I am not aware of any performance differences.

There has been a few comments that performance can be hurt if the path name
is very deep and long.

However, you are MUCH better to use a share name, as drive letters can
change, have to be setup in the first place, and are generally a pain in the
neck.

Drive letters go back 20 years ago to dos and cpm days...you would do well
to avoid them.....

Once you have a connection, it will be cached by your local
network client.

People sometimes set their pc's to connect to mapped drives
at startup. If you have done that, then the connection will
already be cached.

Depending on how your shares are set up, the cached connection
to a share may time out when it is not being used.

In contrast, mapped drives get refreshed all the time, one
way or another. (Sometimes you can see this slowing down
work you are doing at the computer)

But mapped drives are likely to be 'faster' only if your network
is really badly managed and your application is incorrectly
designed.

If the time taken to make a connection is bigger than the
time taken to do security checks, lock data, and retrieve,
then your network is not set up correctly.

A stupid 'optimisation' to the Jet 3.0 engine,  means that there
is a lot of overhead in connecting to a Jet database. For this
reason, you need to maintain a persistent connection to your
database. If you are maintaining a persistent data connection,
your cached network connection will never be lost, and you
only ever have to find the share at application start up.

An advantage of mapped drives is that they can be redirected
to other shares.

An advantage of shares is that they don't get re-directed when
the mapped drives change.

For these reasons, mapped drives are often the choice of formal,
very large organisations.

And network shares are often a better choice for informal, small
or medium organisations.

(david)