AndrewP
Unregistered
|
|
Reading your article I noticed an interesting contradiction. You strive for a simpler model, at the same time talking about recovery and self-healing. Hmm....
Anyway, life is not black & white. If you check out Mule (http://mule.codehaus.org), it mostly falls into level 3 of your gradation, at the same time its complexity is well below the level expected.
Still, an article was a word of reason in all this SOA buzz-ocean.
|
hsheil
Unregistered
|
|
Hi
hopefully what came across in the article was my proposal that we have a SOA taxonomy, ranging from a very simple implementation that assumes a lot of things about your application right up to the very complex SOA implementation / enabler that assumes nothing and hence has to provide a lot more features - and it's at the medium or complex layer that I see a real need for retry and error queues. My points about recovery and self-healing are restricted to how they are good features in a Jini or Grid Computing environment but don't have the same applicability in a JEE environment.
Hope this makes sense.
Regards
Humphrey
|
|
Previous
Index
Next
Threaded
Edit
Reply
Quote
