A client of ours has n medical clinics, C1 to Cn.
The clinics are not interconnected. Might be as well that they're in different countries.
Every clinic is running the same version of our software. 1 Server, many clients.
The software is (simplified) running a 3-level deep database structure, like
L0 : Patient 1
L1 : Clinical finding n 1
L2 : Exams n
Every clinic is at 90% autonomous, but in 10% special cases a longtime patient of clinic Cx
must go to clinic Cy because it's the only one specialised on his current problem.
What the client wants:
- a Transporter database, which
1) gets cloned
3) current clinic is selected from a list (we know the local IP addresses of all clinics' servers)
4) first name, family name and birth date of the patient is entered
5) the data (2 levels deep) is imported into the transporter (we assume a single record L0 find)
6) Transporter database is burned on CD and given to patient
7) patient goes into different clinic
8) transporter database is copied from CD on a local workstation and opened
9) destination clinic is selected from same list as 3)
10) one click and the patient data is uploaded to current server and the local doctors access it.
11) changes are made and the process starts again - data goes back to clinic Cx, updates patient record.
What is in your opinion the best way to achieve all this, given that you can't define TO's having variable IP source addresses ?