[Cryptography] Best internet crypto clock

Lodewijk andré de la porte l at odewijk.nl
Tue Oct 7 15:02:46 EDT 2014


2014-10-07 16:21 GMT+02:00 Arnold Reinhold <agr at me.com>:

> I would envision including a good quality internal clock, set at time of
> manufacture and non alterable. (When the clock battery dies, the camera is
> toast.) The camera would periodically or on command output a signed
> certificate containing the current reading of its internal click and maybe
> an external nonce like the NIST beacon, which might then be sent to a time
> stamping service, creating a record of internal clock drift over time.. The
> camera might store a correction factor, so it could output a UTC time, but
> the internal clock would be included in any certificate as well.
>
> It would seem that a camera like this would be useful in a variety of
> applications (besides kidnapping) to create legally provable documents.
> Assuming it had a video mode, it could be used as a notary, recording a
> person's spoken acceptance of contract, or witnessing his handwritten
> signature on a document. Of course one would still have to trust the
> manufacturer.
>

Just put it in the SD card, not the camera. Lot cheaper to replace. Of
course, that doesn't mean you trust the contents, just that the image
existed at some point in time (which is pretty good!). Then you could have
additional certificates from the camera, if the HSM self-destructs there'll
be no more certificates. If I can show such a file + certificate, then it
means a valid camera shot that specific image, just not at what time
specifically.

The failure is that there's more than the eye can see.. An unedited video
of a fake moonlanding is still a fake moonlanding, even if it looked real.
For that matter, something faked at the same time as when something was
supposed to happen still looks real anyhow! So, I guess the in-SD-card
option seems better. It'd be cool to have for other files too, claiming
prior art over an idea that someone's patenting because of a sketch or a
dribble in your (digital) notebook? Pretty cool! Preserves privacy and
solves conflicts in a really positive way. It would also really kill overly
broad patents, because everyone has probably thought about it at some point
:). You can even sell your notebook's page-with-similar-enough-note to a
party that wants to invalidate a patent! That'd really turn things upside
down and around again.


> A signing camera isn’t a new idea, a quick Google search came up with this
> 1992 paper
> http://www.friedmanarchives.com/Writings/Trustworthy_Digital_Camera_Technical_Paper.pdf
> , but camera technology developed for cell phones makes something like this
> much more affordable. Has anyone attempted this? How close could we get
> with an iPhone 6, given Apple's improved security scheme?
>

Secure coprocessors can do just about anything. The wait is on for the NSA
chips to begin rolling out as "Secure coprocessors" :]

Of course, it's also cheating. You're just putting a central entity in
charge of authenticity, and duplicating the whole entity a whole lot.

I suspect that on the iPhone 6 there's no added value, given the amount of
other possible exploits.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20141007/a49923be/attachment.html>


More information about the cryptography mailing list