Hi!

What you want to do is called a replication.

Is your application database is VFP database or on SQL Serevr? If on the SQL Server, you can easy organize this using replication on teh SQL Serevr. With VFP database you need to organize replication manually (I did not seen 3-d party tolls for data replication for VFP databases yet).

Yes, replication requires a lot of effort:
1. Make an automation programs that run on background on each server and run replication routine at the scheduled time. It also should have an ability to lock database from any updates (or even close all client applications) for the duration of replication, so better schedule replication for night hours.
2. Make a replication publishing program that will publich all changes in the database sinse last replication successful running.
3. Make a server replication routine that will get all these changes and apply to the main
database.

There are many other kinds of replication supported by SQL Server, I just concentrated around your problem.

As about performance, it is very depended on the way you're going to use to get data from branches. the most reliable way, IMHO, is ftp transfer - install on the main server and send data from branches to the main server when replicating. Main server than will make updates quickly taking data from the local fpt directory. Note that 'get' data here is just a word, really you pull data from branches through fpt.

Another approach is to use DCOM through internet, but this requires more effort. With DCOM you can make updates on the server just from branches.

the main influense to performance will have an replication algorithm and its ability to separate only changes from database to send. It also depended on how much changes made to database by users during a period between replication running.

Note also that with large number of branches you need also to organize schedule very well for traffic reasons. Imagine what will be if 30 branches will try to upload data simultaneously through fpt to the server!

Just a considerations from my experiense...

>Hi,
> I am developing Multi-user POS/Inventory Application. I faced a problem, that my client hope the application to able to be run in multi-location (branches).
> After couples of discussion among our developers, we decide to recommend my client to manage their sales and stock individually by branch. Eveery branch will have their database and manage/update on the branch's server. Every record of each table will have a unique branch code. Is it acceptable?
> I have a problem, to generate reports which will reporting stock, and sales for more than one branches at the same time. E.g, my client wish to view all of the summary sales transactions which group by branch within a SINGLE report in Headquater.
> How could I send the branchs' transactions to headquater? It may be expensive and time comsuming if request every branches to send their updated data every days or period of time. Any suggestions ?
>
>Pls advise!
>
>Thank you