Re: [TLS] one time passwords from private keys

On 28 February 2010 15:09, Ben Laurie <benl@google.com> wrote:

>
>
> On 28 February 2010 14:58, Story Henry <henry.story@bblfish.net> wrote:
>
>> Hello,
>>
>>  I am looking around to see if anyone knows an algorithm to allow one to
>> produce a one time password [1] from the private key part of an asymmetric
>> key pair.
>>
>> The article "Public-Key Cryptography and Password Protocols"
>> http://www.cparity.com/projects/AcmClassification/samples/322514.pdf
>> shows how to do the inverse.
>>
>>  The idea is to make it possible to authenticate into any web site using
>> this one time password. The public key would be tied to a WebId/OpenId,
>> following something like the following steps:
>>
>> 1. A server should be able to know for any webId/openid the public key of
>> that id
>>
>> 2. the user will have
>>  - the private key
>>  - using its private key (and perhaps a time stamp, or a nonce from the
>> server, ... ) the user's software would calculate a one time password which
>> the user could then send with its WebId/OpenId to the server
>>
>> 3. the server which gets the openid/webid find the public key securely and
>> use
>>  that to verify the one time password (somehow)
>>
>> This would allow us to make it easy to create RESTful authentication for
>> devices with broken (Safari on desktop and iphone) or non existent client
>> side certificates (Android, and most other phones).
>>
>> For devices with working client side certificates we have foaf+ssl that
>> works reasonably well currently - though we would love the browser vendors
>> to implement user interface improvements.
>> http://esw.w3.org/topic/foaf+ssl
>>
>>
>>  What we are looking may be impossible. But we can't tell before looking
>> :-)
>> Also if you know of lists that are open source/open standard spirited that
>> you think I should send this question to, please let me know.
>>
>
> One way of doing what you want is a Schnorr signature. One nice property of
> a Schnorr signature is that the private key can easily be derived from a
> password, or could just be a random blob. I'm doing some work in this area,
> so feel free to contact me offlist if you're interested.
>

Also, of course, if you have an RSA keypair you could just sign the nonce
with it :-)


>
>
>>
>>        Henry
>>
>> [1] http://en.wikipedia.org/wiki/One_time_password
>>   http://code.google.com/p/mod-authn-otp/wiki/OneTimePasswords
>>
>> Social Web Architect
>> http://bblfish.net/
>>
>> _______________________________________________
>> TLS mailing list
>> TLS@ietf.org
>> https://www.ietf.org/mailman/listinfo/tls
>>
>
>