indeed Peter but for the small scale users Andy was talking about it's
the beez neez. Clustering would have stopped them from adopting long
ago never mind databases...
On 11 Mar 2010, at 14:36, Peter Schober wrote:
> * Alistair Young <[log in to unmask]> [2010-03-11 14:48]:
>> There are technologies that can make the persistent store
>> such as an embedded database like Derby which can run inside the IdP
>> with no extra administration other than a one time config file and
>> should ideally be mostly invisible to the sysadmins. As it runs
>> the IdP there's no network connectivity to worry about and it gets
>> backed up with the IdP.
> That's fine until you need to cluster the IdP for availablility
> reasons, e.g. if you're using SAML also for campus WebSSO, not just
> for federated access to publisher's resources.
> Unless, of course, such an embeddeed database can reliably be used
> over a network filesystem (which I doubt) aand shared between the
> And for the whole excercise of clustering the IdP to make any sense,
> then your database system where you store those presistentId (whataver
> it may be) needs to be made highly available as well. Which is where
> complexity and/or costs really come in.