Re: How to use the SAME transaction in a distributed environment

allond_at_fts-soft.com wrote:
> Oracle replication is not a good solution since 'A' and 'B' are
> different applications, with completely different tables and
> data-structures.
> I can make them run on the same instance, and even the same schema if
> so needed.
> Please see some more details in my posting below.
>
> Thanks,
>
> Allon.
>

Fail to see the argument; I would say on the contrary:

- App A (your Front-end) commits (rollbacks don't save data...)
- Replication awakes (9i R2 and after: on commit, perhaps?)
- Data in All-Important-Table-of-A is replicated to same

   table in B.
- All-Important-Table-of-B has triggers, calling stored procedures,

   that will handle processing of newly arrived data in table    All-Important-Table-of-B, ending with the flagging of data in    All-Important-Table-of-B as being processed. Commit this.

What's wrong with this? Pretty easy to implement (you wanted it today, your management called: they want it yesterday :-)), pretty easy to maintain (replication is not very complicated, there *are* safe points (flagged as processed)), can run in the same instance, but also between instances.

And even so, did you think about InterConnect?

-- 
Regards,
Frank van Bortel