Shibboleth Developers

["Scott Cantor" <>]

> Okay, so we'll wait and see how you implement persistent identifiers
> before we try to add anything on top of 1.3.  Btw, have you settled on
> a persistence mechanism for 2.0?  We were looking at derby but now I
> see grouper is using hibernate + hsqldb so I'm not so sure any more. 
> What are you gonna use?

The design probably needs to be a persistence interface. The sample might
include Derby, but nobody in their right mind is going to deploy that in
production.

There *still* isn't going to be a production implementation in the box.
There can't be, because production in this case is defined as "support for
clustered use via whatever database is already used in your shop". In our
case, that's likely to be Oracle, for example, or possibly DB2 since that's
free now.

But the sample will be much closer to what a real deployment would be based
on than the current one. That's the best we can do.

An analogy would be the SP, which will include an ODBC-based cache plugin.
That doesn't give you a clustered deployment. It just works properly once
you add in your site's choice of database and all the surrounding tools.

Personally, my requirement is that I can drop in a clustered driver like the
c-jdbc stuff. If hibernate can support that, that's fine too, but I won't
introduce a single point of failure into my deployment.

-- Scott