Shibboleth Developers

["Scott Cantor" <>]

> I've heard the word "database" mentioned a couple of times
> recently...has there been any discussion about including a database
> with the Shib 2.0 IdP?

Persistence is virtually mandatory for federated IDs, and helpful for other
features.

> Our project requires a database so we might as
> well use whatever Shib intends to use.

There is no realistic way we can ship something people will actually use in
production. That isn't really a problem, since the gap from "download" to
"production" for something like Shibboleth is huge.

What we can do is properly abstract the interfaces, create files/resources
with replacable SQL code for porting, and possibly look at including some
cluster-aware software out of the box.

I'm quite interested in this:
http://sequoia.continuent.org/HomePage

I've evaluated it in the past, and it seems like the only practical way to
avoid database-specific clustering, which I don't trust all that much
anyway. I think using a single database is a horrible idea when you have
such an otherwise cluster-friendly system.

I also noted this:
http://carob.continuent.org/HomePage

That seems very promising for the SP so I can dump that MySQL code.

-- Scott