Your welcome ! I'm always interested in design
problems (opportunities ...). Thanks for the
extra amount on information.
Maybe the best approach would be to have the
three systems installed and some heavy tests
done.
>Our manufacturing enviroment demands a high degree of tracebility, therefore we must maintain large amounts of historical data. Our current database size for the Item Master is approaching 2 million records. If we explode our Bill of Materials to the level of 150 parts per Work Order, we could be approaching 1.5 to 2 million records per year.
>
How critical is this application for your
business ? To me it looks like mission
critical !
>Currently, all historical data is maintained in one master file for ease of accesss. It's kind'va shotgun approach, but purging historical data offline would seem to be cumbersome.
>
It's more then a shotgun approach. I hope the
new system have a better database design.
>I've read the general Foxpro reviews on the Chunnel project and the database design seems to handle huge volumes, at what level of maintenance and what degree the software logic is written in seems to be the crux of the issue.
>
You're right. This is the crux. Other systems
handle a lot of this logic internally. I don't
want to suggest you any product, and I'm a VFP
fan, but if it's mission critical, with a very
big amount of data, it think a real database
server would be a better solution.
Another issue: if you plan to use the application
over a network (the dbfs on a server), the the local
vfp client may load in the local memory the accesed
table, not only the result of a selection (like a real
c/s system - fox as client, oracle as server). With
very big tables and indexes it's not really performant.
There are some ways to solve this problem, but again
over coding, design and so on.
Wish you a nice selection ...
Vlad-Georg