Remote Data Update! Any Idea?

[EzLogic]

Hello people!
here is the scenario:
I have an application (POS Application) running in 2 different sites (cities).
Application A (The Warehouse)
Application B (The Retail)

Application "A" needs to transfer some records (Items) to Application "B"

What is the best way and how can one accomplish this

I thought of this:

I developed an FTP upload/Download and I built it into the Application "A" (The WareHouse)
The user selects the Agent (Which I will know the IP address and the server for that agent) and then Scans and Enters all the Items he needs, he hits Transfer/Save Button and in which it will upload the table (Temp Table that I will create) to the Host where Application "B" (The Retail)resides.

On the server where application "B" is on, there is a small function running and watching for any file that comes
(Received Directory), and once the file is there, the function will take all the Items from that table and appends it into Application "B" and stamps each record as done, so, the app will not append it again.

Is that a good idea? Is there a better way? to access the tables of the other application if I know the 'Static IP' address of Server "B"?

I am running Dedicated Lines 256K and some sites I have DSL 768K Down/128K Down?

What do you think?
Please let me know.
Thank you in Advance Ali Koumaiha -

 

[hdc]

Ali Koumaiha,
That is exactly the things I am working for one of my client. At last I choose to use a MySQL server to handle the data (i.e. A is MySQL, B is VFP). These are some of my considerations.
1) Application B does not know the time data arrives. Therefore it must scan the hard disk again and again. If this is supposed to be done 24 hours a day, the hard disk may break down very soon. If you know something about programming i/o port, this can be avoided. However this is a very tough topic.
2) Usually a POS system just need to get 1 or 2 kilobytes of data to process a transaction. A 256K DL (my client only gives me a 56K telephone line, too bad) should be fast enough for a SQL query.
3) Centralizing data is much easier to maintain.

Hope this can help.

Hetzer

 

[EzLogic]

We Have centralized data now, but, My company wants to go into franshizing business, so, each store will be owned independatly. Therefore, they might have only DSL lines.

Actually. Host A. Will only scan for a file only when A request a PO Number. Then the function will be triggered and It will scan
if file('....')
do updatestuff
endif
every 1 hour from (store hours defined in a table)
what do you think? Ali Koumaiha -

 

[hdc]

Ali Koumaiha,
As update is needed only when PO comes, your method will be feasible. Actually you are setting up a supplementary database so that update can be made from the warehouse to the outlet with limited data flow. That's a good idea if the update is not frequent (frequent=once per minute) and the update does not contain a lot of bytes.
If you will consider to upgrade later so that communication is bidirectional (A->B and B->A), you may need to reconsider your algorithm.

Hetzer

 

[EzLogic]

Hetzer,

Thanx for your reply. I have the following:
1 warehouse distributes to 20+ locations:

So the update will be from the warehouse to any client 1 per week only. Not alot. Maybe 3 times a week maximum.

And it is not just A->B only. It is A->(Anylocation) but, not daily. And once per po only! Ali Koumaiha -