Vendor independent data source

[smitty_in_vancouver]

I have an application that I deployed to WebSphere using the default
Struts datasource from it's action. I've been reading about this now
and it sounds like this is not the prefered way to make a database
connection pool. They suggest using the containers, but I want to
remain vendor independent and be able to deploy my application in any
container with little or no changes.

I know there must be a standard non-vendor specific way to create a
connection pool. Can someone tell me what that standard is? Searching
the web just provided me with about 100 different ways to do it, but I
want the "Java Standard" way to do this.

Thanks,
Smitty

 

[anonymous]

I have an application that I deployed to WebSphere using the default
Struts datasource from it's action. I've been reading about this now
and it sounds like this is not the prefered way to make a database
connection pool. They suggest using the containers, but I want to
remain vendor independent and be able to deploy my application in any
container with little or no changes.

I know there must be a standard non-vendor specific way to create a
connection pool. Can someone tell me what that standard is? Searching
the web just provided me with about 100 different ways to do it, but I
want the "Java Standard" way to do this.

Thanks,
Smitty

FWIW, we have given up using the data-sources.xml for Oracle even when
using their OC4J containers. We have written our own based on their OCI
driver. Is a singleton and works very well.
Our implementation uses abstraction and it is easy to create another
instance using IBM's or MS or Mysql methods.
We have not found any way to have one do all.