9i license fee) is for something I can get for free? Is Oracle support really THAT good? :-)
2) Weblogic, when it was a standalone company, had already a large marketshare, hence was production-tested. Orion, which theoretically performs better and is more stable than WL, does not have the marketshare or penetration to back up these numbers in large scale production environments. Can it do what it purports in the real world?
Also, at JavaOne a lot of good things were said about IBM's WS 4.0. I will definitely give them the benefit of the doubt and download an eval when it comes out. And finally, if Iona spend just a fraction of its resource in developing its app server as it did at JavaOne for its lavish party and contest promotion, then there's high hopes for this product as well! :-)
Please share why you think Orion is open-source, or where you got your information.
I was under the impression that Orion (at one time?) is an open-source server. OK, then I hope at least I'm batting .500 and am right with my second point, which is that Orion currently holds a small market share, especially in comparison to the market share Weblogic held when it was acquired by BEA.
I have seen the comparison thrown around of BEA and Weblogic - however another difference from that scenario is that Oracle has not bought out Orion - they are licensing their (Ironflare's) software. I also suspect that the deal has been in the making for quite some time and that there has been a lot of testing both live and stress that helped seal the deal.
I think this is a good move for Oracle and a good move for the App Server market - it gives developers and their customers a choice when it comes to big market app servers that support EJB 2.0 spec (albeit in varying degrees - but hey its a moving target) - not JUST BEA - plus Oracle products are free to develop with :-)
think this is a good move for Oracle and a good move for >the App Server market
I think a better move for their customers would have been for Oracle to license WebLogic -it definitely is the easiest to use and has the most support among developers, ISVs and Tools vendors.
Competition is better than consolidation. BEA is a fine product - I use it a great deal, but it is not always the best product. Its got the name, the developers, and the various vendors and it is easier for new developers to get an edge up but my clients want a range of choices. A choice between Oracle/Weblogic and Weblogic isn't a choice. Oracle and Orion provides a technically solid (and in some aspects better), fully supported solution that offers a real choice for my clients. The developers will come. And with Oracle and its name and reputation solidly in place - ISVs and Tool vendors will not be far behind.
Orion is *not* open source. It is also not free. Its developer license is free but it is $1500/box for commercial/production use. Note that license fee is per box and not per CPU.
has all of the details.
Orion is a open source server
This is incorrect. Check your sources.
Orion is free (no cost) for non production uses. Sources are never available.
WebSphere 4.0 looks really promising. Although a version behind, IBM is very careful not to leave existing customers in the dust. Weblogic took a bit of a hit with the EJB 2.0 spec, which is a cost of being bleeding edge. IONA's Orbix was as slow as a turtle so I can't imagine iPortal being any better. I can't say much about Oracle and IPlanet because I haven't used it.
I see Websphere overtaking BEA soon. IBM does a good job playing both sides of the fence(Microsoft/Sun) and is very advanced in web services and XML. Websphere 4.0 will be a very powerful tool. Although just being EJB 1.1, they will have Message driven beans and a JMS listener backed by the best messaging middleware in the industry.
Weblogic took a bit of a hit with the EJB 2.0 spec
Do you mean we're getting high while developing the EJB 2.0 container? Who told you?
More seriously, can you be more specific?
My experience is that customers really like the EJB 2.0 bleeding edge stuff, but it's only because we are providing a strong EJB 1.1 container behind, so they get to develop both production-ready EJB's while getting ready to transition painlessly to EJB 2.0.
WebLogic was done implementing the initial EJB 2.0 spec(persistent manager, etc..). IBM led the push in finding the hole in the spec, therefore causing the change to local interfaces and so-on.
Give WebSphere 4.0 a chance before you knock it. 3.5.2+ is an excellent release.
BEA was NOT done... where were the dependent object?
They finished implemented it but never released it.
IBM led the push in finding the hole
> in the spec, therefore causing the
> change to local interfaces and so-on.
I don't know where you got this information from. The discussions on the EJB 2.0 experts mailing-list clearly show that the convergence toward Local Interface was a collegial effort, not an idea pushed by one single vendor.
> Give WebSphere 4.0 a chance before you knock it.
I guess that's pretty much the point: it's not out yet. But yes, we're all waiting for it. However, it won't have CMP 2.0 as far as I know (and much less other PFD2 features such as Local Interfaces).
I remember someone in Bea said that Websphere wasn't competition because the application server didn't scale comparing with WebLogic.
Does anybody have any performance test comparing Websphere and WebLogic. I see a lot comments on Websphere, but seldom any comments that can give me some performance numbers that can "prove" that Websphere scales.
A JBoss developer said that JBoss "beat the crap out of Websphere" in a performance test.
My problem with application market it is a lot opinions on every product, but not hardcore facts.
Any input is appreciated.....
--- Trond Andersen
>WebSphere 4.0 looks really promising.
Even the 2.02 version sounded promising before I actually used it.
Even though our company uses WL and we have a good rapport BEA folks, I'm also on the lookout for a better and cheaper server. That's the whole point of Java and open standards!
Hence at this JavaOne I was furtively visiting all of the server vendors booth and getting demo copies and kits. (I felt the need to do this stealthly since BEA was a Platinum sponsor and most likely had many eye-in-the-sky cameras, tracking costomers that may become turncoats... :-)). I agree, IBM WS 4.0 seem to have a lot of potential, and I will look at it once it comes out at the end of the month. And Iona had a cool developer's kit, flashing LED light and all.. I'll even give Oracle a chance, especially since Ellison was SO CONVINCING in his keynote speech!
Competition is good. It's good for the vendors, and its definitely good for us customers. Don't you just love win-win situations?
What would be your key buying criteria if you were to choose an AppServer for a large project? Different people have different opinions. To me, it would be in the following order:
1) Stability is paramount - period. Although it's hard to measure
2) Skills availability in the marketplace - an AppServer is just a tool that enables the developers to build business apps. After all, it's the quality of the people that matters the most.
3) Documentation and support - on the same point, developers can only be effective with sufficient docs and support.
4) Performance & Scalability
5) Standards support - this is important, don't get me wrong here.
6) Ease of use
What is your take???
I would add a cost-benefit analysis. For instance if I just went on skills availability in the market place I would always choose something like BEA or IBM app servers. Big project, but what about cost contraints? They are showing up a lot more, even in large companies. I find clients want to look at some of the smaller market app servers like Sybase's EA Server and now we're starting to get some requests for Oracle's new app server version with Orion. Development costs and sometimes even deployment costs are cheaper on these products and depending on what you need, can solidly satisfy the requirements of the client.
Also - along with App server skill sets, I look for one or two developers who are strong in J2EE independent of a particular app server - always important in my mind - and they can pick up an app server pretty quickly - it just offers me a different perspective than we can get from the app server experts.