How to fix authentication (was LinkedIn)

AP NANOG nanog at armoredpackets.com
Fri Jun 22 14:24:15 UTC 2012


I used the example I did based on YubiKey, I own one and use it on a 
regular basis.  The real issue I am trying to make is the fact that even 
in the scenario I placed forward it still requires trust.  Trust of a 
person or trust of a company.  This reminds me of a quote:

                     Only two things are infinite, the universe and 
human stupidity, and I'm not sure about the former.
                         - Albert Einstein

By no means am I saying any of us, or the majority of the world is 
stupid or uneducated.  However, the inherent nature behind trust is just 
that, relying on some sort of other party is the weak link here.  It 
only takes a single person who has a bad day, or just wants to slack off 
for that day, to create a vulnerability in any password, key, 
encryption, or authentication process hundreds if not thousands of 
people work so hard to solve.

While I used YubiKey as my original example, and use it on a regular 
basis, it still has its downfalls.  It cannot be used with Active Sync, 
so ultimately you can not use it for your Active Directory log in 
because of a small thing called Exchange.  There have been other areas 
were YubiKey has failed but not by it's design, but by the design of the 
application itself.

How can any of our solutions over come the human factor?

-- 

- Robert Miller
(arch3angel)

On 6/21/12 10:53 PM, Christopher Morrow wrote:
> On Thu, Jun 21, 2012 at 10:48 PM, Randy Bush <randy at psg.com> wrote:
>>> That's basically the Yubikey. It uses a shared key, but since you're
>>> relying on a trusted third party anyway
>> there are no trustable third parties
> note that yubico has models of auth that include:
>    1) using a third party
>    2) making your own party
>    3) HOTP on token
>    4) NFC
>
> they are a good company, trying to do the right thing(s)... They also
> don't necessarily want you to be stuck in the 'get your answer from
> another'
>
> -chris




More information about the NANOG mailing list