Re: Self-service password reset approaches

[Kevin Shalla <kshalla () UIC EDU>]

We're not using Telesign, but it reminds me of a system we use here for 
authentication during telephone calls.  When someone calls us and we 
need proof that the person is who he says he is (like wants to talk 
about confidential matters), we ask him to go to a certain Web page, 
login there, then read the one-time-use generated token.  The telephone 
operator then enters this on a different Web page, and it indicates 
which netID generated that token.  It's much more secure than asking a 
bunch of questions, and hoping that only the correct person can answer 
those questions correctly.

On 2/14/2012 8:41 AM, Theresa Rowe wrote:
> We've recently looked at a product that does use the phone to send a 
> separate code:  Telesign (telesign.com <http://telesign.com>).  Any 
> comments from a current user out there?
>
> Theresa
>
> On Tue, Feb 14, 2012 at 9:37 AM, Roger A Safian 
> <r-safian () northwestern edu <mailto:r-safian () northwestern edu>> wrote:
>
>     In my mind, there’s no doubt that security Q & A’s are considered
>     effective by both the users and the help desk team.  OTOH, from
>     the security point of view they are a huge risk.  Especially in
>     education.  Our constituents are too young for many questions to
>     be effective, so things like “what street did you grow up on?” or
>     “What was your first car?” may in fact be the street they live on,
>     or the car they drive, now.  In addition, people choose poor
>     questions if they can make their own.  “What color is an orange?” 
>     The number of compromises of celebrity accounts shows just how
>     risky these questions are.  Personally I’d love to ditch them,
>     but, I would need something more effective, and some way to pay
>     for it.  I’d like to see something tied to their phone.  Maybe
>     they register their cell, and we sms them a temp password.  Not
>     ideal, as people can be overseas, but I think it would be safer.
>
>     *From:*The EDUCAUSE Security Constituent Group Listserv
>     [mailto:SECURITY () LISTSERV EDUCAUSE EDU
>     <mailto:SECURITY () LISTSERV EDUCAUSE EDU>] *On Behalf Of *Theresa Rowe
>     *Sent:* Tuesday, February 14, 2012 8:00 AM
>     *To:* SECURITY () LISTSERV EDUCAUSE EDU
>     <mailto:SECURITY () LISTSERV EDUCAUSE EDU>
>     *Subject:* Re: [SECURITY] Self-service password reset approaches
>
>     We are looking at these processes, too.  I am surprised to read
>     Steve's response about phasing out security questions and
>     answers.  We just implemented this in 2011 and it has been very
>     helpful.  With multiple campuses and online learning, we can't
>     expect our constituents to visit campus.  We accept a faxed photo
>     identity, along with other security information, and will call
>     back with IDs and a pin reset that is forced on first login.
>
>     Account claiming - specifically providing the student ID number -
>     is our biggest challenge.  How are you folks handling that?  We
>     used to have a discovery web site, but we were told it wasn't
>     FERPA compliant to display student ID like that.  Then we switched
>     the site to email the ID, but that didn't work because the
>     individual didn't have access to email if they hadn't set it up,
>     and they needed the ID to set it up (catch-22).
>
>     Appreciate the discussion -
>     Theresa
>
>     On Thu, Feb 9, 2012 at 4:53 PM, Steve Werby <steve.werby () utsa edu
>     <mailto:steve.werby () utsa edu>> wrote:
>
>     Dave,
>
>     It’s good to see others considering the progressive approach that
>     other industries have already adopted. Security questions are
>     fraught with problems and put the users’ accounts with other
>     organizations at risk.
>
>     We’ve been designing and developing a system to move from password
>     resets via answering security questions to resets via unique code
>     sent to an alternate email address or mobile phone number via SMS.
>     It’s entirely optional, but since we’ll be phasing out (and
>     deleting) the security questions and answers, the next available
>     disclosed alternative will be a physical visit to an authorized
>     office who will require ID to be displayed. We’re bundling the new
>     process with a change from a typical password
>     complexity/composition policy to a 15+ character passphrase. We’re
>     doing usability testing with a range of users right now and our
>     pilot starts in March.
>
>     -- 
>
>     Steve Werby
>
>     Information Security Officer
>
>     Office of Information Security (OIS)
>
>     The University of Texas at San Antonio
>
>     *From:*The EDUCAUSE Security Constituent Group Listserv
>     [mailto:SECURITY () LISTSERV EDUCAUSE EDU
>     <mailto:SECURITY () LISTSERV EDUCAUSE EDU>] *On Behalf Of *David Curry
>     *Sent:* Monday, February 06, 2012 8:18 AM
>     *To:* SECURITY () LISTSERV EDUCAUSE EDU
>     <mailto:SECURITY () LISTSERV EDUCAUSE EDU>
>     *Subject:* [SECURITY] Self-service password reset approaches
>
>     It's been a few years since this has come up on the list, so here
>     goes.
>
>     For various administrative reasons having nothing to do with
>     security we need to make some big changes to our self-service
>     password reset approach, and I'm trying to capitalize on the
>     opportunity to improve its security at the same time. At the
>     moment, we do what (we think) many other schools do -- provide
>     student id number, netid (username), and date of birth, and you
>     can reset your password. The problem with this is, of course, it
>     was never that hard to come up with that information in the first
>     place, and the combination of students doing more and more stuff
>     online and the growing use of social media makes it just that much
>     easier.
>
>     So... what other approaches are you taking?
>
>     There is of course the "pick a few security questions" approach.
>     But it's hard to come up with a set of questions whose answers
>     aren't trivial to guess (either because they have little if any
>     entropy or because the answer is on Facebook). And if you do
>     manage to come up with a set of hard questions, people can't
>     remember what their answers were. Do you use this approach? If so,
>     how have you addressed these problems?
>
>     We've been tossing around the idea of using something similar to
>     the "email confirmation" links you see many forum-type websites
>     use. In this approach, we would ask the user for some identifying
>     information (netid, student id number, etc.) and then look up the
>     email addresses we have on file. The user could choose any
>     non-university email address in the list, and we would send a
>     randomly-generated URL to that account, which the user could then
>     click on to reset his/her password. Users for whom we have no
>     alternative email on file (or for whom all the ones we have on
>     file are "no good") would have to call the help desk. Does anybody
>     use an approach like this? How well is it working (or not working)?
>
>     Any other "interesting" approaches out there?
>
>     Thanks,
>
>     --Dave
>
>     --
>
>     *DAVID A. CURRY, CISSP* • DIRECTOR OF INFORMATION SECURITY
>
>     *THE NEW SCHOOL* • 55 W. 13TH STREET • NEW YORK, NY 10011
>
>     +1 212 229-5300 x4728 • david.curry () newschool edu
>     <mailto:david.curry () newschool edu>
>
>
>
>
>     -- 
>     Theresa Rowe
>     Chief Information Officer
>     Oakland University
>
>
>
>
> --
> Theresa Rowe
> Chief Information Officer
> Oakland University