On Sat, Mar 1, 2014 at 1:07 AM, Bill Burke <bburke@redhat.com> wrote:


On 2/27/2014 11:31 PM, Travis De Silva wrote:

As per your future plans, if we can get a stateless keycloak co-location
option and also enable external config in a DB when you refactor the
adapter code, that should cover the needs of most developers who want to
go beyond the out of the box solutions.

BTW, I hope with the above changes it would be possible to associate one
war with multiple realms and this is not a core keycloak structure
design issue.


How soon you need this by?  Yesterday?  ;)

In our project, I was going to build the security model with social login and was on the verge of using an open source social login library to start building it when like god sent the keycloak project appeared :) So I am not the one to demand and happy with the little miracles that come my way. Having said that, yesterday would be great :) But seriously if your Jira roadmap is sort of an indicator and beta 1 would be released end of Match, that timeframe is fine for us :)

Like I said earlier, I don't think colocation is necessarily a requirement if we a) provided an option for public clients (don't require a client secret) or b) you had a shared secret between clients for all realms.  The adapter would just extract the realm name from the request, invoke on the keycloak server to get the public information about the realm (i.e. public key), then cache this information locally.


I guess a shared secret would do. Just wondering why we can't use the keycloak-admin realm as the top level realm and use it's secret to get the realm info to be cached locally and from that point onwards, it falls into the current keycloak flow.

I am assuming that the individual keycloak realm admins (as per the change done by Stin on KEYCLOAK-292) will not be able to view the keycloak-admin realm info.
 
Bill


--
Bill Burke
JBoss, a division of Red Hat
http://bill.burkecentral.com