Friday, 7 October 2011

Announcing the WAS V8.5 alpha

The year was 1911. While most were trying to understand how to create accounting machines and punch card readers, one man had a vision. He created a company with the sole intent of constructing a lightweight, fast, consumable application server that was so fast and easy to use, a developer could run it on even the slowest of punch card readers. "The server must start within 5 seconds" he exclaimed! "The server footprint cannot be bigger than 50 megabytes!" he mandated. While some of his more forward-thinking colleagues questioned how a production-ready server with such strong enterprise qualities of services could run in just 50 MB of space, others were puzzled by what an application server was and baffled by this talk of a "megabyte". They called him a dreamer and questioned whether this was even possible. "I want this application server technology to be available, even if it takes a hundred years!" he replied.

The rest is a history that is very much left out of the IBM Centenary literature. However, WebSphere Application Server V8.5 Alpha provides all this and more, using the new Liberty profile. Want to know more? Why not take a look round this new development community, WASdev. You can download the alpha server and tooling. We'd love you to let us know what you think on our forums, and to collaborate with us to help create a next generation development platform that works for you.

Thursday, 23 June 2011

The link between WebSphere Application Server Community Edition and Apache Geronimo

In previous posts I've written on this blog, such as Get OSGi samples running on your machine in 10 minutes and OSGi Bundle Lifecycles I've referred to Apache Geronimo as the light weight application server I've been using.  Geronimo is an open source application server that consumes the Apache Aries project.  However the food chain does not stop there.  An IBM product called WebSphere Application Server Community Edition consumes the Geronimo project.  So what can we infer from this food chain in the future?

Well, Geronimo currently offers full support for the Java EE 6 and enterprise OSGi applications (via integration of Apache Aries).  I asked Kevan Miller, an IBMer who is the PMC Chair of Apache Geronimo and the WebSphere Application Server Community Edition Architect, to clarify the intent for the Community Edition product.  Here's a slide Kevan often uses which carries a clear statement of WAS CE future intents.

I'm looking forward to seeing these functions in WAS CE, particularly the Java EE 6 support and the new OSGi enterprise programming model to WAS CE, in the OSGi platform.

IBM’s statements regarding its plans, directions, and intent are subject to change or withdrawal without notice at IBM’s sole discretion.   Information regarding potential future products is intended to outline our general product direction and it should not be relied on in making a purchasing decision. The information mentioned regarding potential future products is not a commitment, promise, or legal obligation to deliver any material, code or functionality. Information about potential future products may not be incorporated into any contract. The development, release, and timing of any future features or functionality described for our products remains at our sole discretion.

Monday, 13 June 2011

External Bundle Repositories

An external bundle repository is a location that can host a number of OSGi bundles and an xml file, which describes the bundles in the OSGi Bundle Repository. The xml file specifies the capabilities and requirements of a number of bundles together with their URLs. The bundles can live in the same location or different locations from the xml file.

Apache Aries, WebSphere Application Server V7 + OSGi and SCA Feature Packs and WebSphere Application Server V8 all support external bundle repositories. However, a common frustration among those who use external bundle repositories is the creation of the repository. Emily Jiang, a member of our development team, recently developed a repository generation tool that's worth sharing (instructions can be found on the Apache Aries website). The tool takes a list of URLs and generates a repository xml file describing the URLs. If the given URL is a local file directory, the tool will generate metadata for all valid bundles in that directory, which saves the user significant effort, as they do not need to explicitly list every bundle in the target directory. Subdirectories will be recursed for any bundles that exist higher in the directory structure.

For more instruction about how to administer external bundle repositories, check out the docs for the WebSphere V8 beta infocenter:

Wednesday, 18 May 2011

New Enterprise OSGi YouTube video: Blueprint

Blueprint provides a dependency injection framework for OSGi and was standardized by the OSGi Alliance in OSGi Compendium R4.2. It is designed to deal with the dynamic nature of OSGi, where services can become available and unavailable at any time.

A new video on the Enterprise OSGi YouTube channel gives a brief introduction to the Blueprint container. 

Tuesday, 10 May 2011

Contributing to the Enterprise OSGi YouTube channel

I've uploaded a new video to the Enterprise OSGi channel, contributed by @iocanel.  The video demonstrates basic Karaf features.

If you would like to contribute a video based around OSGi please contact me (@sjmaple/on this blog) and we can work to get it on the channel.

Wednesday, 20 April 2011

OSGi Bundle lifecycles

One aspect of the OSGi model which distinguishes it from say Java EE, is the OSGi bundle lifecycle.  During the last few OSGi sessions I have presented, the bundle lifecycle has always been of interest and has often been the source of many questions, so I thought I’d talk about it in a blog post :o).  The easiest way of describing the model is to dive straight into the state diagram that describes the lifecycle.

The diagram, as with any state transition diagram, shows the possible states an OSGi bundle can be in at any time within an OSGi framework, along with the possible transitions it could make to other states.

There is one way into the state diagram (signified by the black blob at the top), and one end state (UNINSTALLED).  These are the ways which a bundle enters and exits an OSGi framework.

I’m going to use the Geronimo OSGi runtime that was created in an earlier blog post to show a bundle going through the various lifecycle states (feel free to run on just the OSGi frameworks if you wish - although the commands will be subtly different).  First, I’ll remove the bundles that were added in the last tutorial by running the following command from the /opt/geronimo/geronimo-tomcat7-javaee6-3.0-M1 directory (alter the file locations to suit your dir structure).

./bin/deploy –u system –p manager undeploy

Right, we now have a clean deployment, so lets install a bundle out of the aries blog sample.  Extract the bundles from the .eba file and install the biz bundle into the framework.  Run the following command from within the Geronimo console:

install file:///opt/geronimo/geronimo-tomcat7-javaee6-3.0-M1/

You should have been returned a bundle id.  Type list and the biz bundle will now be the last in a long list of bundles.  Notice it’s in INSTALLED state.  We’ve reached the first state in the lifecycle diagram, above.  If we try to resolve the bundle by running the resolve command, the bundle remains in the INSTALLED state as the current OSGi framework does not contain the required bundles to satisfy the biz bundle’s requirements.  To check which dependencies the bundle is missing, try to start it by running start <bundle id>

Lets now install the API bundle:

install file:///opt/geronimo/geronimo-tomcat7-javaee6-3.0-M1/

If we view the list of bundles again we notice both our new bundles are in the INSTALLED state. However if we now try to resolve all bundles in the framework, by calling resolve. We notice that both bundles have now transitioned into the RESOLVED state.  This means that the bundle dependencies from both bundles have been met by the existing bundles in the OSGi framework. 

Install the final two bundles and resolve them:

install file:///opt/geronimo/geronimo-tomcat7-javaee6-3.0-M1/

install file:///opt/geronimo/geronimo-tomcat7-javaee6-3.0-M1/


Next we can actually start the bundles, by running start <bundle id> for each bundle in turn. Once all have been started, run the list command to ensure they are all in the ACTIVE state and then go to the blog sample webpage to make sure the blog is running.  You are also able to start a bundle from the installed state if it can resolve its dependencies. The framework will resolve the bundle first and then start it.

Right, so lets have a bit more of a play… lets stop a bundle and see what happens - First let’s stop the persistence bundle.  We’d expect the server to get to the web content as the web bundle still exists, however it will likely fail much beyond that.  Go to the webpage again, and you should see an exception from the web bundle part of our application stating the BloggingService is unavailable. This is because the BloggingService requires another service that was provided by the persistence bundle and therefore no longer exists in the service registry.  We should restart the persistence bundle and check the webpage to ensure we're all working again.

Let’s now stop the API bundle, by calling stop <bundle id> and they try to access the webpage again. Still working?  OK, lets go a stage further and uninstall the API bundle, by calling uninstall <bundle id>.  Check the webpage… Still working? Good isn’t it :)  This works because the consumer bundles that use the API bundle have previously cached the code it requires from it's dependencies, the API bundle.  The API bundle does not register any services, so the other bundles will not miss them, meaning the app can continue to run without the API bundle existing in the environment until a refresh is called on the OSGi framework.  The refresh wires up the new dependency map, refreshing the caches, which now will not include the API bundle.  This test was designed to show that the dynamism of OSGi is in the services rather than the bundles.  Yes, they are linked, without bundles there would be no services.

We can also uninstall bundles that are in the ACTIVE state.  Again, the framework will call stop on the bundle under the covers to get it into the RESOLVED state first and then uninstall the bundle.  The bundle is then back to the black blob at the start of the lifecycle :o)

Tuesday, 5 April 2011

Enforcing OSGi Modularity

Tom Watson is an IBMer based in the U.S. and a co-lead of the equinox project in eclipse.  He has kindly put together a presentation for our EnterpriseOSGi YouTube channel on how OSGi uses modularity.  Do go and take a look here.

Thursday, 31 March 2011

Get OSGi samples running on your machine in 10 minutes

I've seen a few blogs that show how to get Apache Geronimo up and running, followed by how to install the Apache Aries Blog sample to use the neat OSGi functionality. However, that's typically where it stops!  So this post will hopefully be the end to end tutorial for a developer wanting to play with OSGi from scratch in about 10 minutes.

Let's start from scratch... We're going to Install the Apache Aries Blog sample on Apache Geronimo.  Not sure how many people already want to google 'Apache Aries' or 'Apache Geronimo', so I'll explain...

Apache Aries

Apache Aries is a set of Java components that can be added to an existing OSGi framework, such as equinox or felix, to provide an OSGi developer with Enterprise functionality, which a typical EE developer would be used to, such as JNDI, JPA, Transactions etc. To read more on Apache Aries, visit

Apache Geronimo

The Apache Geronimo project produces an application server runtime that uses other open source components to create a rich application environment.  It provides application containers for Java EE applications and OSGi applications (via Apache Felix). One of the open source components it contains is Apache Aries. For more information about Geronimo, check out the project website

So our picture of Geronimo so far is as follows:

Our OSGi bundles essentially run in the Apache Felix framework.  However with the Apache Aries functionality that sits alongside the Felix framework, they are able to connect with some of the typical JEE functionality as well as use other OSGi application features.

You'll need a Java runtime installed on your machine (Make sure the JAVA_HOME and JRE_HOME system variables are set), but I assume you all mostly have this already.  So start your stopwatches... Download Geronimo v3.0 M1, as this has the Apache Aries functionality, from you'll want one of the drivers from the "Non-Certified Web Container Releases" section.  You can choose which web container you wish to use, I chose Tomcat. 

Once your download (65MB ish) is complete, gunzip and extract your tar to an appropriate location on linux or just unzip somewhere in windows.  Your Geronimo server is ready to go! Simple as that (hopefully).  To start your server, go to the "geronimo-tomcat7-javaee6-3.0-M1/bin" directory or equivalent and type "./geronimo run". (Ensure you have root privileges)

You should see this screen:

Don't get too excited yet though, as it processes a little more before you should start playing. Wait till you get to this screen:

Right, at this stage our Geronimo Application server is up and running as you can see from the last line of output. Hit enter to get a prompt back.  We can see from this output that we have a few app modules installed including an EAR, a couple of RARs and some WAR files.  Let's now install our own OSGi application.

We need to install a datasource for the application, so firstly, download a rar file that is used to install the datasource here. This will be included in Geronimo in future versions.  Next download the actual data source descriptor here. Once those files have been downloaded, put them in the root directory of your extracted Geronimo server.  From a new command prompt, navigate to the geronimo rooot directory and run the following command, with the server running, to install the datasource:

./bin/deploy -u system -p manager deploy tranql-connector-derby-embed-xa-1.6.rar aries-datasource.xml
We're now ready to install the application.  Download the Apache Aries application binary from  I downloaded the Blog JPA app. Click the 'eba' link to download.  With the .eba file in the root Geronimo directory, run the following command, again whilst the server is running:

./bin/deploy -u system -p manager deploy
To view the blog sample, point your browser to http://localhost:8080/blog.  You should see the following:

Well, you've now got a running Geronimo server and Apache Aries Blog Sample running on your machine in (hopefully) about 10 minutes.  Feel free to have a play around with this sample - It's not massively interesting but is fit for purpose :o)  

Lets now look under the hood in the runtime :o)  From the Geronimo console, type 'list'.  This returns the full list of bundles in the Geronimo framework, including system bundles.  Quite a few, aren't there!  At the end you'll see the blog bundles we just installed:

What we see above is a quite a bit of info on each bundle we have in the environment.  Lets take the first bundle in the highlighted list and look at it in more detail. Important columns to mention are id, state and description columns. The first element is the bundle id.  This is a unique number given to every bundle in the framework so that we can use the number in commands, like stopping and starting a bundle.  The next element is the state of the bundle.  'Active' means the bundle has been started and can fully interact with other bundles.  The description column at the end is how to visually see which bundle is which.

So, now you can play :o)  hitting <tab> gets you a list of commands you can call from a geronimo prompt.  Commands of interest from an OSGi perspective include headers, list, refresh, restart, start, stop, update, exports and imports.  You can get more info about these commands by typing '--help' after the command, although they're all pretty straight forward.

Given this is quite a long post already, I'm going to leave it there :o) Well done for those that got this far!! I'll follow up very soon with a blog post which really plays around with the bundles and pulls apart the blog sample.

Please let me know how you got on (comment or tweet @sjmaple)

Wednesday, 30 March 2011

WebSphere User Group, Bedfont, UK

Last week was the WebSphere User Group UK event held at Bedfont, UK.  I helped to organise this with a group of other people on the WUG board.  I'm pleased to say it was a great success and the sessions went down very well indeed among the 200-250 attendees

If you'd like to see some of the foils, they will be uploaded here shortly.

This year marked the 10th year of the UK WUG since Alan Chambers created the group back in 2001. We celebrated with an amazing cake, made for us by Cath at  I know she did amazing cakes as she made our wedding cake!

I'm pleased to say that there were a good number of developers from Hursley attending, which allowed us to set up a helpdesk like stand where we had some great discussions with customers and we were able to show live demos on some of our new OSGi functionality in the WAS v8.0 beta.

Feedback from our WUG members was very good. In general the timetable and quality of sessions and speakers was Excellent/Very good on feedback forms. The day ended with a meet the experts panel with many great minds up on stage:  (half the panel, pictured left to right - Andy Piper, Ian Robinson, Alasdair Nottingham, Leigh Williamson)

If you're interested in WAS and haven't been to a WUG event before, I would strongly reccommend you attend.  Meeting people and talking to others using WAS and IBMers between sessions are often as valuable as the sessions themselves! :o)

How OSGi can be dynamic

A new video is now available at the Enterprise OSGi YouTube channel. In this new video, Valentin Mahrwald goes through the effects that removing a bundle in an OSGi environment can have in a dynamic OSGi framework.

The new video can be seen here

Monday, 14 March 2011

The Enterprise OSGi YouTube channel is a big hit!

We've had some great responses from our new Enterprise OSGi channel.  Notably the channel has had over 650 hits, and over 20 subscribers in under a week!  I've still got a number of videos going through various stages of editing, some on WebSphere Application Server, some on the core OSGi technologies.

Why not look at our latest video on the WAS v8.0 Beta, March refresh, which I blogged about earlier today.

One of the responses I got was regarding video contributions from non-IBMers.  I think this would be a great idea to build up an Enterprise OSGi community for us all to share knowledge and information through.  If you would like to contribute to the channel, please get in contact with me, either via this blog or twitter @sjmaple and we can work on new content for the channel.

Going forward, it'd be good to make the channel more technology focused rather than product based. (I'll put up the already recorded and edited WAS videos, which are already in the pipeline).  Videos should be around 5 minutes in length and could include concepts, best practices or experiences of Enterprise OSGi.

Thanks for your support, and I look forward to working with you to make this channel a success!

WebSphere Application Server v8.0 Beta, March refresh

IBM have released a WAS v8.0 Beta refresh which includes some great new OSGi Applications functionality, including: In-place application update, a feature that allows administrators to manage and migrate their OSGi bundles at runtime, without application restart. Application extensions allow administrators to extend their applications at runtime with a Composite Bundle Archive (CBA), again with no application restart. There are also performance monitoring, security, session management, servlet 3.0 support for OSGi applications/components.

To download the WAS v8.0 Beta and get playing with all the new features, visit the download site here ->

Tuesday, 8 March 2011

Introducing the new Enterprise OSGi YouTube channel.

Over the last couple of weeks, I've been putting together a list of OSGi related talks that act as a useful resource for OSGi beginners.  I'm lucky enough to work in a development team full of talented people who are very happy to share their knowledge and experiences to customers and at conferences.  I wanted this youtube channel to provide a web resource so that we can share this information to a wider audience.

The short overview presentations are designed to giev the viewer a high level understanding of concepts without going into too much depth.  I'll put up links and follow up material which goes into more depth at our main OSGi home at

The new channel can be found at

Videos available right now on our new Enterprise OSGi channel include:

Big EARs and OSGi - Part 1, by Ian Robinson, IBM, Distinguished Engineer
Big EARs and OSGi - Part 2, by Ian Robinson, IBM, Distinguished Engineer
An Interview with Zoe Slattery, by Zoe Slattery, IBM, OSGi Technical Evangelist
JPA in OSGi, by Tim Ward, IBM, OSGi Applications developer in WAS.

More will be uploaded soon, so why not subscribe to the channel to get notification of when new videos are uploaded.  To subscribe, you'll need a YouTube account.  Click on the subscribe button on the top left of the Enterprise OSGi channel, shown below.

Thursday, 10 February 2011


Firstly, I feel I should apologise for an awful pun.  If you hadn't guessed, this is a post that looks at why people might want to use and adopt OSGi technology.

Adopting technologies because you can, or because they're the latest or because they seem the flavour of the month often adds complexity to a project for little gain.  A technology should address and fix a problem and justify it's existence.

OSGi has already had a reasonably long and established life, since 1999.  I want to discuss why OSGi is around today and the look at the existing problems it addresses.

In complex software projects a properly modularized system enables:
  • Parallel development of modules by teams who need no understanding of the internals of other modules.
  • Reuse of modules by different apps.
  • Maintenance of 1 module without affecting others.

The Java language goes a long way to solving these problems, however, problems do still exist.

The OO encapsulation model is very focused on instance variables and methods. By restricting access to variables and preferring access via getters and setters, a developer can enforce a usage policy over a classes content. However, this only helps at the class and object level.  Post compilation, classes are usually aggregated into JAR files for deployment into a Java environment, when at runtime, it is considered to be another modular unit. There isn't any granularity in terms of which parts of a JAR can or cannot be used by other JARS . A JAR is either on the classpath and visible, or not. Nothing at runtime is policing modularity/accessibility/visibility between JARs in a similar way to how instance variables and methods are between classes; i.e. there is no meaning or mapping of public and private visibility for JAR files.

JAR files can consist of multiple packages and, if the JAR represents a cohesive function, there is typically a need for classes in one package to access classes in another which then requires public accessibility. An example would be package 'x.y.z' and package 'x.y.z.impl' in the same JAR, an approach often taken to try to isolate the implementation away from it's contractual interface.  However, this is only convention, there are no restrictions here, so immediately, all classes in both packages are visible to classes in any other JAR.

As code and projects grow in size and function there will be an natural interface split, where some interfaces will be considered/declared as external and are to be used by other components/modules, whereas others are indended to be internal and should only be used by internal code. Java has no easy way of controlling the split between API and SPI interfaces and as a result, they are handled by Java in the same way, meaning SPIs are an artificial definition applied by the developer but policed/enforced by nobody, as anyone can use them.

JARs lack the capacity to scope the visibility of what they contain, they also lack the capacity to declare their own fine grain dependencies, such as packages. Many jars have implicit dependencies on other jars that means these jars cannot be installed or moved around independently. If they are installed without dependencies being present then the first time there is any indication of a problem is at runtime when the code is actually invoked. 

The global Java classpath is a flat structure that doesn't allow for classes with common names. If two classes with the same package and class name exist in different JARs the client cannot be sure which one they will get back as the classpath is a flat list and will return the first matching classname simply because it was higher on the global classpath.  A common example of the global Java classpath's inability to accommodate multiple classes with the same name and package structure is when multiple versions of a class/component are present. There can be multiple versions available on the classpath, but again, only the first will ever be loaded.
So, the above list of problems with the existing Java behaviour can all be fixed by features of OSGi.  I will describe how in my next blog post (perhaps 'OSGhow' :o) ), however, some/most/all of these may not be problems in everyone's deployment environment today.  Essentially, I believe that using OSGi allows a developer to better achieve the Parallel development of modules, reuse of modules and the maintenance and management of modules.

Tuesday, 25 January 2011

The London Java Community

The London Java Community (LJC) is a great network of people in London who mix the sharing of knowledge and experiences in a social environment.  I was asked to help out with an Open Conference in November at IBM South Bank, last year and have since offered assistance with the community.

The LJC Associate meeting was held in London to decide what kind of topics we wanted on the agenda over the next few months.  This was a great meeting and we managed to get a bunch of topics lined up for the coming months.  The current agenda is below:

January Events
25th   London Clojure Dojo (full)

February Events
1st              London Python Coding Dojo
2nd - 3rd     Linux & Open Source Expo / Cloud Expo Europe (free entry)
3rd              Gojko Adzic: Breaking the TDD mould
8th              London Ajax User Group Meetup: HTML5 Gaps
9th              Cloud computing with Scala and Gridgain
9th              DevTank
21st            Groovy & Grails meetup
28th            Kanban Vs The Maffia - John Stevenson

18th            Functional programming exchange (£)
24th            Cuke Up

Zoe and I are currently working on an OSGi pitch to give at the LJC over the next few months also.  If you are interested in joining the LJC or attending some of their meets, why not give it a go! More info about the LJC can be found here.

Wednesday, 12 January 2011


Firstly, an introduction!  My name is Simon Maple and I work for IBM's development labs in Hursley Park, Winchester.  I joined IBM 10 years ago in a test role for WebSphere Application Server (WAS) v5.0.  Since then, I moved into development where I worked on many areas for many releases of WAS, including the Transaction Service, Compensation Service and most recently the OSGi feature pack which was delivered on top of WAS v7.0.

I've recently split my role to undertake my existing development role and an OSGi technical evangelist role, in equal share.  I will be taking on some of the work Zoe Slattery is currently doing when she retires from IBM in April.

I'm sure my opinion of what a good technical evangelist does/should do will change with time/experience in the role.  However, my initial vision is that an evangelist should enjoy working with and believe in a technology so much that they want to increase awareness to generate support for the technology.

OK, let's assume that interpretation is correct for now... How can I achieve that?  Well, in IBM we have a wealth of technical minds and talent.  It's important not to lock that talent up in the labs, rather unleash it onto a number of mediums, such as conferences, blogging, workshops, tweeting, articles, sample code, webinars etc.  I interpret my role as one that helps build the support and following of a technology by increasing awareness via the described mediums or by networking to join up technical resources in the lab to talk and share at conferences etc.

I'll be updating this blog with some of my work and also interesting things I see on my way :o)