Discussions

News: Oracle becomes sponsor and co-spec lead of EJB 3

  1. Oracle today announced that it will join Sun Microsystems as the co-specification lead for the Enterprise Java Beans (EJB) 3.0 specification, and will provide the persistence technology in the Reference Implementation of Java Enterprise Edition 5.0.

    The reference implementation will be derived from Oracle TopLink, a Java object-to-relational tool and deployment platform that provides advanced object-relational mapping capabilities suitable for mission critical enterprise applications. Oracle will also contribute components of the reference implementation to Project GlassFish, Sun’s Open Source implementation of the Java Enterprise Edition 5.0 Software Developer’s Kit.

    The EJB 3.0 Preview is available free for download at http://www.oracle.com/technology/ejb3 .

    Threaded Messages (42)

  2. Seems both parties are desperate, though in a different sense. Oracle is desperate to grab some J2EE market share, while Sun, a hardware company, is desperate to hold its relevance in the Java space.
  3. Seems both parties are desperate, though in a different sense. Oracle is desperate to grab some J2EE market share, while Sun, a hardware company, is desperate to hold its relevance in the Java space.

    George,
    I'm really surprised at your ignorance and your comment that "Oracle is desperate to grab some J2EE market share". Oracle is growing its market share than any of its competitors.

    I'm sure you have seen this news earlier and it was posted in a thread that Oracle is a close third with 19 per cent market share last year.

    So you do not spread the FUD that Oracle is desperate. The reality is Oracle is gaining market share faster than its competitors.

    have a look at the news report ..

    http://news.com.com/IBM+continues+gains+in+server+software/2100-1012_3-5210844.html


    regards
    Debu Panda
  4. If EJB3 spec had been closer to Hibernate, it would have been the death of TopLinK. I think Oracle may win to Hibernate this way, although they must open their TopLink toolkit.

    Regarding Sun, they will try that Oracle was more supportive to their hardware platform, I suppose.

    Joserra
    Mi bitácora tecnológica
  5. If EJB3 spec had been closer to Hibernate

    Hibernate3 will implement the EJB3 API
  6. If EJB3 spec had been closer to Hibernate
    Hibernate3 will implement the EJB3 API
    related to persistence
  7. Pity[ Go to top ]

    Having in mind the previous software quality of Toplink I have pity on the reference implamentation. Hopefully the major players (IBM, BEA) will have their own implementation.

    I agree that Oracle tries to grab as much of the market as possible. It almost looks like a (final) struggle for their existence in the J2EE sector...
  8. Pity[ Go to top ]

    Having in mind the previous software quality of Toplink I have pity on the reference implamentation.

    What quality problems have you found in Toplink? We have been useing it for years very successfully. I was worried when Oracle first got its hands on Toplink, from WebGain. But IMO they have done a good job extending and expanding it, and kept it open to work on all databases.

    If EJB3 spec had been closer to Hibernate, it would have been the death of TopLinK. I think Oracle may win to Hibernate this way

    Have you built a EJB3 bean based on Hibernate3 yet? The EJB3 spec and Hibernate3 w/annotations could not be any "closer" to eachother. I have no idea what you are talking about here. Plus neither product is going to "Win" over the other. Both products are so far out in front right now, with respect to EJB3. Plus Oracle shops are going to keep using Toplink, and open source developers are going to keep using Hibernate. I don't think Oracle working with Sun is going to change much. Unless you are worried that everyone is going to start using the Sun Application Server har har har. I think it in the end it will be a big win for both.


    The thing that I find interesting about this article is that if Oracle is going to provied the reference implementation to an open source application server, then are they going to have to open source parts of Toplink.

    Also I wonder if the reason that sun went with Toplink is because they have gui support for their ORM mapper? Personally I just write my EJB3 Entity bean by hand, but maybe sun thinks EJB3 will be better recieved if people have a gui to build them.

    Also TTS please provied a link to this anouncement. I would like to read exactly what it says, as sometimes things don't get paraphrased correctly here.
  9. May I ask which version of TopLink you tried and what type of problems you have been running into? We have been using top link internally on some very wide scope and large scale use cases and we have been rather impressed with the results.
    Edwin
  10. There goes any reason for me to keep using EJBs. Rather use Spring and Hibernate for my J2EE applications.
  11. There goes any reason for me to keep using EJBs. Rather use Spring and Hibernate for my J2EE applications.

    Same here - we were creating "J2EE without EJBs" (a-ka web applications...) applications but now we're starting to migrate to EJB, coordinating the full development to 2006 when EJB3 will be out. EJB3 promises to be the next big thing with regards to server developments - and just thinking about the delights Spring will add to it makes me grin :)
  12. There goes any reason for me to keep using EJBs. Rather use Spring and Hibernate for my J2EE applications.

    The largest database vendor, who develops a respectable ORM product is helping sun with the reference implementation for a persistance spec.

    Yes by all means let run for shelter, the sky is falling!! Thanks for the insight Chick Little.
  13. "The largest database vendor, who develops a respectable ORM product is helping sun with the reference implementation for a persistance spec.

    Yes by all means let run for shelter, the sky is falling!! Thanks for the insight Chick Little. "

    Look at Oracle's Java offerings from the past and current, yes this is a concern. Oracle delivered an awful J2EE application server and I had to work with it Oracle had to get a copy of Orion's application server to get a working product out. So, the largest database couldn’t deliver software on J2EE specs.

    Now, Oracle is a spec-lead of EJB 3? My excitement on this great next release is over, time to start lookung around for other options.
  14. George,

    When was the last time you use OC4J? I would be interested in following up on this discussion. If you have the time, please send me an email to edwin.khodabakchian @ oracle.com

    Thanks,

    -Edwin
  15. Hi Edwin,

    I sent you email concerning my issues with Oracle's quality of service regarding Java tools.

    You have not responded, no surprise, Oracle's strategy of putting R&D money over to India is factual evidence of agenda.

    I hope the USA market will vote against Oracle's products and support good old USA hard work.

    Folks, just compare what Oracle has to offer, I got tired of India support, not able to support my concerns serving my USA clients.

    Don't get me wrong, I'm all for international market competition, but cheap service is a strategy decision Oracle has chosen.

    Don't settle for this, Oracle has not earned the right to contribute to the J2EE community!
  16. Oracle has not earned the right to contribute to the J2EE community!

    There is no need for this type of elitest crap. Unlike you I'm sure most people would love to see the whole world involved in the development and evolution of java.

    and support good old USA hard work

    Who are you talking about anyway? Sun, IBM, Apache, JBoss, JoAns, etc are all international companies or groups made up of internation developers.
  17. "There is no need for this type of elitest crap. Unlike you I'm sure most people would love to see the whole world involved in the development and evolution of java."

    Of course, the development and evolution of java by everyone is great!

    My critizism was for Oracle, if you look at the low quality of Oracle's Java products with relationship to its database incenses revenue - you clearly see, contribution to java is really weak.

    I may be wrong, tell me, anyone - what java innovation has Oracle created?

    "Who are you talking about anyway? Sun, IBM, Apache, JBoss, JoAns, etc are all international companies or groups made up of internation developers."

    Agreed, international contribution and competition are vital for all.

    Again, of all the money Oracle is making from the database market (love that cpu license count) the strategy to offshore develop a product, is yet another example of Oracle trying to cheap their way in.

    I think before somebody becomes a co-spec of EJB 3, they should have some kind of "actual" credibility. If money is the only criteria, then let Wal-Mart become a co-spec of EJB 3 as well.
  18. I think before somebody becomes a co-spec of EJB 3, they should have some kind of "actual" credibility.

    As a member of the EJB3 expert group and a direct competitor of Oracle's offering in the ORM arena, I can assure you that Mike has years of ORM expertise. He has been an influential member of the EJB3 expert group to date, both in terms of his technical input and his diplomatic attitude towards problem resolution. In my experience on expert groups, strong technical founding in the problem domain and a will and aptitude to build consensus are invaluable. He's been helping Linda with persistence-related administrivia more and more over the last months, so I think that codifying his efforts in this new role is a good thing all around.

    -Patrick

    --
    Patrick Linskey
    http://solarmetric.com
  19. I think before somebody becomes a co-spec of EJB 3, they should have some kind of "actual" credibility. If money is the only criteria, then let Wal-Mart become a co-spec of EJB 3 as well.

    Your issue is with Oracle, and yes I agree Oracle has made some very shitty Java products, but I would disagree that *all* of their Java products are crap. Sun and IBM have made plenty of crappy Java products too. JBoss, JonAs, and Tomcat have contributed their fare share of crap also. Once again I'm not sure who you are trying to use as your standard.

    With respect to the EJB stuff, it is the Toplink group that is working with Sun on the spec. Toplink has a much longer, and (with respect to Java) more successful track record then Oracle. Mike is also well respected in the industry of persistance. Throwing their experience and contribution to the community in with a bunch of rants about Oracle as a whole is not justified.
  20. With respect to the EJB stuff, it is the Toplink group that is working with Sun on the spec. Toplink has a much longer, and (with respect to Java) more successful track record then Oracle. Mike is also well respected in the industry of persistance. Throwing their experience and contribution to the community in with a bunch of rants about Oracle as a whole is not justified.

    This once again indicates ORM should be in a spec of its own, instead of being part of EJB. It is ridiculous to say "ORM is not EJB, but EJB is an ORM standard".
  21. My critizism was for Oracle, if you look at the low quality of Oracle's Java products with relationship to its database incenses revenue - you clearly see, contribution to java is really weak.I may be wrong, tell me, anyone - what java innovation has Oracle created?
    I think before somebody becomes a co-spec of EJB 3, they should have some kind of "actual" credibility. If money is the only criteria, then let Wal-Mart become a co-spec of EJB 3 as well.

    George,

    It sounds like you may have been the victim of some of Oracle's very early forays into the Java market. Oracle is one of the first people to admit that they made some mistakes when they started with Java (who didn't?) especially in the area of Java experience. It's just that I think that you may have a stale view of the new reality of Oracle, Java and J2EE.

    Not only did Oracle acquire the Orion server quite a while back and use it as leverage to build a high performing full-featured server, but they also agressively hired lots of very bright and experienced people to work on the Enterprise Java platform. This platform has emerged as a major product in the J2EE marketplace. Look at statisics, market share and what ALL the analysts are saying about who are poised to be the leaders in the near future.

    As for what innovation Oracle has contributed I could list many technologies and point you to a number of JSR's that we have submitted, led or currently play prominent roles in. In fact two people that have posted to this thread are good examples of other technologies: Cameron is the spec lead of JCache (JSR 107) which was initially proposed, led and implemented by Oracle, and Edwin created a BPEL server that is the envy of the BPM world and that is now part of Oracle. I have seen that server held up in multiple talks by all sorts of BPM afficionados as an example of how well this stuff can work.

    Finally, and I guess this is probably the most important correction being that it is the topic of this thread, I want to update you to the fact that by virtue of Oracle owning TopLink it actually has more credability and experience on persistence than any other vendor -- bar none. TopLink has been deployed on more enterprise applications and for longer than any other persistence product. We were teaching courses on persistence 10 years ago when the product first became a hit on WebLogic (before it was owned by BEA) and we were the first vendor to provide POJO persistence and EJB CMP within the same product.

    Java is a fast moving business (yes, it is a business, not just a platform), and things change very quickly. I realize that it is hard to keep up, and I wish it didn't take so much time to keep on top of it all, but that is the business we are in and partly what makes it so exciting.

    If you have any technical issues with the EJB 3 spec I would be happy to address them either personally or through the EJB 3 feedback group.

    -Mike

    PS Thanks Matt and Patrick for standing up in my absence :-)
  22. George, while I suppose I await the more flames in this seemingly degenerating thread, I am following up with Edwin as well - he was doing a good samartin service on behalf of the OC4J team (his team BPEL Process Manager runs as a J2EE app on OC4J) and once I get either your e-mail to him or from you directly (mike dot lehmann at oracle dot com) I will see how best we (the OC4J team itself) can answer your issues.

    As a by the by, development for OC4J is in fact done globally - yes, like many companies we have folks in India but actually most of the team is in fact here in the US though a another major chunk reside in Canada where, for those familiar with TopLink know, TopLink started its life.

    Mike.
  23. version?[ Go to top ]

    What version of OAS?

    I find that most developers complaining about Oracle J2EE "anything" were using early 9 versions. I also had problems with it.

    After moving to 10g, we've had much fewer problems.
  24. If it's so great then why is Oracle telling Retek to rip out Hibernate but to not implemnt Toplink? Doesn't that seem odd? Oracle just bought Retek a couple months back. If it's not good enough for Oracle why would it be good enough for me?
  25. This seems to turn into a classic TSS fud-fest. I'd like to add our experiences to give hopefully some substance.
    Using Toplink for a year and a half now I must say that the jdeveloper toplink mapping userinterface caused a lot of hairpulling in my team.
    But when we switched to the standalone mapping workbench everything was fine: I like the way the workbench gives you feedback immediately as you enter a bidirectional relation for instance and you have not defined the other side of the relation. That quick feedback without having to run unittests I find very helpful.
    The production version of Toplink (9.0.4) has a large API that you need when doing anything non-trivial, that is undocumented. The way to find out what to do is by scouring forums etc. That took us some time to figure out.
    The next version (10.1.3) seems to be documented a lot more extensively though, but that's not production-ready.
    I was very disappointed when it turned out that toplink does not have a distributed caching option out of the box. This has to be developed for you by consultants.
    All-in-all I'm quite happy we chose toplink as our persistency engine. Although we have not gone in production yet, so I can't comment on throughput and availability yet.
  26. JDeveloper and TopLink[ Go to top ]

    I have also pulled a few hairs re: JDev and TopLink.

    All of my problems ended up going to the TopLink and JDev teams. They were very interested in ironing them out for 10.1.3.

    So I have high hopes for JDev 10.1.3...
  27. If it's so great then why is Oracle telling Retek to rip out Hibernate but to not implemnt Toplink? Doesn't that seem odd? Oracle just bought Retek a couple months back. If it's not good enough for Oracle why would it be good enough for me?

    Brain, can you post a link to this info, not that I don't believe you I would just like to read more about it.

    Thanks,
  28. If it's so great then why is Oracle telling Retek to rip out Hibernate but to not implemnt Toplink? Doesn't that seem odd? Oracle just bought Retek a couple months back. If it's not good enough for Oracle why would it be good enough for me?
    Brain, can you post a link to this info, not that I don't believe you I would just like to read more about it.Thanks,

    Heard it from a guy that works at Retek.
  29. If it's so great then why is Oracle telling Retek to rip out Hibernate but to not implemnt Toplink? Doesn't that seem odd? Oracle just bought Retek a couple months back. If it's not good enough for Oracle why would it be good enough for me?
    Brain, can you post a link to this info, not that I don't believe you I would just like to read more about it.Thanks,
    Heard it from a guy that works at Retek.

    What did Oracle recomend they do?
  30. EJB3[ Go to top ]

    Maybe EJB3.0 is so great that Oracle is going to eventually recommend CMP over O/R mapping? (because it actually works?)
  31. EJB3[ Go to top ]

    Maybe EJB3.0 is so great that Oracle is going to eventually recommend CMP over O/R mapping? (because it actually works?)

    EJB3 and ORM are becoming the same thing.
  32. EJB3[ Go to top ]

    And the next two specs co-leaded by Oracle will be: EJB3 Forms and PL-EJB3 :-).
  33. Good for Oracle[ Go to top ]

    I'm another one that has had success with TopLink. And in fact, since the 10g release, both Oracle's App Server and JDeveloper IDE have improved exponentially.

    I'm really looking forward to what Oracle will bring to the JEE space in the next couple of years.
  34. Why does everyone using Hibernate feel the need to bash EJB 3.0 before it has even been released? Feeling a little insecure or just astroturfing?
  35. Wait and see EJB will die in a few years..Ofcourse there would be legacy stuff in EJB
    People are moving to Spring and Hibernate in tonnes.
    Who is that senseless at Oracle to go this way?:)
  36. Wait and see EJB will die in a few years..Ofcourse there would be legacy stuff in EJBPeople are moving to Spring and Hibernate in tonnes.Who is that senseless at Oracle to go this way?:)

    I agree with you Spring is in fasion. I have written a few Spring apps, and I like Spring too. But that has nothing to do with *who* is working with EJB3 with respect to persistance. JBoss/Hibernate are pushing in this direction just as hard as Oracle/Toplink are. Plus Spring now supports Toplink too. You make it sound like these products are at polar oppists of eachother when infact they are becoming more and more similar and intermixable.

    (Yes I know I suck at spelling.)
  37. My biggest gripe with EJB(1,2 or 3) is that you need a high priced container to stage things(well if you go the Geranimo way or something free and stable...no).
    And EJBs force you to write senseless object oriented code.
    And clustering with entity beans is a ..well try it!

    The only good thing Oracle did for us J2EE lovers is that they went out and bought Orion and incorporated it as their engine and called it OC4J. That was an excellent decision on the part of Oracle. When compared to Websphere, Weblogic etc Orion is an excellent sturdy server which is easy to install and work with. But I think them(oracle) going after TopLink was a dumb decision. (TopLink is too cumbersome to use, if you ask me) Likewise this decision to co-spec EJB3 (which is definitely gonna die) is dumb. Waste of their energies IMO
  38. My biggest gripe with EJB(1,2 or 3) is that you need a high priced container

    I think you should visit this site here if you are concerned about high priced containers. These guys have been around for 6 years ;-) and have more market share than Oracle, IBM, or BEA.
    And EJBs force you to write senseless object oriented code.And clustering with entity beans is a ..well try it!

    Try out this online tutorial. I think you will be pleasantly surprised by how nice EJB3 is.

    Bill

    P.S. Sorry Debu, Mike. Couldn't resist....
  39. Bill -
    I think you should visit this site here if you are concerned about high priced containers. These guys have been around for 6 years ;-) and have more market share than Oracle, IBM, or BEA.

    For someone who complains a lot about FUD, you're pretty damned good at it.

    First, for businesses that require support and indemnification, JBoss is a very "high priced container". More than one company I've talked to claims to pay well over $10,000 per year per server for your container.

    Second, please stop citing an online poll (the BZ study) as a market share indicator.
    P.S. Sorry Debu, Mike. Couldn't resist....

    I'm only sorry that they're too polite to respond to your FUD. Bill, you are a shameless snake oil salesman.

    Peace,

    Cameron Purdy
    Tangosol Coherence: Clustered Shared Memory for Java
  40. My biggest gripe with EJB(1,2 or 3) is that you need a high priced container to stage things(well if you go the Geranimo way or something free and stable...no).

    There are almost as many free application server as non-free ones: JBoss, JonAs, Geronimo, or Sun Application Server. How many options do you what?

    And EJBs force you to write senseless object oriented code. And clustering with entity beans is a ..well try it!

    Are you making this comment about EJB3? If so then you have not tried them out? For the most part it works like any other ORM that uses POJOs.
  41. Curious to know...[ Go to top ]

    Will Oracle allow me to use Hibernate3 instead of TopLink?

    Will Oracle now promote TopLink in favour of BC4J (aka ADF)?

    Will development for this application server be closely tied to their IDE, JDeveloper?
  42. Curious to know...[ Go to top ]

    Will Oracle allow me to use Hibernate3 instead of TopLink?

    yes, they have to. The EJB3 spec requires app-server vendors to support pluggable persistence providers. Review the schema for persistence.xml.
  43. All the Java products Oracle have done are extremely low quality (bc4j and adf comes to mind to name a few). Oracle involved in the ejb implementation makes me shudder.