<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">On 21 October 2015 at 04:59, Natanael <span dir="ltr"><<a href="mailto:natanael.l@gmail.com" target="_blank">natanael.l@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><span class=""><p dir="ltr">Den 20 okt 2015 19:01 skrev "Jeff Burdges" <<a href="mailto:burdges@gnunet.org" target="_blank">burdges@gnunet.org</a>>:<br>
> Formally, there should be a function KG(s) that returns a tuple<br>
> (k_1,..,k_n) and a function P(k,d) such that if d_i = P(k_i,d_{i-1})<br>
> then d_n=d_0 but there are no known relationships between strictly<br>
> fewer than n of the k_i.<br>
></p>
</span><p dir="ltr">Secret Sharing on the symmetric key? Do you need it to be integrated into the encryption algorithm itself? Your terminology is a bit unclear to me, what exactly are you trying to achieve?</p>
</blockquote><div>I think the request is for a keyed permutation P(k,d), and list of n independent keys. Such that applying all of the n keys returns the original plaintext, but that all the intermediate results (keys k_0 .. k_x for x < n) are confidential.<br><br></div><div>I feel things like this are easier served with encrypting with a random fixed length key K, then doing this n keys thing on the K instead of the whole plaintext.<br></div></div></div></div>