Re: SQRL (Re: [SECURITY] Image, word, and password login)

["Mclaughlin, Kevin (mclaugkl)" <mclaugkl () UCMAIL UC EDU>]

Ben and all:

I would like to hear a bit more on the idea of biometrics asserting identity and not authorization.  I am curious as to 
the distinction and what would then provide authorization.   This seems like a change in the tried and true 1,2,3 
factor of authentication which is used to provide proof of your identity but which is then used to provide the 
identified user with authorization to the items they are entitled.  If that is the case wouldn't that mean that any of 
the forms (password, biometric, token, etc.) are for both asserting identity and authorization.

- Kevin

________________________________________
From: The EDUCAUSE Security Constituent Group Listserv [SECURITY () LISTSERV EDUCAUSE EDU] on behalf of Ben Marsden 
[bmarsden () SMITH EDU]
Sent: Friday, December 06, 2013 4:07 PM
To: SECURITY () LISTSERV EDUCAUSE EDU
Subject: [SECURITY] SQRL (Re: [SECURITY] Image, word, and password login)

I hate passwords as an authentication tool...

and I side with the concept that biometrics assert identity, not authorization....

So, is anyone else looking at SQRL as a possible implementation option?  IMHO it's a pretty nice concept, but it's in 
the early stages of development.

In a grossly oversimplified nutshell, it uses site-specific asymmetric key pairs to identify / authenticate an 
individual with pretty minimal user interaction -- seems both easy to use (user friendly) and robust (PEBKAC averse + 
compromise resistant).

  https://www.grc.com/sqrl/sqrl.htm   and
   http://www.sqrl.pl/  -- for a more graphical description of the process

 fwiw,  -- Ben
============================================
Ben Marsden : Information Security Director, CISSP/GISP
ITS, Stoddard Hall, Smith College, Northampton, MA 01063
bmarsden () smith edu<mailto:bmarsden () smith edu>     (413) 585-4479
---------------------------------------------------------------------
=--> Any request to reveal your Smith password via email is fraudulent!



On Fri, Dec 6, 2013 at 3:36 PM, Karl Bernard <karl.bernard () gmail com<mailto:karl.bernard () gmail com>> wrote:
A colleague mentioned this thread to me and I'm a consumer of this same technology (site authentication) at a couple of 
financial sites. Until today, I'd always thought it was pretty cool until I was trying to find the official name for 
this kind of thing and found some less than stellar articles and studies about using them:
http://www.finextra.com/news/fullstory.aspx?newsitemid=16469
http://security.stackexchange.com/questions/19155/effectiveness-of-security-images

Karl Bernard
UTHealth, Academic Health Center at Houston



On Fri, Dec 6, 2013 at 1:18 PM, Joel L. Rosenblatt <joel () columbia edu<mailto:joel () columbia edu>> wrote:
Hi,

I have an account at a vendor that uses a system like this - I picked
a picture and a word, and when you enter your account (before your
password) it takes you to a page that displays the picture and word
and prompts for the password

It makes the login a 2 screen affair, which may bother some of your
users who think that everything has to be done in subsecond time.

Our web login to our mail system displays

Greetings, Joel Rosenblatt (or your own name :-)

after you type in your account, but before you type in your password -
similar idea, but less pages and it doesn't require the user to do
anything except recognize their name :-)

Good luck!
Joel


Joel Rosenblatt, Director Network & Computer Security
Columbia Information Security Office (CISO)
Columbia University, 612 W 115th Street, NY, NY 10025 / 212 854 3033<tel:%20212%20854%203033>
http://www.columbia.edu/~joel
Public PGP key
http://pgp.mit.edu:11371/pks/lookup?op=get&search=0x90BD740BCC7326C3


On Fri, Dec 6, 2013 at 2:08 PM, Derek Diget
<derek.diget+educause-security () wmich edu<mailto:derek.diget%2Beducause-security () wmich edu>> wrote:
> We are thinking about creating a login process where user's pick a picture
> and/or word before getting a password entry box.[1]  The main driver is to
> prevent phishers from copying our "static" login pages.
>
> The process would go something like....
>
> 0) Training, Training, Training...and other carbon based life form user
> issues...... :)
>
> 1) User gets to our login page
> 2) User enters login ID
> 3) login process retrieves user's picture and word choice
> 4) login process displays user's picture with 8 (or 11) others randomly
> 5) User selects their picture
> 6) If correct, login process displays user's word with 8 (or 11) others
> 7) If correct, login process give user a password text box to finish
> authenticating.
>
> (Yes, a phisher could duplicate the pictures and words and disregard what
> the user picks...so the user would always get to the password box, but our
> current thoughts is that it would take to much "work" for them to duplicate
> this new login process and there are other easier fish in the sea to phish.
> :)
>
> I have two questions to the group....
>
> 1) Is there an industry term for this type of authentication process? (It
> kind of is two-factor, but we want to avoid using that term as most people
> think of two-factor having a physical component...token card, key fob,
> phone, etc).
>
> 2) Does anyone know of any research on a multi-step authentication process
> like this?  Be it usability issues, increased security, etc.
>
>
>
> Note 1: We vet the user.  As part of the process of setting a password, they
> also pick a picture out of ~12 (with a library of 100+) choices and store
> their choice.  They then pick a word out of ~12 (with a library of 100 or so
> words) and store their choice.  Then they finish setting a password.
>
> --
> ***********************************************************************
> Derek Diget                            Office of Information Technology
> Western Michigan University - Kalamazoo  Michigan  USA - www.wmich.edu/<http://www.wmich.edu/>
> ***********************************************************************