<div dir="ltr"><div>While implementing Session Management, Back and Front channel logouts i find myself struggling to find concrete details on the "sid" claim in general.<br></div><div><br></div><div>Questions i find myself asking</div><div><ul><li>if the client's subject_type is public, is it okay to just send my internal OP session id? </li><li>if the client's subject_type is public can i send the same sid to every client, afterall the sub claim will be the same for all these clients so where's the point in different sid claims</li><li>if the client's subject_type is pairwise i suppose i SHOULD calculate the sid values unique for each client, else instead of matching users by sub clients can do so by sid.</li></ul><div>My current proposed implementation is to return the same OP session ID as sid to all public clients and use the same mechanisms for calculating sid for pairwise client claims as i do for sub.</div></div><br clear="all"><div><div class="gmail_signature" data-smartmail="gmail_signature">Kind Regards,<br><b>Filip Skokan</b></div></div><div class="gmail_signature" data-smartmail="gmail_signature"><b><br></b></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><pre style="white-space:pre-wrap;color:rgb(0,0,0)">One of the things blocking us from taking the Front-Channel and Back-Channel logout specs to Implementer's Draft status is that they currently use the Session ID ("sid" claim) differently. The Front-Channel spec expects "sid" to be globally unique. The Back-Channel spec only expects it to be unique in the context of an IdP - just like "sub" is.
In talking with John and Nat and William at IETF 95, we discussed the possibility of unifying the definitions by changing the Front-Channel usage to match the Back-Channel usage. In the case when logging out only one session at an RP using the front-channel spec, this would necessitate adding an additional "iss" (issuer) parameter to the logout request, so that it would become:
logout_uri?iss=issuer&sid=session
In the simple case, one could still use only:
logout_uri
without any parameters if it is OK to log out all sessions at the RP.
Are people OK with this change? It could provide us a way forward to bring both the Front-Channel and Back-Channel specifications to Implementer's Draft status.
-- Mike</pre></blockquote>
</div>