JBoss seems to be undergoing some generational pains as it strives to morph from an open source products company to an enterprise open source products company. So its recent formal announcements covered the enterprise tack: something called Enterprise Acceleration that performs the basic blocking an tackling to show enterprises …
Another article, another way to spread some critique of the JBoss salvo in to enterprise software: there is analysis and then there is 'momentum'-analysis, with JBoss doing well in spite of its inability or unwillingness to release a Java 5 server...
The Red Hat acquisition was a singular move to keep the at-that-moment leader in the open source ranks, and therefore, a superior cost model to proprietary Java platforms, truly one of the shrewdest moves in software...
This Accelerator program is a non-issue, though nice to see, with pre-built best practice environments, Muzilla needs to demonstrate that he knows how to ship product integration that can get to the 95% of non-app server experts...
Basically, this is a safe article, but it is getting easy to pile on JBoss for not yet taking "50% of all enterprise deployments" as Muzilla himself declared, even though the current state of middleware means that JBoss, and by virtue, RHAT, control its destiny as a deployment platform....
Two problems for JBOSS
Firstly the existing commercial J2EE customers of Websphere, Weblogix etc have for the most part invested substantial resources to development standards, change management, operational procedures, working fail over procedures and after years of effort are in a "comfort zone" with these products.
They are not going to go through all that pain and effort a second time just to save some small cahnge on license fees.
Secondly, nobody likes EJBs! If you look at any java development site/blog etc. the overwhelming opinion among experienced developers is "go for POJOs", if you must invoke remote processes use JMS, if you really think you need object relational mapping use Hibernate. For this type of development Tomcat, Jetty or Glassfish work just fine.
So Jboss will find itself in a very small niche market -- those sites that mandate or really require EJBs but dont have a bigbucks commercial infrastructure in place.
What you don't seem to realise is that there are a hell of a lot of companies out there using non-open source enterprise application servers (i.e Websphere or Weblogic).
Bea completely screwed up their 9 and 10 releases of weblogic, leaving only websphere or jboss as alternatives. Oracle's takeover of Bea means change and an uncertain future for the product - two things that large companies relying on enterprise solutions can't live with.
Currently JBoss is the tried and tested enterprise application server for EJB3 (based on POJOs and annotations [the java blogs and sites that you're reading are referring to EJB 2.x]) and it doesn't cost a company as much as an IBM enterprise solution. Thus, making it, in my opinion, the only alternative out there.
This makes Red Hat's debt of communication to the current JBoss community huge and very necessary for their own survival.
I have to say, I've always admired (what I presumed to be) their business model: get the geeks hooked on the open-source version, let the geeks impress management with all that the OS version can do, let management buy in Red Hat consultants because they don't trust their own IT. Everyone's happy: Red Hat get their pound of consulting flesh, management get a fully supported product and the IT get the open source infrastructure that they've always wanted.
- Product Round-up Smartwatch face off: Pebble, MetaWatch and new hi-tech timepieces
- Geek's Guide to Britain The bunker at the end of the world - in Essex
- FLABBER-JASTED: It's 'jif', NOT '.gif', says man who should know
- If you've bought DRM'd film files from Acetrax, here's the bad news
- VIDEO Herschel Space Observatory spots galaxies merging