Our clustering solution for IdP v3

[Tom Scavo]

On Mon, Feb 29, 2016 at 8:27 AM, Peter Schober
<peter.schober at univie.ac.at> wrote:
>
> AFAIU computedIds are not reversible (i.e., the IDP cannot map
> them back to a subject when presented with one in-bound) so those
> would be useless for SPs that relied on attribute queries. You'd still
> need to be able to issue (and probably keep announcing support for)
> transientIDs.

Good point. This begs the question: Why would one ever assert a
computedid as a SAML2 Persistent NameID (as recommended by saml2int)?

Tom