Hello databasers,

I'm working on a reasearch project dealing with an experiment about a
"fault-free" database system. The main idea is to have a couple (say 3)
RDBMS, so if one fails, the rest won't. The biggest deal is that the RDBMS'
are from different manufacturers, so eventualy they won't crash at the same
time under the same circumstances, so the system will continue to work. This
is what our theoretical team is calling "software diversity". Our goal is to
prove it works.

In order to do our experiment, we actually need to exploit BUGS in the
different RDBMS'. If we know that a certain situation or SQL query will
crash the RDBMS, we want to use it and see if the other servers will crash
too.

SO, I'm appealing to all of you, database gurus and everybody else, if you
know about any bugs in any of the following RDBMS, to mail a description of
the problem to me:

We're using:

1. Interbase 5.5 (!)
2. Oracle 8.05
3. Microsoft SQL server

The above servers will run on all the platforms they could be ported to -
Unix/Solaris/Linux/Windows 95-98/Windows NT/Windows 2000

If any of you could help me with any links to something like "known bugs",
it'll be highly appreciated.

I'm especially looking for bugs that will actually CRASH the whole server,
or will produce incorrect results, or will damage the database or in general
will make the server or parts of the server unusable or unoperateable.

I'm interested in the limitations of any of the above mentioned servers,
like number of users, number of records, database size, etc. For example if
anybody knows if Interbase would crash if 500 users connect to it
simultaneously, please let me know.

We're working on an Internet gateway to our experiment. As soon as it's
ready, EVERYBODY is invited to try to crash our system. Any help will be
greatly appreciated.

Anyone who's wishing to participate or would like to learn more is welcome.
This is not a comercial product, but a real university research.

Best regards,
Dimitar Selensky

selensky@...