Integrating idp3 with cirqlive

[Joseph Fischetti]

Thank you all for the input and examples. I have enough info now to choose a route (which will probably be the separate, shared private key route). 

For what it's worth, during my testing the hardest part of the whole integration was the auto provisioning. I had very few issues getting the user to simply authenticate as all webex really wanted was their username (email address). Auto provisioning was a different story as, as already noted, webex wants certain attributes and does a very poor job outlining that.

On top of everything else... we're also using the idp for authorization to webex. That is, we have a separate webex based on role, and I had to build attribute filter to release to the separate webexes based on the user's attributes. The major downside to this is that rather than receive any actual error message, webex just throws an error that a valid username wasn't found in the saml assertion. 

Joe


Joseph Fischetti 
Linux System Administrator 
Marist College 

E-mail: joseph.fischetti at marist.edu 


-----"users" <users-bounces at shibboleth.net> wrote: -----
To: Shib Users <users at shibboleth.net>
From: "Cantor, Scott" 
Sent by: "users" 
Date: 08/22/2017 10:51PM
Subject: Re: Integrating idp3 with cirqlive

On 8/22/17, 10:34 PM, "users on behalf of Rhian Resnick" <users-bounces at shibboleth.net on behalf of rresnick at fau.edu> wrote:

> CircLive uses these integrations to create, provision, and organize WebEx functions in a way that integrates into Canvas, 
> Blackboard, or D2L (etc...). Without them WebEx just doesn't have the organization and permissions/filtering to work with
> a learning management system.

I don't begrudge them the organizing piece or charging for that value; just the authentication part that they don't need to be doing, since the user wouldn't notice the difference.

> Note setting up WebEx with Shibboleth pretty difficult as they have very poor documentation related to the custom attributes
> they require to be sent.

There's obviously nothing we can do about that, but if we don't document these things after we reverse engineer them, it's never going to get better for the next person. All of us share that problem.

-- Scott


-- 
To unsubscribe from this list send an email to users-unsubscribe at shibboleth.net