notes on efficiency

diff --git a/raddb/mods-available/dpsk b/raddb/mods-available/dpsk
index d487f0a63d2719bd4835ba267876e4ef8e93dd71..3cd8411403b3ddd5ad6c65186cae9f91b847bd78 100644 (file)
--- a/raddb/mods-available/dpsk
+++ b/raddb/mods-available/dpsk
@@ -89,7 +89,16 @@ dpsk {
        #  double quoted: "psk".
        #
        #  The mac field is optional.  If it exists, then that PSK
-       #  will be used.
+       #  will be used.  It is highly recommended that the MAC *not* be placed
+       #  into the CSV file.  Instead, the MAC and PSK should be placed into a
+       #  database.  The server can then be configured to look up the MAC in the
+       #  database, which returns the PSK.  That way this module will only ever
+       #  check one PSK, which is fast.
+       #
+       #  i.e. the CSV file should only contain the small number of PSKs where
+       #  you do not yet know the MAC.  As soon as you know the MAC, you should
+       #  put the MAC and PSK into a database, and then remove the MAC and PSK
+       #  from the CSV file.
        #
        #  NOTE: the file is opened and read from top to bottom for every
        #  new request which comes in.  This process can be very slow!