Discussions

News: IBM Releases Developer Preview: J2EE 1.4 Compliant Server

  1. IBM has announced WebSphere Application Server Technology for Developers V6. The code provided is an early release of the WebSphere Application Server with the latest J2EE Version 1.4 for Windows specifications. IBM claims that this is one of the first J2EE 1.4 compliant servers out there.

    The J2EE 1.4 compliant WebSphere Application Server Technology for Developers V6 will help you:

    - Improve integration with customers, suppliers, and partners with complete Web services support, including:
    . JAX-RPC 1.1 API that supports service endpoints based on Servlets and Enterprise Beans and provides interoperability with Web services based on the WSDL and SOAP protocols.
    . Web services for J2EE specification (JSR 109) that defines deployment requirements for Web services and utilizes the JAX-RPC programming model.
    . Platform Web services interoperability via the WS-I Basic Profile 1.0.

    - Leverage existing assets with new functions with Java Connector Architecture (JCA) 1.5:
    . Resource adapters provide bidirectional connectivity (synchronous and asynchronous) between EISs (Enterprise Information Systems) and J2EE-comliant applications.
    . JMS pluggability allows customer to use existing database software instead of a closed, proprietary means to provide message persistence.

    - Take advantage of server enhancements that enable faster, easier development with new Java Servlet and Java Server Pages (JSP) technologies:
    . Servlets now support request listeners and enhanced filters.
    . Simplified JSP technology page and extension development models make it easier than ever for developers to build JSP-enabled pages.

    Learn about the new WebSphere Application Server Technology for Developers V6

    Threaded Messages (20)

  2. A DAO?[ Go to top ]

    Why not use a DAO, like jBoss has Hibrenate and others (me included) use iBatis, others use Cayene, etc.
    I realy think if IBM offers something KISS in DAO, it be great for us, to have a commercial alternative to Eclipse + Tomcat 5 + iBatis + Struts.

    .V
  3. I agree[ Go to top ]

    I'm a developer in Brasil, and alternatives like Eclipse, JBoss and Tomcat are very good in a Country where investments are hard.

    Regards,

    WarMath
  4. I agree[ Go to top ]

    Marucio,

    Especially in USA, the big rich profitable companies hate spending money and love geting value, so they use open source.

    BUT... a few companies that are not as profitable, decide to use commercail products. Sometime I have to consult those as well. It be nice if they did not have to suffer, all I am saying for these poor folk, let them have simiplicity.

    .V
  5. I agree[ Go to top ]

    Marucio,

    >
    > Especially in USA, the big rich profitable companies hate spending money and love geting value, so they use open source.
    >
    > BUT... a few companies that are not as profitable, decide to use commercail products. Sometime I have to consult those as well. It be nice if they did not have to suffer, all I am saying for these poor folk, let them have simiplicity.
    >
    > .V

    Yes, Vic, big, successfull companies use open source and small, unsuccessfull companies use WebSphere and Weblogic. Excellent analysis coming from a sharp mind. I do wonder which planet you are from.

    Br - J
  6. I agree[ Go to top ]

    Johan,

    The evidence I have is anecdotal. I have a poor client, a state, that is using expensive comercial sofware, they are cutting back services for the poor.
    I have a few commercial clients, one of which has almost doubled in stock price. They only use open source.

    In information age, IT departments that add value (cost vs benefit) help the organization. Not so hard to understand. I bet that I can show your organization's managers that OS is better support, better value, etc. etc.

    All I said with orginal post is the some commercail vendor should do KISS so developers that are sometimes forced to use commercial stuff do not have to suffer. For example change an action in java and restart the server to test this new action. Or EJB, when DAO's are a better value.

    Profitability is important, else CXO outsource or cut back in other ways.

    .V
  7. I agree[ Go to top ]

    Johan,

    >
    > The evidence I have is anecdotal. I have a poor client, a state, that is using expensive comercial sofware, they are cutting back services for the poor.
    > I have a few commercial clients, one of which has almost doubled in stock price. They only use open source.

    I dont doubt that this is true, but I doubt if it is representative for the situation in large.

    >
    > In information age, IT departments that add value (cost vs benefit) help the organization. Not so hard to understand. I bet that I can show your organization's managers that OS is better support, better value, etc. etc.
    >
    > All I said with orginal post is the some commercail vendor should do KISS so developers that are sometimes forced to use commercial stuff do not have to suffer. For example change an action in java and restart the server to test this new action. Or EJB, when DAO's are a better value.

    I dont see EJB´s and DAO´s as competing technologies/patterns, I see them as complimentary. We use stateless session beans with daos for almost everything. The beans gives us declarative transactions, object pooling, security etc. Sure we could write these services ourselves, but why would (did) we?

    In any case, the licensing fee for an application server is a very small part of the total cost for developing and maintaining an application over the whole life-cycle. Even if one buys an expensive IBM or BEA product.

    Br - J
  8. I agree[ Go to top ]

    Johan,

    John said:
    I dont see EJB´s and DAO´s as competing technologies/patterns, I see them as complimentary. We use stateless session beans with daos for almost everything. The beans gives us declarative transactions, object pooling, security etc. Sure we could write these services ourselves, but why would (did) we?

    Vic: EJB's are a fraud. There are several thread on TSS that say something like:
    "At a seminar I was at, 9 out of 9 experts said that they can think of a possible scenario when they would use EJB" . No need for a new thread of DAO vs EJB, I among others posted a lot on it.
    Lets say you have a factory and a vendor sells you a machine that should have scalability and simple services; if that turns out to be a lie you migh sue that vendor.
    Just use a DAO like iBatis, you won't ever look back.


    John said.
    In any case, the licensing fee for an application server is a very small part of the total cost for developing and maintaining an application over the whole life-cycle. Even if one buys an expensive IBM or BEA product.

    Vic: Lets say one has a $300,000 project. One could buy vendor propriatory db (for development, staging and production), buy vendor propriatory app. server (for development, staging and production), buy IDE per developers, buy vendor propriatory servers (for development, staging and production)... it's like a fever. Plus the more complex the sofware, the more HW you need.
    You could eat up 1/3 of the budget.

    Or you can do open standards DB (pgSQL - free and fast), open standards container (Tomcat 5), Eclipse, Linux... and buy a mass produced HW like a NewISys2100. Then you can have more expereienced developers and you have not touched the budget, you have room for ROI.
    Once you are done developing, with J2EE as cross platgorm, you can deploy your war file on many contianers. You could just do OpenSta stress test, and see what container is most effective. Even if you develop WebShpere, you want to develop in a way that it would work on WebLogic... or other open standards. A big benefit of J2EE is competition. So if you develop on Tomcat, and IBM is faster or more reliable or cheaper or whatever... you can deploy there.

    Vendor propriatory does have a cost.

    But that was not my point at all. I was saying some developers are forced to work with vnedor propriatory tools. I was just saying I wish IBM would come up with something that works simpler, like jBoss did with Hibrenate instead of EJB and complexia.
    If IBM would use a DAO, for example, that would be great. Becuase some of the time I do consulting at sites that do vendor propriatory.

    .V
  9. I agree[ Go to top ]

    Lets say you have a factory and a vendor sells you a machine that should have scalability and simple services; if that turns out to be a lie you migh sue that vendor.


    Well, that just isnt my experience with ejbs in general and websphere and weblogic ejb containers in particular. I have participated in performance tests both for applications running on weblogic and websphere, and there has always been performance problems. The problems have never been related to the performance of the ejb container though, but rather been found in either application implementation or design.

    > Vic: Lets say one has a $300,000 project.

    What would be the cost of maintaining the application during the whole life-cycle? The total sum of maintenance and development will be much higher than 300K - in comparison the cost of the license for the app server will be quite a small share.

    >Even if you develop WebShpere, you want to develop in a way that it would work on WebLogic... or other open standards. A big benefit of J2EE is competition.

    Totally agree.

    > If IBM would use a DAO, for example, that would be great. Becuase some of the time I do consulting at sites that do vendor propriatory.

    Well, WebSphere or Weblogic is no more proprietary than JBoss or Jonas.

    Anyway - we have differing opinions, lets rest at that,

    Br - J
  10. Postgres?[ Go to top ]

    Vic: Or you can do open standards DB (pgSQL - free and fast) ...

    How do you define "fast"? I'm just curious. Maybe Postgres is fast enough for the projects that you work on, but it is at least 90% slower than Oracle under heavy load, and even worse when it crashes. It's not particularly stable when it's under load. And frankly, saving the customer a couple of dollars but putting their data at risk sounds like a bad trade-off to me.

    (On the other hand, maybe the project you described is using only read-only database replicas or something. I don't want to be critical of your use of Postgres, just to point out that it isn't comparable to something like Oracle for a transactional system or for safeguarding data or for up-time etc. And for read-only database replicas, I prefer MySQL ;-)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  11. Oracle?[ Go to top ]

    Camerron,

    Caemron: Maybe Postgres is fast enough for the projects that you work on, but it is at least 90% slower than Oracle under heavy load, and even worse when it crashes. It's not particularly stable when it's under load. And frankly, saving the customer a couple of dollars but putting their data at risk sounds like a bad trade-off to me.

    Vic:
    http://jamesthornton.com/acs/benchmarks-ora817-pg703.html
    - this says a lot above.

    Maybe you are confusing MySQL as unreliable:
    http://www.phpbuilder.com/columns/tim20001112.php3
    SourceForge used to run MySQL, not it runs...
    A busy site, would you say?

    http://advocacy.postgresql.org
    http://www-css.fnal.gov/dsg/external/freeware/mysql-vs-pgsql.html
    I use a caching disc controller, and my OpenSta test show pgSQL plenty fast.
    No mater how much money you spend, one should stress test.

    Oracle is a little money to acquire and operate? Oracle is very expensive, something like $45K per CPU. SO if you have a 4 cpu server (newisys 4100 for ex) and a fail over or a stageing, there goes the money, like $300,000. I can develop and operate a large web site for that.
    Oracle is funky JDBC, it needs a lot of DBA time to set up, and it is most expensive DB by far. There are people that feel that the more money they spend, the better it is.

    I think you mixed up your facts, pgSQL runs circles ROI arround Oracle, and so does M$-SQL. Stored Procedures, Ansi SQL, stability, ... if your profitability is important.

    I do agree with you the MySQL is a toy and it's not real open source either. But pg SQL rocks.

    .V
  12. Oracle?[ Go to top ]

    Hi Vic,

    Well, I'm not a PostgreSQL expert, and my negative experiences are second hand (I wasn't the one tuning / maintaining / complaining.) I have relatively good experiences with it directly, but I've never tried to use it for a db server handling several thousand transactions a second either.

    OTOH, I have seen Oracle in those situations many many times, so I learn to trust it. (Also Sybase, but it is mostly a dead product line, even though it is widely used for financial systems.)

    Peace,

    Cameron Purdy
    Tangosol, Inc.
    Coherence: Clustered JCache for Grid Computing!
  13. Oracle?[ Go to top ]

    Cameron,
    >
    I've never tried to use it for a db server handling several thousand transactions a second either.
    >

    Sites like SourceForge.net are huge, and no problem with pgSQL. Even when using propriatory vendor solutions I assume we both do:
    1. Stress test to find the breaking point of performance.
    2. Write cross platform - so that you can re-deploy ANSI SQL to another ANSI SQL DB.

    So if we test and write ANSI SQL, no issues. Just like a .war file, deploy against a faster and a more cost effective container.
    I found this to be true on OSS:
    http://www.opensource.org/advocacy/case_for_business.php

    Back to IBM thread for a second: I wish that a commercial vendor, such as IBM, would offer J2EE as cost effiecent to use as what we have in open source, for example with a DAO. (What I am saying in between the lines is I gave up on Sun doing J2EE right, so I hope IBM ... or BEA offers something for those that wish to pay.)

    .V
  14. Oracle?[ Go to top ]

    Sites like SourceForge.net are huge, and no problem with pgSQL. Even when


    Vic, I can assure you that this interesting fact will have absolutely no impact on most companies that currently use Oracle, or Sybase, or DB2 on iSeries or zOS.

    Br - J
  15. Oracle?[ Go to top ]

    John,

    :-)
    That's cool, it won't happen overnight!
    Do most companies use use some OSS? Linux, Struts, Apache, .....
    http://news.netcraft.com/archives/web_server_survey.html

    http://www.opensource.org/advocacy/case_for_business.php
    http://www.technewsworld.com/perl/story/31901.html

    ROI is good. OSS goes trough a back door in open department, and then it spreads. No budget meetings, no legal review,...
    Anyway, just write an ANSI SQL app and you'll be set, it does not matter how it gets deployed. You can save the organization some $, just start playing with it.

    .V
  16. Oracle?[ Go to top ]

    You're not the only one with those experiences. I did some informal db testing at a previous company. Fastest to slowest: sapdb, oracle, mssql, pgsql. pgsql was almost twice as slow as anything else. Everything was running out of the box defaults.

    People also say mysql is super-fast but turn on transactions, foreign keys, etc and watch what happens to those benchmark times. I'm usually an open-source fan but the core oracle database is one of the very few closed-source components that has never let me down.

    mike

    > Well, I'm not a PostgreSQL expert, and my negative experiences are second hand (I wasn't the one tuning / maintaining / complaining.) I have relatively good experiences with it directly, but I've never tried to use it for a db server handling several thousand transactions a second either.
    >
    > OTOH, I have seen Oracle in those situations many many times, so I learn to trust it. (Also Sybase, but it is mostly a dead product line, even though it is widely used for financial systems.)
    >
  17. Where are the benchmarks?[ Go to top ]

    Vic:

    >http://jamesthornton.com/acs/benchmarks-ora817-pg703.html
    >- this says a lot above.

    Is it just me, or does the above page have no Oracle benchmark results due to some Oracle restriction? And no conclusion section, so you can't tell anything from the above page other than the performance of pgSQL itself?
  18. Postgres?[ Go to top ]

    Vic: Or you can do open standards DB (pgSQL - free and fast) ...

    >
    > How do you define "fast"? I'm just curious. Maybe Postgres is fast enough for the projects that you work on, but it is at least 90% slower than Oracle under heavy load, and even worse when it crashes. It's not particularly stable when it's under load. And frankly, saving the customer a couple of dollars but putting their data at risk sounds like a bad trade-off to me.
    >

    Probably you have used some of "old" pgsql and your expierence is out of date (It needs some knowlege to tune postgres too). It is very stable and fast RDMS and it is used for "mission critical" applications too. I do not say oracle is a toy like MySQL, but most of things are better on pgsql. Have you ever tried tools like psql,pg_dump and "interfaces" like jdbc ?
    Both databases implement advanced concurrency control and recovery, query optimization.
    Looks like garbage collection is better on oracle (vaccum), but I am sure it will be fixed by pgsql too.


    > (On the other hand, maybe the project you described is using only read-only database replicas or something. I don't want to be critical of your use of Postgres, just to point out that it isn't comparable to something like Oracle for a transactional system or for safeguarding data or for up-time etc. And for read-only database replicas, I prefer MySQL ;-)
    >
    > Peace,
    >
    > Cameron Purdy
    > Tangosol, Inc.
    > Coherence: Clustered JCache for Grid Computing!
  19. Well, just to be different I thought I'd reply to the original message rather than the record? sub-thread.

    Hi Billy, good to see WebSphere at the front of the technology wave for a change, I was rather expecting to see the J2EE 1.5 spec out in beta2 before IBM released their 1.4 version but I'm impressed, are they turning over a new leaf or have you just replaced the usual programmers (all of them)? :-)

    I for one will be trying it out, I must admit that most of my clients (banks) seem to be a little cool (not cold) on EJB architecture recently but a lot are using Web[Sphere|Logic] so I s'pose I'd better learn my way around it.

    I agree with other people's comments on JBoss, I'm a great fan but only a select few banks are using it in production and the same is sadly true for PostgreSQL, MySQL and other very respectable OS products. The only thing that seems to be in common use is Apache's "Apache" and Tomcat.

    Let's see what J2EE 1.4 does for the J2EE world in 2004, I'm going back to play with my JavaSpaces.

    Keep up the good work Billy, BTW, are you still doing all this on that OS/2 lunch box you used to have? :-)

    -John-
  20. Mr Davies...[ Go to top ]

    Nope,
    The lunch box is dead but I still have the case ;)

    Billy
    WAS 6.0 HA Lead
  21. Websphere 6.0[ Go to top ]

    1) Will Websphere 6.0 implement the Service Data Objects specification?

    2) Will Websphere 6.0 implement the WorkManager for Application Servers
       specification?

    3) Will Websphere 6.0 implement the Timer for Application Servers
       specification?

    http://www-106.ibm.com/developerworks/java/library/j-commonj-sdowmt/