IdPv3 and its metadata ?

[Cantor, Scott]

On 1/8/14, 3:21 PM, "Ian Young" <ian at iay.org.uk> wrote:
>
>On 8 Jan 2014, at 19:49, Tom Zeller <tzeller at dragonacea.biz> wrote:
>
>> Will IdPv3 consume its own metadata ? (I assume not)
>
>I'm fairly sure we have discussed this at least once, and came down hard
>against it doing so. I'd certainly prefer that it doesn't, as this
>introduces some really hard to debug behaviours when the metadata fed
>back from the federations you're a member of clash with the "real" values.

I don't have a lot of dealbreaker opinions, but this is one I won't budge
on.

>That's one option, but I'd prefer that we think in terms of something
>equivalent to what the SP does, if at all possible. Producing a static
>file and publishing doesn't allow the automatically generated metadata to
>adapt when you make configuration changes or perform upgrades, which I
>think helps a lot with things like additional profiles and algorithm
>agility metadata.

Replicating the generation of the file during installation should be easy
enough that it's not worth skipping, but I don't think generating it at
runtime is realistic for this version. There's a lot of instrumentation
that would have to happen for that, or you'd have to build in metadata
awareness into lots of components (which is how the SP does it).

>> If yes above, should the IdPv3 metadata file be the same or similar to
>> the idp-metadata.xml file the v2 installer creates, or should it look
>> like example-metadata.xml in the SP distribution ?

I don't know what the differences are at this point, they aren't likely
very important.

-- Scott