[OpenID] Combining Google & Yahoo user experience research

John Panzer jpanzer at acm.org
Sun Oct 19 17:12:27 UTC 2008


The OpenID model is scalable, in the sense that it scales well for both 
low security situations (the vast majority of interactions, allowing for 
wide deployment) and allows for trust mechanisms to address higher 
security situations.  It's quite possible for a given RP to trust my IDP 
without verification for me to contact customer service, but if I 
request an (outbound) transfer, I'm going to need to be using a well 
known & trusted IDP or provide alternate authentication.  But I don't 
need to require that until I need to do something that requires such 
security.

Andrew Arnott wrote:
> I don't think Shane was saying it's great for low-security needs.  But 
> if you don't feel you can trust a random OP, Shane was saying you as 
> an RP can choose to trust only certain OPs so they are not random. 
>  Microsoft HealthVault is an example of an RP that chooses merely 3 
> OPs to trust.  The OPs on the other hand don't have to do any extra work.
>
> On Sun, Oct 19, 2008 at 9:45 AM, Brandon Ramirez 
> <brandon.s.ramirez at gmail.com <mailto:brandon.s.ramirez at gmail.com>> wrote:
>
>     So it's great security if you need very little security?
>
>     Transactions of value are precisely where we need federated
>     identity.  I have different logins for my bank, credit card
>     company, car insurance, every everything under the sun.  Except I
>     can share identity between my blog and a site like Flicker.
>
>     - Brandon
>
>
>     On Sun, Oct 19, 2008 at 12:36 PM, Shane B Weeden
>     <sweeden at au1.ibm.com <mailto:sweeden at au1.ibm.com>> wrote:
>
>
>         Brandon:
>         > [...]  Why should I trust a random OP?
>         >
>
>         You shouldn't, and nobody is claiming you should for any
>         transaction of value. What does excite me about OpenID (and
>         InfoCard for that matter) over other SSO protocols like SAML
>         is the zero cost of onboarding additional RP's if I am acting
>         as an IDP. All the RP needs to do (besides following a
>         best-practices secure deployment model) is define that they
>         trust the IDP (e.g. for OpenID define a trusted list of OP
>         endpoints) and the IDP need do nothing in particular.
>
>         Sure, there are dynamic extensions to SAML like those defined
>         by Shibboleth for dynamic metadata sharing, but out-of-the-box
>         nothing I've been exposed to thus far quite matches the
>         simplicity of the OpenID model.
>
>         =shane
>
>
>
>     _______________________________________________
>     general mailing list
>     general at openid.net <mailto:general at openid.net>
>     http://openid.net/mailman/listinfo/general
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> general mailing list
> general at openid.net
> http://openid.net/mailman/listinfo/general
>   

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.openid.net/pipermail/openid-general/attachments/20081019/dfed8113/attachment-0002.htm>


More information about the general mailing list