<html>
<head>
<meta content="text/html; charset=windows-1252"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
<br>
<br>
<div class="moz-cite-prefix">On 6/1/15 7:16 PM, Cantor, Scott wrote:<br>
</div>
<blockquote cite="mid:F57ECC02-FFC2-44B3-9DA8-7DC5B2000410@osu.edu"
type="cite">
<pre wrap="">On 6/1/15, 11:12 PM, "Brent Putman" <a class="moz-txt-link-rfc2396E" href="mailto:putmanb@georgetown.edu"><putmanb@georgetown.edu></a> wrote:
</pre>
<blockquote type="cite">
<pre wrap="">
I forgot to comment earlier on that specific detail. I don't know what
is being envisioned here, but: doesn't anything involving use of SAML
metadata fundamentally assume an entity identifier of some kind?
</pre>
</blockquote>
<pre wrap="">
Not necessarily on the wire. The common case is when you have to reverse
map from an endpoint, which is the case here.
</pre>
</blockquote>
<br>
Given we're envisioning a move from batch to dynamic metadata:
Doesn't dynamic imply that say an MDQ server (or similar) needs to
index on multiple properties of the metadata, and then be able to
answer queries accordingly?<br>
<br>
I don't recall off-hand, but does the MDQ protocol spec really
account for this use case? I remember it having a general notion of
identifiers in the query, but is it sufficiently flexible right now
to support this, e.g. query by endpoint (or a an endpoint-matching
regex) and so on?<br>
<br>
Otherwise I see Marvin's CAS stuff as well as SAML artifact, etc
only being supportable when the entity is in a metadata batch.<br>
<br>
<br>
</body>
</html>