Per-Entity Metadata Working Group

["Cantor, Scott" <>]

On 7/27/16, 4:53 PM, "Tom Mitchell" 
<>
 wrote:

> If the MDQ protocol really is straight up HTTP documents, which it appears 
> to be by
> design, there should be ample opportunity to make those documents highly 
> available via
> CDNs or similar.

I'm not familiar enough with how they work, and I'm getting maybe this is one 
reason Leif's been pushing the SAMLBits idea.

There's still the complexity of getting a web server to map an entityID to 
the right document, and even if that was just a 302 to a more static 
location, you've just put a redirector into the middle as the service that 
has to be up.

But if we're blue-skying, when I think about hosting metadata at the 
endpoints, I don't believe in, nor do I think we can use (because of all the 
URNs and so forth), a well-known location. I think InCommon would have to 
distribute a mapping file of the indirection from entityID to "something", 
and that something could conceivably be hosted anywhere.

That's not what's implemented now, but it's been in my mind as what may have 
to get done.

-- Scott