There are several points:
* independant: will the replication between the same systems or cross-systems (ie can the replication always be between the same RDMBS types (A->A, B->B), although supporting different types, or replication from type A to type B)?
Note that for A->A you have already everything in place, why try to write something as complex as replication to be independant, and finally BAD in performance.
For A->B, you will FAIL miserably as each RDBMS has it's own specifics, with more and more for each version. You will end up upgrading your application to take care of all the possiblities. Leaving items out will make your tool be thrown to the dustbin quite fast...
Now, the above is my opinion, and I know that a true DBA will never be able to trust a tool that pretents to replicate between ANY rdbms type to another one, and won't even give it a try as the own one works fine...
CHeers
Main Topics
Browse All Topics





by: crescendoPosted on 2003-06-10 at 15:01:57ID: 8694864
Independent? Is this theoretical stuff? Why re-invent the wheel when every major RDBMS has replication built-in, in several flavours, with GUI interfaces?