Hi All,

General observations & questions really.

Has anyone experimented with SSDs and VFP tables?

I have a biggish system which is starting to strain a bit.
Some general info:
15 simultaneous users
4 gb total across all tables & indexes.

Just as an FYI
All 15 users need to run a big job on the same day, which was really hitting the network and server hard - this job was taking up to 3 hours to complete (for each user).
As part of my analyses, I found the indexes were being read 30 or 40 times for each .dbf read (when this was multiplied up to 15 users and thousands of dbf reads it became crazy).

Therefore, I had the idea of taking a local filtered copy of certain indexes before starting any of the big processes (index on xx to yy for zz)
This reduced these 3 hour job down to 10 minutes for all 15 users - Bliss!
Network and server contention was dramatically reduced.

Next steps are to move the DB to a dedicated file server (at the mo it's on a shared fileserver with 200+ normal fileserver users).

I was also thinking that it may be a good idea to put in some SSD's on the new server. I understand that drive seek latency is a big slowdown in physical drives. Could this make a difference?

Another related question: Does anyone know of a windows / raid solution to host a .cdx file on a different drive than the dbf while it is still in the same folder? I guess this could help too.

That's it for my performance ramblings for the moment!

Cheers
Will