Myth: ODBMSs Are Slow
The myth that ODBMSs are slow has no basis in reality. ODBMSs have always excelled at high performance on complex data. It is not unusual to see private benchmarks that show ODBMSs running anywhere from 10 to 1000 times faster than an RDBMS when accessing complex data. See complex data.
I have been able to think of two possible sources of this myth:
- Some people who have a strong background in relational modeling often, at first, have trouble designing an object model. It is sometimes quite hard to not use intersection entities and type codes that are used so often in relational models but can nearly be eliminated in object models. So, a poorly design object model has the potential of making an ODBMS run slowly.
- Some people believe ODBC stands for Object Database Connectivity. (I have seen press releases that use ODBC this way!) Well, ODBC stands for Open Database Connectivity and it does have a reputation for being slow. So, if you think ODBC refers to object databases, you could understandably have the wrong idea that ODBMSs are slow.
Don't just take my word for it, listen to what Dr. Michael Stonebraker said about object database performance at the Object/Relational Summit in August 1997. At the time, Dr. Stonebraker was the Chief Technology Officer at Informix Corporation. In this audio clip he mentions ODBMS performance with C++. Java had not as yet come on the scene. Audio clip: Stonebraker.mp3.
More on the general topic: ODBMS Myths
- Myth: ODBMSs Have a Steep Learning Curve
- Myth: ODBMSs Do Not Support Queries
- Myth: ODBMS Products Do Not Scale