Yes, the problem still exists with any other mechanism of password
authentication that is not SAML based, Kerberos (in some cases depending
on who manages the key store) or PKI based. Microsoft MAPI works
similar to IMAP in that a password is exchanged. While both have
encrypted transport mechanisms to prevent snooping in transit, the
password is delivered to the mail provider. The worry is that should a
compromise take place at the provider's end, the SSO password for access
to institutional resources would require additional work =$$ to clean
up. So why risk putting the institutions SSO password at risk. The
solution is to segregate access through these protocols by requiring a
different credential. Some institutions force this action via
technology by funneling password changes through an in-house web based
application that stores authN credentials in a database, then compares
them to the SSO password and disallows the use of the same password in
I hope this helps
On 03/18/2011 08:02 AM, Bowen, Hugh E. wrote:
> Hi Barry,
> The items listed are not only informative, but also thought provoking. Specifically the item associated with IMAPS, and smart phone service providers. With that said, do you happen to know if this type of exposure also exists with Microsoft's cloud email offering, and the smart phone service providers?
> Hugh Bowen
> Tufts University
> -----Original Message-----
> From: Identity Management Constituent Group Discussion list [mailto:[log in to unmask]] On Behalf Of Barry R. Ribbeck
> Sent: Thursday, March 17, 2011 8:34 AM
> To: [log in to unmask]
> Subject: Re: [IDM] Google and password synchronization
> Hi Chris,
> The sync or no-sync discussion was one that we looked at heavily.
> Here are some of the conditions and factors that we thought were pertinent.
> 1) What is the value of the data that will be stored in Google long term?
> 2) Is there a difference in Level of Assurance for the credential in
> Google and the credential we offer?
> 3) What is at risk if we make our credentials available to an external
> service provider? Has this already been done with other ASPs (online
> hiring, Hippa training sites etc.)
> 4) If we use SSO via Shibboleth, then we would not expose our passwords
> through web based access but if a user chooses IMAPS through a phone or
> thick client, then the password is exposed at least to Google. Is the
> added complexity, confusion and effort of differentiating the
> authentication credential for IMAP equivalent to the value of a password
> compromise through Google and if so what is the chance of that occurring?
> 5) Can we do better and if so what is the cost?
> This is the direction we are leaning. I know that each institution has
> different answers to the questions above and therefore may take a
> different path after answering these questions so we welcome comment.
> There are a number of new services and chances to expose data in Google,
> this is not just about email so there is a slightly elevated exposure of
> additional data that does not exist today in our internal services, but
> not much as we currently offer chat, blogs, web sites, file sharing,
> email etc.. so we feel there is little additional data at risk. The
> value of the data is defined as student owned data. We currently use
> our internal email service to communicate FERPA protected data from the
> registrar and faculty to students, there will be no difference with
> Google, so we want to make sure that we use at least the same LOA
> credential that we use internally. This is a vote for integration for
> SSO. We already have a few applications from ASPs that leverage our
> directory services data and authentication services, so this is not a
> new threshold for us and we should use the knowledge gained in those
> experiences and our reasoning for those decisions should extend to an
> agreement with Google. Lastly, the IMAP issue. I am currently of the
> opinion that if the user exposes their Rice SSO password through IMAP to
> Google, we have as much control over this as we would if they used the
> same password in their personal bank service. While I do not condone
> this use, it seems that whatever technical effort we put in place there
> would be more of a barrier that people would feel as unwelcome and
> overly burdensome. So we use 2 factor authN for things we really want
> to protect the authentication with and SSO for things of lesser value
> like email and use policy and education to enlighten the users on the
> best practices for credential management.
> Can we do better? I believe we can. Google and others such as ebay
> etc. are moving toward OTP as a means of getting around the pitfalls of
> passwords. While in my opinion the Google implementation is not quite
> baked yet it is a step in the right direction. It still does not solve
> the IMAP issue however and this is a harder problem that involves
> software vendors and phone manufacturers.
> I hope this helps
> On 03/17/2011 03:56 AM, Chris Brezil wrote:
>> We are in the process of moving our students from our present in house GroupWise email system to Google for email and application services. In doing so, the question of whether to synchronize student passwords, currently stored in our directory servers, should be synchronized to Google as well. Clearly others have gone through this same exercise as well and looked at it from technical, policy, security, privacy, end user experience, etc. points of view. We are interested in what decision you made about this as well as hearing about the driving elements behind making the decision. Also, after the test of time, do you believe that your decision was still the right one?
>> Chris Brezil
>> Assistant Vice President/ Information Technology
>> The New School