Re: Relationals vs. Objects Databases I

On Mon, 2 Feb 1998, Andrew McClure wrote:

> Date: Mon, 02 Feb 1998 22:31:06 +1300
> From: Andrew McClure <andrew@!amac.com.au>
> Newsgroups: cern.oracle, comp.databases.object, comp.databases.objects,
> comp.databases.theory
> Subject: Re: Relationals vs. Objects Databases I
>
> > >As I suggested before, I think it comes down to the type of system
> > >you're developing. I'm working on the population of a sizable data
> > >warehouse at the moment. We use COBOL to embed our SQL. That's not a
> > >big deal. There would be little benefit in moving to, say, C++ to embed
> > >SQL in - the structure of the programs is so simple that I find it hard
> > >to see how OO programming techniques could add value.
>
> I have been following this thread with some interest, and being relatively
> new to Object Databases, I won't fly into the current debate, however -
> I have one simple question. If one was to take a variety of running
> RDBMS applications, and port them to an ODBMS on similar hardware -
> Is there a rule of thumb that would predicate the performance
> difference, would it favour ODBMS, and if so, by how much.

This may give you a clue:

S. Baker and M.A. Salman (1991) Performance comparison between an object-oriented, a relational and an object management database. Proceedings of Advanced Information Systems, London, UK, 1991, pp. 61-67.

Caveat: It is old, but is a "classic" Telecoms Network problem.

On the flip side, see this:

A.B. Larsen (1992) A test evaluation procedure for object oriented and relational database management systems. Master's Thesis, Institute of Informatics, University of Oslo, Norway, 5 February 1992.

which shows that OODBs can actually do worse than RDBs for certain applications, when the working set gets large.

Also:

U. Hohenstein, V. Plesser and R. Heller (1997) Evaluating the performance of object-oriented database systems by means of a concrete application. DEXA '97 Workshop on Object-Oriented Database Systems, Toulouse Cedex, France, September 1997.

is worth looking at, since it compares 3 "major" OODBs for a data warehousing-type application and only one product does well enough when compared to an RDB.

>
> Further to this, if one did the same for RDBMS systems enhanced with,
> and using Object extensions, would these results differ dramatically?

For my research, one product I used did badly at everything. It's only saving grace was the bulk loader, which was able to get large quantities of simple data into the database very quickly. However, once data complexity began to get high (i.e. a deeper inheritance hierarchy and number of interconnections betwen objects), I gave-up, since generating the input data set was just too hard.

>
> Thanks in advance
>
> Andrew
> ____________________________________
> Andrew McClure <andrew_at_amac.com.au>
> Software Engineer - Wellington NZ
> Phone +64 479 2002
> _____________________________________
>
>
>
>

• Akmal

        akmal(at)bigfoot(dot)com Received on Fri Feb 06 1998 - 00:00:00 CST