The death of bloated, heavyweight, monolothic enterprise Java …

… and the rise of the lightweight, agile, dynamic, modular Enterprise Java.

Java EE has long had the perception of being slow, bloated and monolithic. While the Java EE specification doesn’t prescribe a lightweight, modular implementation; neither does it preclude one. So the negative perception of Java EE is more to do with certain implementations than the standard itself. The two dominant commercial vendors of (Oracle and IBM) have, over the last 10 years, done a lot to re-enforce the perception that Java EE has to be complex, bloated and expensive. While at JBoss we’ve been pushing in a different direction for many years. Each of the last major release of JBoss AS has incrementally pushed modularity closer to an ideal where you only pay for what you need (in memory, complexity, CPU).

junk in the trunkPhoto Credit : “Junk in The Trunk”, zappowbang (CC, some rights reserved)

The latest version (JBoss AS 7) – which is also the underlying technology for Red Hat’s commercially supported product (JBoss EAP 6) has pushed the lightweight theme as far as to completely erode the difference between lightweight web containers (like Tomcat) and “heavyweight EE servers”. As far back as 2008 (through the advances in mobile devices and JBoss miniaturization) it’s been possible to run a full Java EE server on something as small as a smartphone; last year at Red Hat summit we ran the entire JBoss keynote demo (starts about 35min) on a cluster of cheap plug-top computers (far less powerful than today’s smartphone).

So it’s great to see IBM following our lead and pulling their own miniaturization stunt – Websphere running on a Rasberry Pi – very cool. Great to see IBM joining us to help change the perception of Java EE.

JBoss Data Grid – Making Big Data Work

Screen Shot 2012-04-13 at 10.11.51 AM.png  

This week Red Hat extended the Early Access program for JBoss Data Grid with the availability of the BETA – available to existing customer and future customers.

Traditional (typically relational) spinning-rust data stores have become one of the biggest economic and technical impediments to extracting the true value from the increasing amount of data that is available to organizations today.

JBoss Data Grid is a distributed, in-memory, fault-tollerant key-value store that is architected for large scale, mission critical applications. JBoss Data grid is built on the Infinispan open source project which is the natural evolution of JBoss Cache – which has been a core part of JBoss products for many years.

There’s a huge amount of activity around BigData and NoSQL as people look for more appropriate solutions to store, manage and analyze data – JBoss Data Grid and in-memory solutions in general provide orders-of-magnitude performance and scalability benefits without necessarily having to completely re-architect the data tier. The intrinsic distribution also provides a high-degree of fault-tollerance without the additional cost, complexity and overhead involved in traditional data stores.

So, give it a spin; and give us some feedback.

Red Hat is looking for some awesome Product Managers

Looking for a new challenge in a fast growing business ? Familiar with Java, JBoss and Open Source ? Able to kick competitor butt and have fun ?

Product Manager, JBoss User Experience (UXP) Mobile Platform

The Product Manager, JBoss User Experience (UXP) and Mobile Platforms Group is looking for a Product Manager, to be responsible for the overall product life cycle of Red Hat’s User Experience Platforms including Mobile and Portal. The Product Manager will be responsible for determining the market requirements for a successful product and driving the overall product strategy. This Product Manager will also work closely with the product development and engineering teams to determine product priorities and plan product releases and determine the product roadmap.

Sr Product Manager, JBoss Application Platforms

JBoss Enterprise Application Platform is the market leading platform for innovative and scalable Java applications. Integrated, simplified, and delivered by the leader in enterprise open source software. Responsible for the overall product life cycle of Red Hat’s application platform products marketed under the JBoss brand name. Responsible for determining the market and technical requirements for a successful product and driving the overall product strategy. Will work closely with product development to determine product priorities and plan product releases and determine the product roadmap.

Sr. Product Manager, JBoss Enterprise SOA

The Sr. Product Manager will be responsible for the overall product life cycle of Red Hat’s JBoss Enterprise SOA platform, Enterprise Service Bus, BPEL engine, related tooling and integration with related technologies. The individual will be responsible for determining the market requirements for a successful product and driving the overall product strategy. This person will work closely with product development and engineering to determine product priorities and plan product releases and determine the product road map.


Here follows answers to the the majority of the questions we had during the BETA webinars last week. I’ll add the remaining answers in he next few days. There were lots of questions so I’ve grouped them into categories – if you have additional questions or require more clarity, please leave a comment.

Note any mentions of specific features, component versions and dates in these answer are subject to change.

I’ll post the results of the EE 6 Poll in a separate post.

Finally some links referenced in the Webinars for you convenience :

The Webinar replay (will be available in a week or so)

Red Hat’s PaaS – OpenShift

EAP 6 BETA Download

EAP 6 BETA Documentation

The Q&A

Developer Tools

Q: the current JBDS 5 beta does not include EAP 6, when will it include EAP 6?

A: A brief time after EAP6’s GA date – JBDS will GA prior to EAP6 GA – so post EAP6 GA will be the first time for us to get them sync’d up. Between now and then, simply download both, it is easy to add EAP6 to JBDS5.

Q: Will Zip file with Maven artifacts be provided in EAP 6 distribution or it will be a separate deliverable?

A: It’s a separate download on the Support Portal.


Q: does the openshift flex support postgresql database?

A: yes it does.


Q: What about the migration from EAP 5 to EAP 6, is it transparent, or is it different?

A. There are definitely some major differences in configuration and management but applications should be largely portable from EAP 4 / 5 to EAP 6. We have created a migration guide and we also offer migration services.

Q: Would you please spend a few word regarding the migration path for Seam applications that currently target the EAP 5 release?

A: We have documentation for how to make Seam 2.2 applications run on EAP6. Furthermore, we will be delivering Seam 2.3 which will maintain the Seam 2.2 programming model but allow you to take advantage of JSF2.

Q: Being already partners, is there a path or redhat offer to help poc’ing the migration from EAP 5 to EAP 6 ? Thanks

A: Red Hat Consulting have a full set of migration services.


Q: will you publish the questions/answers ?

A: Yes.

Q: When EAP 6 GA will be available?

A : Subject to the outcome of the BETA, performance and quality – we plan to release within the next 4 months.

Q: You said you wanted to make sure you don’t do a vendor lock in, but how do you make sure that other vendors won’t lock you in?

A: Avoid proprietary technologies (languages, frameworks, tools) that are only available from a single vendor. Choose open standards and open source. Ask yourself this question about your technology stack – if the (single source) vendor goes out of business or pushes their prices up significantly what is your contingency plan ?

Q: Any idea when the beta’s will become stable?

A: EAP 6 GA is planned within the next 4 months but we may distribute a second BETA release before that.

Q: Will the support Lifecycle of Jboss EAP 4.3 be extended because of the late release of Jboss EAP 6?

A : We have a an extended lifecyle program for EAP – contact your local Red Hat representative for more details.

Q: Pros-Cons Domain Architecture versus standalone Architecture : will you better support the first mode ?

A : They will both be fully supported. Domain mode if you want to run multiple instances for HA or scale-out and don’t have other management tools. Standalone mode if you’re developing or don’t want our management wiring to get in the way of other management tools you use or have developed.

Q: What are technical differences between EAP 6 and AS 7.1 ? ( management tools, fixes, deploy tools, etc ? )

A : At the point of the BETA the difference between AS 7.1.Final and EAP 6 BETA are pretty limited – some branding and some critical bug fixes. As we progress through the BETA and our testing we EAP 6 will start to diverge from the last available AS release and that divergence will continue as we start the maintenance stream for EAP 6. More information on our model here.

Q: when will the Jboss Training courses start covering EAP 6, most spefically, the JBoss Application Administration course (JB336)?

A: JB366 (JBoss EAP 6 Admin.) will be available at GA – in fact you’ll be ale to take the course at Red Hat Summit / JBoss World in June. We’re also developing a new core developer course for EAP 6 that will be available at GA.


Q: Have you enable data replication to hornetQ in EAP 6

A: It’s being worked on upstream but not planned for EAP 6.0 GA. It is on the wish-list for EAP 6.1

Q: When will HornetQ implement AMQP?

A : There are no plans for HornetQ to implement AMQP. Red Hat has a separate messaging product (Red Hat Messaging) that does support AMQP – Red Hat Messaging can be used as an alternate JMS provider for EAP.

Core App Server

Q: Will EAP5 bootstrap be updated to be compatible with Java 7? Or do we need to upgrade to EAP6 in order move forward in Java?

A : The plan is to certify EAP 5 with Java SE 7 after the GA release of EAP 6 – so within the next 3-6 months.

Q: Does EAP 6 contain tomcat 7.

A : EAP 6 embeds a modified version of Tomcat 6 with support for many Tomcat 7 features.

Q: When is EWS expected to uptake Tomcat 7 ?

A : JBoss EWS 2.0 will support Tomcat 7 and will be available later this year.

Q: Is there any EJB 2.1 support?

A: Yes but definitely in maintenance mode – you should consider moving to EJB 3.1

Q: Downloaded beta 1, can’t seem to find a way to “install it as a service” for Windows. Is this going to be introduced later?

A : The Windows Service Wrapper didn’t make the BETA – it will be in the GA and in BETA 2 (if we do one).

Clustering / Caching

Q: Why do not integrate The full Infinispan product in EAP 6 (not only for 2nd level cache or session replication)?

A: A number of reasons. The cache instance used for Query caching and session replication is tuned for a specific use-case, enabling it for general purpose use would likely impact both session replication and query caching. We have decided to offer a separate product (JBoss Data Grid) that is specifically for use a a general purpose, in-memory key value store. It can be run in embedded mode and also standalone and satisfies a broader range of deployment topologies and designs.

Q: Is JBoss Data Grid announced for end-March derived from Infinispan project ? Or any other relation between them ?

A : Yes JBoss Data Grid is based on Infinispan.

Modules / Class-loading

Q: Can we replace class on demand on working instance, without redeploying?

A: Possibly. Not out of the box. You’d need something like JRebel or Stuart’s Fakereplace to make it work. And of course there are limits to what kinds of changes can be made with that sort of mechanism.

Q: Will you support to access a Java module inside of an EAR via JNLP?

A: Sounds like a cool idea, though you can probably do this already using simple servlets.

Q: Can we point to application classloader what jar version should it use when we have many version of the same jar?

A: It depends. If you have more than one version deployed as deployments, or your JARs are installed as modules (either with different names, or with the same names but different versions) then yes, you can choose which one to link to simply by using the right Class-Path or Dependency entry, or their jboss-deployment-structure.xml equivalents. We do not support more than one version of a JAR within the same EAR’s lib/ directory or within the same deployment, and each deployment can only refer to one version of a JAR. You can however have an EAR with two subdeployments in it, with each sub-deployment linking against a different version of the same JAR (as long as the conflicting dependencies are never brought into contact with each other).


Q: What is the status of PicketLink in EAP6?

A : PicketLink is included and fully supported.

Q: Is there any internal API to confirm x.509 certificate validity.

We do not provide any internal API to perform the x509 validation. You can use the standard JDK Classes for that. Usage of X509 depends on the subsystem:

If you are using JSSE via the JDK, then you can install your own Trust Manager

For the Web subsystem, you can use CLIENT-CERT and use any of the x509 based login modules.

Management Questions

Q: Can the domain controller be started from the command-line interface (CLI) as well?

A: No – you have to start it manually (or run as a daemon / service)

Q: What kind of scripting tool will be available in EAP6 ? Can all the tasks can be performed via the scripting interface ?

A: JBoss CLI was cover in the BETA Webinar, there are some more details in the docs.

Q: Is JON 3 packaged with EAP6 then? or still separate?

A: All JBoss Enterprise Platforms have a managed option that include entitlement for JBoss ON; though the products have their own distribution / installation.

Q: Will EAP6.0 contain more streamlined patches/fixes and will there be a better way to apply patches instead of copying jar files at different locations ?

A : Local patching is scheduled for EAP 6.1

Q: Does EAP 6 CLI able to deploy a web app on many instances at once ?

A: Yes – the CLI is fully domain aware.

Q: Can we see/configure all application servers under domain controller from the admin console of “domain controller” ?

A : Yes there is a physical view of running instances within the domain.

Q: Where web console keeps our modification? Does it change the XML files or what?

A: Yes. Any config. modification made through the management APIs are persisted in the underlying confg. files.

Q: Would twiddle still be supported?

A: No – we no longer support Twiddle but you can use something like JConsole to get at the underlying JMX MBeans or use jboss-cli.

Q: Does jboss-cli compatible with previous Jboss EAP releases?

A: There was no equivalent to jboss-cli in previous release of JBoss EAP.

Q: Will Domain Controller introduce a single point of failure (either at runtime for applications or for management operations) ?

A: No it isn’t involved in application request flow. In fact it’s quite common to disable the management infrastructure in secure production environments. We chose not to encumber the Domain Controller with an overly complex HA design, when the simple solution is to re-start it if it fails – or run it as a daemon / service so it is auto-started.

Q: is it possible to use JON3 to make automated deployments of jboss applications?

A: Yes., more info. in the JBoss ON 3 Documntation.

Q: If we can change configuration of other instances remotly, can it be done in transactional way – I mean all of the node confirm the change, or none of them in case of some failure

By default, if you make an administrative change that affects multiple servers, the Domain Controller will apply it on all affected servers and will roll it back on all affected servers if it fails on any of them. This constraint can be relaxed when using the CLI by adding an additional header to the operation request that specifies a “rollout-plan”. Among other things, rollout plan can specify that a certain number or percentage of servers in a server group can fail to accept the change without triggering an overall rollback. (See “Operations with a Rollout Plan” on for some conceptual information about rollout plans.

JBoss EAP 6 BETA Webinar


Join me next Wednesday 3/14 – I’ll be talking about how to get involved in the JBoss EAP 6 BETA program. World-wide-friendly times :

  • Wednesday, March 14, 2012 | 14:00 UTC / 9am (New York) / 2pm (Paris) / 6:30pm (Mumbai)
  • Wednesday, March 14, 2012 | 19:00 UTC / 2pm (New York) / 7pm (Paris) / 11:30am (Mumbai)
  • Thursday, March 15, 2012 | 01:00 UTC / 6:30am (Mumbai) / 9am (Singapore) / 12 noon (Sydney)

Sign-up for the free Webinar here.

With all the enthusiasm of cleaning up a dog turd …

… Heroku finally got around to supporting Java. But they couldn’t do it without first piling on some hate.

Why then, if Java is such a miserable platform to develop on would Heroku bother ?

Here are a couple of thoughts :

1. Huge Developer Base


2. Massive Adoption

Screen Shot 2011-08-26 at 4.22.48 PM.png

3. Large, Active Ecosystem

Only Java gives developers such a broad range of tools, technologies and APIs – both commercial and open source. Only Java gives you Open Standard enterprisey features like Transactions, Object Persistence, Messaging, Security, Integration, scalability and high availability for when you need them.

Basically, most professional developers use Java (if they aren’t beholden to Microsoft that is) and they, to a degree, decide how to spend money on deployment and long-term care and feeding of applications. And Heroku, like any other company wants to make money.

But why would a professional Java developer choose Heroku given their very out of date and poorly informed opinion of Java ?

Surely – Red Hat’s OpenShift is a better choice ? Instead of whining about Java’s shortcomings over the years – Red Hat / JBoss has put a huge amount of energy in fixing Java’s shortcomings – and doing it in an open and collaborative way so the entire ecosystem can benefit. Red Hat has a deep understanding of Java technology and open collaboration – more than anyone else in the industry. OpenShift’s support for Java EE 6 is a recent example of this – we didn’t sit around complaining that Java EE didn’t fit with the new deployment paradigm that PaaS represents – we simply did what we had to do to make it work. And you can try it for free.

Finally, as Isaac makes clear, it’s time for the folks at Heroku to wise up about Java and maybe trade in their 2004 copy of “Beginning J2EE 1.4: From Novice to Professional” and have a look at some of the advancements in Enterprise Java over the last decade. Hey – Adam – leave a comment and I’ll buy you a new book 🙂

Rethinking Enterprise Java

JBoss AS 7 has been out for a week or so – probably not enough time for opinions to be formed but the feedback I’ve seen so far has been overwhelmingly positive. But that isn’t the subject of this post.

You probably didn’t see Red Hat’s press release as those things are typically only read by the press so I wanted to draw to you attention a single paragraph :

“JBoss Application Server 7 represents a major milestone in the evolution of Java application servers from complex and monolithic to more lightweight, modular and agile. This release will enable developers to re-think how they develop and deploy enterprise Java applications.”

I wrote that and I meant it. Over the last 5 years there has been a significant difference between the Java EE servers like JBoss, Weblogic and Websphere and Apache Tomcat. Tomcat has been the poster-child for the lightweight container movement (but we shouldn’t forget Jetty and Resin – both very capable servers) and has established itself as probably the most popular Java run-time.

But I think we’re at the point where there is no-longer a lightweight division between Java EE servers and Tomcat (and other Web Containers) – some good blog posts here and here that discuss typical developer requirements like startup time and deployment speed (make sure you read the comments). When we’re at the the point where we’re discussing sub-second differences between startup or deployment times then I think we have convergence. I think we’re at a point where you can no longer paint Java EE servers into the big, slow and heavy corner and Tomcat into the lean and fast corner.

Developers have more choice today than ever before – they can choose a lightweight container but no longer have to make a tradeoff between footprint and features. Or start with a basic server like Tomcat and incrementally build a full featured application server from the ground up as more features like caching, persistence, transactions, messaging, view layer are required.

OK before the Glassfish fanboys chime in – yes Glassfish did a great job of addressing light-weight needs for Java EE some time ago but by any available measure Glassfish still doesn’t represent a mainstream choice like Tomcat, JBoss, Weblogic or Websphere.

Not all free advice is good advice

Mike Gualtieri at Forrester has a rather sensationalist post with a rather sensationalist title “Stop Wasting Money On WebLogic, WebSphere, And JBoss Application Servers”. I agree with some of what Mike says and have been giving the same advice to developers, customer and users for at least the last decade. Specifically – choose the run-time that best suits your needs. If all you require is a Servlet container then Tomcat is very likely the way to go; very likely the least complexity you need. If your application has greater needs – Transaction coordination, caching, clustering, ORM / Persistence, messaging, a modern UI framework then you probably need to look at one of the commercial or open source Java EE Application Servers (Weblogic, Websphere, JBoss, Glassfish).

The mistake that people make (Mike included) is that vanilla Tomcat addresses the full spectrum of enterprise web / app platforms. It simply doesn’t. There is only a small class of applications that need nothing more than a servlet container. The vast majority of Tomcat deployments require a whole bunch of additional technology – the de-facto Tomcat stack is Tomcat + some view layer tech. (eg. Spring MVC) + some persistence / ORM (eg. Hibernate). That is the basis for the alternative to Java EE. Most people end up adding a lot more stuff than that and find themselves somewhat accidentally maintaining a full featured application server. I know of customers who have been very successful doing this; they’ve invested in very knowledgeable people to manage the stack; I’ve also seen customers fail or simply decide that they aren’t willing to become an app. server vendor.

Finally, full disclosure – I do work for one of the vendors Mike mentions – specifically Red Hat and we would be delighted to sell you a subscription to a full featured Java EE Application Server (and a broad range of middleware) but if you have chosen Tomcat – we can help you with that too – we provide a distribution of Tomcat and even certify and test other elements of the Tomcat stack like Hibernate and Spring. Or as most customers actually demand – we can help you with both.

Lightning Strikes !

Screen shot 2011-07-05 at 8.02.30 AM.png

Just 6 months after JBoss AS 6 was released, JBoss AS 7 (codename Lightning) is now available. Congratulations and a big thank you to the JBoss AS team and community. JBoss AS 7 is a major release in every respect and will become the technology underpinning for much of what we do at JBoss for the next decade. I believe it also represents a shift in the way developers will think about enterprise Java and it opens up new possibilities for deployment that were unthinkable 5 years ago due to technical and economic limitations.

If you’ve been following the AS 7 candidate releases (and AS 6 before it) then you already know that AS 7 includes some significant new features. I’m not going to list them all; but here are the highlights:

Developer Productivity

  • Startup-time and memory utilization have been significantly reduced which leads to a much more productive developer experience – no more coffee breaks during deployments and restarts.This required some significant rethinking and a fair amount of innovation (something we’re good at apparently)
  • The Java EE 6 Web Profile provides a much leaner, less complex platform for developers who focus purely on the web-tier – less to learn, fewer design choices – increased developer productivity
  • More flexible and powerful modular classloader – less time debugging and configuring classpaths; more time writing applications
  • Testable by Design with Arquillian with out of container testing for the business logic so developers can be more productive while delivering better quality applications.

Price / Performance

  • It’s probably a little early to claim significant performance gain over the competition right now but request path performance is a goal and the hard work of tuning and performance improvements starts now, One early indicator will hopeful give you a sense of what we’d like to achieve is the recent SPECjms2007 submission from Red Hat. SPECjms is a pretty narrowly focussed benchmark and not all the JMS vendors are represented, that said this is pretty significant for us as it is the first public benchmark submission from JBoss and good practice for future activities

Operational Ease of Use

  • Some of the more significant advances in JBoss AS 7 are around the operational ease of use. The configuration has been completely refactored around a multi-node domain model, though the simple single-instance view has been maintained for developer use as well
  • There are stable, easy to use management APIs – so AS 7 deployments can be completely automated from Java or any other scripting environment.
  • New shiny, task oriented domain console that also allows you to manage multiple, distributed nodes.

Anyway – time to stop reading and start playing : learn more about JBoss AS 7 here and get the bits here and provide feedback on community site.

Next post – how AS 7 relates to Red Hat’s commercial, fully supported distribution – JBoss Enterprise Application Platform 6.

Thoughts on G+

My opinion on G+ (based on about an hour of tinkering) :

  1. G+ will be of limited utility until the majority of people I know or care about is using it (ie. it has to catch up with FB and Twitter)
  2. Circles – great start but needs finishing – need more set theory – subsets, intersections, compliments, etc. will make it very powerful though maybe a little nerdy for the non-technical.
  3. UI is very slick – amazing demonstration of what is possible in a standard browser
  4. I represents a personal dilemma. I’ve been using Facebook for mostly personal stuff; Twitter for mostly work stuff. With a little natural overlap. I can manage two social networks; three will be tough.

This is my profile if you want to connect.