Hello Daniel,
thats exactly what we like to do with our new csp-based-Application.

Denver has mentioned some important security benefits. In addition you
can also backup the data from each customer seperately and you can also move a
customer-db from one to another server, etc.

I think this is a very recommend structure.

For our Application we want one central application-database/namespace APP where
all app-classes/ routines/csp-sources are hold. Changes in the Application-Logic
can take effect imediately in all customer areas.

There can also an ALL database/namespace if there must some data accessible
from all-customers (some kind of lookups, etc.).

The mappings:

DB/Namepspace CUST1:
--------------------
Standard Global Mapping : APP
Standard Routine Mapping: APP
Global Mapping: <AppPackage>* , Data-Location : CUST1
                <AppAllPackage>* , Data-Location : ALL
                <AppAdminPackage>* , Data-Location : ADMIN

CUST2:
....
CUSTX


We also need an ADMIN database/namespace which implements the Login-Server. All customer-informations needed for login (server, db/ns, user-accounts, etc.) are located there. There can also be an Admin-Console (csp-App., etc.) to add a new customer, move a customer from one to another server, etc.

Because we have a csp-Application all Logins must be routed through
the login-server. (Login-csp-Page)
From the given login-informations (user, password, cust.-no.) a
specific url was builded for redirecting the user to the corresponed
customer-csp-app./namespace. The mapping for this namespace (see above)
will do the job ... i hope so ;-)

Beware in mind that this is theory for me at the moment!!!


Does Anybody has experience in this and a working/running csp-App. with that kind of structure?

I am very interessted in any comments and suggestions regarding this,
especially in someone who has done something like that with success and is
willing to share his experience with me/us.
(db/ns/mapping-restrictions, performance-loss because of extensive mappings,
 configuration-issues, ...)

Thanks and Regards...

Bernd!
-SHD-

[EMAIL PROTECTED] wrote:
Hello all,

We are trying to re-desing our database solution. Currently we have a setup that is similar to this:

DB1:
  * Setup information all Companies
DB[n]:
  * Specific information for Company[n]

where all the DB[n]s share the same schema, but differ in the data they hold. The rationale for that was that a company specific DB could grow quite large. If all the Company Specific data was held in one big database for all client companies, a join with the company would always be required, taking a performance hit since some of those companies have quite a bit of data.

Does it make sense in Cache to split data like that? Is it just a much better solution to have the data all in one DB, instead of spread out?

I know I might not be explaining myself very clearly, I could try to explain it in a different way if it would make it easier to understand.

Thanks for your thoughts,

Daniel.



Reply via email to