[EMAIL PROTECTED] wrote:
> Hi Stefan,
>
> I'd like to add to this (as I have been working in a DWH context the
> last half year): I am very interested (just have not had the time
> yet) in the question of how SOA based integration can be usefully
> applied to an environment where data storage is so dominant that
> interfacing with other applications is primarily done through the ETL
> (extract-transform-load) processes and the DWH.
>
> In a sense this is 'unidirectional' integration, as information just
> flows from the integrated systems to some consuming system. The DWH
> (and the DWH's physical model) become the central integration spot
> (with all the coupling consequences).
>
> The question is not if that approach is good or not (since that is
> just the way things are and will be done) - the question is how SOA

Well, maybe but I think actually it does matter whether the approach is
good or not.

Or more specifically, is this approach "SOA friendly" or not (i.e. good
for SOA).  If it isn't you can't use SOA.

And if you can't use SOA with that approach does that mean we should
bend our SOA to fit or should we change our approach?  And if we bend
our SOA to fit, are we SOA any more?

> concepts can be usefully applied or how SOA ideas can be explained to
> people bound to such an environment?
>

Dunno, maybe you have to show them how an SOA-based version of the
system would be better than what they do now?  Perhaps you have to
demonstrate that it'll make their jobs easier - maybe they end up with
fewer custom reports to develop or something?

Dan.





SPONSORED LINKS
Computer software Computer aided design software Computer job
Soa Service-oriented architecture


YAHOO! GROUPS LINKS




Reply via email to