From:  Mike Jenkins <[email protected]>
Date:  Wednesday, October 24, 2012 12:21 AM
To:  <[email protected]>
Subject:  Re: [therightkey] Improving Trust

> I don't think trust is that nebulous an idea for many things that people do
> with browsers. We've got to stop convincing people that "this certificate is
> okay", and start informing them "it's okay to enter a credit card number
> here".
> 
> I often wonder why the browser doesn't have the ability for me to label my
> trust anchors with labels like "bank" and "school", and then indicate to me
> when a certificate validation has terminated in a labeled trust anchor. How I
> apply those labels is my policy - based on prior experience, or obtaining a
> "fingerprint" from a letter from the bank, or what have you. It doesn't even
> prevent me from working with other banks, unless I want to limit myself.

This is along the lines of a browser feature I've wanted for a while ­ the
means for the user to indicate the nature of what they are doing ­ i.e., I'm
shopping or banking or doing medical stuff, etc.  A narrow trust anchor
store associated with purposes like these could allow the posture of the
browser to vary based on what the user says they are doing.

> 
> This would be even more useful for device apps, that could be initialized to
> only use one trust anchor from the list.
> 
> There have got to be ways to do this that are fairly straightforward, not too
> tedious for youth and not too complicated for elderly folks. [Have I insulted
> everyone yet? Alrighty then.]
> 
> Mike
> 
> On Tue, Oct 23, 2012 at 2:46 PM, Bill Frantz <[email protected]> wrote:
>> On 10/23/12 at 10:35 AM, [email protected] (Rick Andrews) wrote:
>> 
>>> ... IMO the browser vendors have not helped much to improve the status quo.
>>> They created a trust system in which all CAs are trusted equally, and they
>>> are reluctant to remove any CA's trusted roots from their browser. I'd like
>>> to brainstorm how those issues can be addressed.
>> 
>> Trust is such a nebulous idea. For example, I could trust by cousin to abuse
>> alcohol before he died from alcohol-related causes.
>> 
>> What I think we are trying to do is ensure that the user can rely on the
>> normal clues to people use to correctly identify a web site (or other
>> computer-mediated communication). Probably users use different visual, and
>> possibly audible clues, but probably the primary clue they use is the look
>> and feel of the web page. How do we get our public key based trust system to
>> separate out the fishing sites?
>> 
>> We have tried to train users to look at the browser's crome. To the extent we
>> have been successful, we can now discuss how to improve both the use
>> interface from the crome, and the underlying trust model that supports that
>> interface. If we are going to assign differently levels of assurance to
>> different CAs, we will need to reflect those different levels in the chrome
>> UI.
>> 
>> One way to improve the underlying trust model is to use more that one
>> identification technique. We currently use only the PKIX/CA system. As Rick
>> points out, all CAs are equal as far as our trust model is implemented. If we
>> could notice that the last time we visited this site, it had the same public
>> key, that would improve trust, giving us two paths of recognition. If we also
>> got the same information through DANE, that would give us three paths.
>> 
>> When these paths disagree, we have a problem. What do we do when the CA has
>> issued a CRL with the site, but the public key is the same as the last time,
>> and the site doesn't use DANE. What if site uses DANE and it says the key
>> belongs to the site? Should we have more paths to establish identity, the
>> number of possibilities will grow rapidly.
>> 
>> 
>> Consider a nearly real world trust establishment problem. My son is driving
>> from California to Pittsburg PA to go to collage. I receive a communication
>> from him saying the car has needed major repairs and to please send him
>> $2000. How do I establish trust in this communication.
>> 
>> If the communication is two-way, a phone call for example, I can engage in a
>> dialog similar to the security questions requested by web sites. If it is a
>> voice call, I can use my built in voice recognition system.
>> 
>> If the "send to" address for the money in on a logical route between CA and
>> PA, I'll react differently than if it is in Nigeria.
>> 
>> Etc. etc.
>> 
>> Cheers - Bill
>> 
>> -------------------------------------------------------------------------
>> Bill Frantz        | Airline peanut bag: "Produced  | Periwinkle
>> (408)356-8506 <tel:%28408%29356-8506>       | in a facility that processes
>> | 16345 Englewood Ave
>> www.pwpconsult.com <http://www.pwpconsult.com>  | peanuts and other nuts." -
>> Duh | Los Gatos, CA 95032
>> 
>> _______________________________________________
>> therightkey mailing list
>> [email protected]
>> https://www.ietf.org/mailman/listinfo/therightkey
>> <https://www.ietf.org/mailman/listinfo/therightkey>
> 
> _______________________________________________ therightkey mailing list
> [email protected] https://www.ietf.org/mailman/listinfo/therightkey


_______________________________________________
therightkey mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/therightkey

Reply via email to