I often hear people say that MySQL is "slow". Which often translates to sub-optimized or under-powered. Consider the following real-world example which shows that MySQL works fine with enormous data sets (Yes that almost 1 billion rows in a single table having 41 columns, 1 primary key, 1 unique key, 9 non-unique keys):

mysql> select count(*) from t_prod;
+-----------+
| count(*)  |
+-----------+
| 952174654 |
+-----------+
1 row in set (0.00 sec)

# ls -lh /var/lib/mysql/prod | grep t_prod
-rw-rw----  1 mysql mysql   11K Jun  4  2007 t_prod.frm
-rw-rw----  1 mysql mysql  222G Apr 18 07:35 t_prod.MYD
-rw-rw----  1 mysql mysql  102G Apr 18 07:35 t_prod.MYI
[download]

And I can tell up operations on that table are very fast as long as the indices are used properly.

Maintenance, however, is a nightmare. For example:

• When there's a power outage or voltage flux and that 222G table's indices become corrupt and needs to be rebuilt. During which the table is locked and inaccessible.
• When one needs to reduce the data set and the delete operation takes hours. During which the table is locked and inaccessible. (Of course, this can be mitigated by using partition tables).