axis-java-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Sagara Gunathunga <>
Subject Re: Axis2-Spring integration
Date Wed, 31 Mar 2010 19:32:50 GMT
Hi all,

I think we all can agree about the usability of having good Axis2-spring
integration module, let me provide some background about Axis2M.

Around 1.5 years ago I had to find a good WS framework for a Spring based
project so i tried to introduce Axis2  but some people pointed out there are
other WS stack that provides 1st class Spring support  than Axis2 but even
at that time I realized it is possible to provide 1st class Spring support
by developing a good supporting module.

In the same time , as I mentioned in Axis2M guide [1] I tried for WSF/Spring
module but in a Spring based developer perspective  it has some good
features also some weaknesses as well. As i mentioned ,Axis2M started taking
only good features from WSF/Spring. As an example Axis2M's Service
deployment architecture based on WSF/Spring but Axis2M added more
improvements to it such as Spring Namespace support , ability to define
service beans as top level beans as other WS stack does , Spring based data
binding etc.

Also Axis2M introduce completely new set of features like Spring based
Module development,Namespace support , POJO based Module development, Spring
OXM support ( this will enable to support two new data-binding as Castor and
XStream) Also provide number of Maven archetypes too [3].

About the axis2.xml configuration , I have clearly answered in my post [2]
about the Andreas's review , it was not going back , it simply current
Axis2M version does not provide any Spring based configuration option, only
for current version it depend on default Axis2.xml file. but I have mention
my future design to handle Axis2 configuration and WSDL generation for next
version here [2].

Exposing a Spring bean that contains business methods as a web services is
one possible use case but lot of Spring developers expect much more features
, If I take a different but good example is Quartz  scheduler Quartz/Spring
integration allows to configure and perform all the Quartz  functionalities
within the Spring context . Spring devs expect same level of support from WS
frameworks too and some of other WS frameworks already offering  such level
of support. And clearly it is possible to provide same level of support with
Axis2 if we can come up with good integration module.

I'm developing this Axis2M module just to share my experience with others
and also i don't like hear that Axis2 doesn't provide good Spring support :)

I don't care about location of this code whether it integrate into
Axis2-Spring module (IMO that is the best option ) or  in a different place.
Also I'm not suggesting to start with my codes but better to start with
correct and uptodate codes .If you guys don't like I can keep my codes on

As the developer of Axis2M I alleyways thankful to WSF/Spring team because
they are real pioneers of this effort and then I took best of it, shaped it
and continue..

I like to participate to this discussion further if we can have it in a
objective manner without being subjective.

[1] -
[2] -
[3] -

Thanks ,

On Thu, Apr 1, 2010 at 12:31 AM, Andreas Veithen

> On Wed, Mar 31, 2010 at 19:59, Amila Suriarachchi
> <> wrote:
> >
> >
> > On Wed, Mar 31, 2010 at 1:40 PM, Andreas Veithen <
> > wrote:
> >>
> >> All,
> >>
> >> Recently there have been some questions and discussions about Spring
> >> support in Axis2 [1] [2] [3] [4]. As mentioned in these references,
> >> WSO2 has developed a framework that provides Axis2-Spring integration
> >> [5]. Also, Sagara Gunathunga has done some very interesting work [6],
> >> but this has not yet been released. In [1] I argue that neither of
> >> these two frameworks provide a definite solution and Sagara actually
> >> agrees with my analysis.
> >
> > I also agree with your analysis :) but let me tell something about it as
> one
> > of the initial designer of wsf/spring.
> >
> > As I learned about the spring (two years ago) it promotes a way to write
> the
> > business logic as POJOs and managed them using spring framework. Spring
> > provides functionalities to let these business methods to access data
> layer
> > and expose them as web applications. But you can test the business logic
> > without being depending on the data layer or web layer as they are
> written
> > as POJOs.
> >
> > The idea of the wsf/spring was to let people expose these POJOs as web
> > services. For this it uses RPCMessage receiver, schema generator and some
> > utility methods of axis2-spring. So some of the new features you have
> > mentioned was not considered by that time as well.
> >
> >
> >>
> >> On my side, after writing my review and in
> >> reaction to [3], I wrote some PoC code to explore how some of my wish
> >> list items could be implemented. This code is available at [7]. I
> >> would also like to mention the interesting contributions made by
> >> Stephan van Hugten, who opened [3]. I would really like to see him
> >> continuing to contribute.
> >>
> >> After discussion with Sagara, we decided to join our efforts under the
> >> umbrella of the Axis2 project.
> >
> > Can we do this? According to apache rules I think it has to start an
> > incubator project. I am +1 on starting as axis project :)
> >
> > For me the ideal place is in the spring code base itself. As I remember
> this
> > is where the cxf/xfire spring integrations lies. Since this is a feature
> for
> > spring users it would give a better visibility to axis2 as well.
> The intention is indeed to make this a module of Axis2, so that an
> up-to-date version is part of every upcoming Axis2 release.
> >
> >>
> >> Probably what you will see happening in
> >> the next days is that Sagara will bring the existing Axis2M Spring
> >> code into a sandbox area in Axis2 [8] and that we will start a
> >> discussion about how to best merge our two codebases (which for the
> >> moment address different parts of the problem) into a single one.
> >
> > Although wsf/spring has short commings that you have pointed out, it
> > describes a better possible way of integrating Axis2 into a spring
> container
> > (I think this could be the reason for forking Axis2M from wsf/spring as
> > mentioned in the your blog). And also it supports ws security as well.
> > Therefore it would be easy to start from the wsf/spring and gradually
> > improve it to suite for new requirement would be an easy approach.
> >
> > So would like to propose to start with wsf/spring. Further Axis2M has
> forked
> > from wsf/spring and it has go bak to use axis2.xml as well.
> If everybody proposes his own code as a starting point, we will get
> nowhere. I think we should start from scratch, and then as we progress
> through the different areas we want to cover, take over those
> components from the three codebases that fit well into the
> architecture, discarding those that need to be redesigned. Anyway I
> think that both WSF/Spring and Axis2M share an architectural flaw that
> will make it difficult to support the standalone (non servlet) case
> and Spring at the client side. I will provide an analysis of this
> issue later.
> We should also establish a list of the requirements and goals that we
> want to achieve, so that everybody has the same vision of where we go.
> Probably a Wiki is a good place to write this down. I think, a new
> Wiki for the Axis project has been created recently.
> >> I think that everybody agrees that it will be beneficial for Axis2 to
> >> provide a good Spring integration out of the box. Of course, this is
> >> also the right moment to look for involvement of others in the
> >> community (developers and end users). I you want to help shaping the
> >> solution, please participate in discussions, provide scenarios that
> >> you would like to see supported, make suggestions, contribute code or
> >> simply test the code.
> >
> > From the given features list I really like the idea of servlet + jaxws +
> > spring. As I mentioned earlier wsf/spring uses RPC message receiver. But
> > this has some issues with performance and support lists maps etc. So if
> we
> > can solve these issues with jaxws message receiver or a new message
> receiver
> > based on jaxb while supporing jaxws it would be a really advantage of the
> > user point of view. IMHO for a normal user the performance and the
> > complexity level of the POJOs that we can support is also very important.
> I think that once the infrastructure is correctly in place, supporting
> JAX-WS doesn't require any new stuff, just plumbing code. Everything
> is already in place in the existing JAX-WS deployer. But it is
> probably too early to start discussing this question.
> > thanks,
> > Amila.
> >>
> >> Andreas
> >>
> >> [1]
> >>
> >> [2]
> >> [3]
> >> [4]
> >> [5]
> >> [6]
> >> [7]
> >>
> >> [8] Note that Sagara was part of the bunch of committers we recently
> >> voted into the new Axis TLP.
> >>
> >> ---------------------------------------------------------------------
> >> To unsubscribe, e-mail:
> >> For additional commands, e-mail:
> >>
> >
> >
> >
> > --
> > Amila Suriarachchi
> > WSO2 Inc.
> > blog:
> >
> ---------------------------------------------------------------------
> To unsubscribe, e-mail:
> For additional commands, e-mail:

Sagara Gunathunga

Blog -
Web -

View raw message