Most of the conferences have a couple of things they want you to take home. Although I understand that last year was already about SCA and policies, it's still pounding here now. And of course Beehive... Well BEA had Beehive as their "making life easy", "Spring like" framework, but is brand new! Or isn't it...
Well actually Oracle Beehive is actually an awful lot like the Liquid Enterprise BEA revealed at BEA World in 2007. Then again, that's one of the reasons why they bought BEA, so no wonder!
SCA is hot at Oracle OpenWorld as well. They are pushing mainly the Weblogic Application Server platform and the BPEL / BPM products forward as their SCA enabling platform, ready for the agile Enterprise future. I'm not so sure yet about the real advantage of SCA over plain old BPM. So mainly what the added value is above having business processes being defined is that you have a good notion of what is in your composition (which services), what the interfaces (dependencies) to other components are and what the properties for the component are. You still have the point (as with BPM) that services are deployed over several systems and you'll have a deployed composition of the services on another system. A good thing is that you're not restricted to having Web Services, with SCA a composition might also contain Java classes or .Net classes. Which actually could be done with BPM using for instance the Oracle Service Bus! Nevertheless SCA is a nice clean way of composing Service based applications and since Oracle is pushing it might be here to stay.
It seems however it is mainly a technology thing, which might not make it up to deciding business level.
The scoop of the conference was actually presented on Wednesday by Larry Ellison (CEO of Oracle) in what you can best summarize as a sales pitch or commercial of about 30 minutes of the new HP - Oracle Exadata Data server.
Cool server technology but I did not like the presentation actually. Rather then setting out the new Oracle way forward he spent the whole presentation, which looked a bit unprofessional("next slide please...") on how good the relationship with HP is (elephants do love elephants) and what a perfect hardware / software combination they created for handling bigger and bigger query data throughput demands. Impressive, but I didn't like the keynote at all!
Just to be complete a little info on the Exadata servers technology. The server consists of several high performing Disks combined with Intel processors for running queries on the data. This together with a high bandwidth data pipe for sending out data puts together the Exadata server. The real trick with the server is that queries are performed on the same server that hosts the data (not on a separate server) and only the result data is passed to the database server that requested the query.
This can be a good thing for some of the Oracle customers, but it will be a limited set. The other thing is: how about inserts, updates and delete? These servers are mainly tuned for high performance on querying, not for those. Maybe next year...