Re: IBM i in the cloud (was social media)

["Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>]

> > From: Pete Helgren
> Joe's last response implied
> that out-of-the-box pool managers will return connections for essentially
> any criteria, including "tenant".

No, my specific statement was that the connections are based on connection
string.  Because you can identify schema in the connection string, that
means you can create connection pools by schema, which in turn gives you
the schema-specific partitioning you had requested.  (To understand how
the connection URL works, I suggest some time spent with the connection
properties section of the appropriate DB2 manual.)

The above paragraph is actually a bit of an oversimplification.  Standard
Java EE environments don't let you specify the connection string; instead,
you define data sources which in turn specify the connection properties. 
You then ask the data source for a connection and typically they do their
own pooling, but as you can see it would be by data source.  In that
scenario, you would create one data source for each tenant and then
request a connection from the data source for the appropriate tenant,
which would in turn retrieve a pooled connection.

Joe