Thanks for the feed back,
No decision has been made regarding PXE. ActiveBPEL has been chosen for POC
because we need to choose something and PXE did not come to our attention early
enough to be considered. A brief look
indicated some challenges but I don't have the
details.
As to ServiceBus, you make a compelling case. My only
comment here is that ALF should only require what it requires. It seem
reasonable at this point that ALF can remain BPEL engine agnostic and I think it
is better for us to maintain that
approach so long as it is viable. While
ServiceBus provides BPEL
functionality it packages a number of
other technologies. If ALF requires
ServiceBus, particularly at an early stage, it will be much harder for ALF to remain BPEL engine
agnostic because of the additional assumptions that will be created none of which are core to
ALF.
That said, if
there is sufficient momentum within eclipse around an SOA technology suite that
includes PXE then it is inevitable that some configuration of ALF will work with
it.
Tim
Another addition to Rupert's comments - ServiceMix is actively
engaged with the Eclipse SOA project (STP), which is targeting JBI as it's initial deployment platform. Leveraging the work of STP within ALF will be much
easier if we commit to a common infrastructure.
Perhaps for the POC it's a bit late to change tack, but for the
subsequent deliverables we should seriously consider aligning our deployment
target with STP's.
Cheers,
Joel Hawkins
thank you for your comments. i did not suggest pxe, but
servicebus, which includes pxe. servicebus develops rather quick, and at that
time when you made your decision, pxe was not yet included. maybe i should
also say why i think servicemix would be a good
choice:
technical:
* contrary to activebpel already
supports both ws-bpel 2.0 and
bpel4ws1.1
* additionally rule based
routing
* additionally
scripting (a very important interfacing issue)
* java business
integrations allowing different vendors to "plug and play"
* integrated in
application server or standalone mode
community:
* by far a larger developer community and
community support than activebpel
commercial background:
* integrated in geronimo (ibm aquired geronimo development company)
* can be integrated in jboss
* ibm supports eclipse
actively and it can be expected it goes a
similar
direction with
geronimo.
summary:
* further technological and
stability leadership of servicemix can therefore be
expected
* imo not choosing servicebus means
implementing/integrating similar technologies
to activebpel which needs resources and
time. even for a poc.
could you detail what servicebus or the included pxe is
missing compared to activebpel? or somehow publish your results ... i cannot
remember you mentioned your choice and the reasons on the mailing
list.
kind regards,
rupert thurner
Thank you for posting this analysis. It is
very useful. I agree with your evaluation.
Just to be clear, Agila was formerly our preferred
choice before we had investigated it fully.
However it quickly became clear that is incomplete is some important
ways. It may be our primary choice in the futured depending
how it develops. Currently the BPEL part (Twister) is being merged with
Agila to become a full apache project so it may be a while before it settle in
and work in the BPEL part results. We looked at a number of Open Source
engines and only ActiveBPEL seemed to be
mature enough for our immediate needs. Consequently, ActiveBPEL was chosen as the
target for the POC. I believe this was discussed on the newsgroup
or in the mail list but I apologies if that was not made clear in
an obvious way.
PXE is on the
list as an example implementation in the future. We have started to look
at it internally but it will not be the target for the ALF
POC. I recommend staying with ActiveBPEL for ALF
related work for the time being.
That all said,
it is a goal of ALF to be BPEL engine agnostic as far as is
practical. This is not the focus of the POC and thus any current
ALF work but it will be useful, going forward, to understand the various
engines out there.
Tim Buss
+1
hi robert,
maybe you want to look into http://servicemix.org/. it should be much
closer to what you want to do. it uses opensource pxe.
best regards,
rupert thurner
The contents of this e-mail are intended for the
named addressee only. It contains information that may be confidential. Unless
you are the named addressee or an authorized designee, you may not copy or use
it, or disclose it to anyone else. If you received it in error please notify us
immediately and then destroy it.
**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. Any unauthorized review, use, disclosure or distribution is prohibited. If you are not the intended recipient, please contact the sender by reply e-mail and destroy all copies of the original message.
|