>From Mark Shepard <[EMAIL PROTECTED]>

>My understanding of Freenet was that data stored at a node was encrypted so
>that node operators could neither censor nor be held accountable for the
>data.  Yet I was just reading
>http://freenetproject.org/twiki/Main/Papers/ieee-final.pdf and the paper
>seems to say that data encryption _in_the_store_ is something the user does
>before inserting data, and is not part of Freenet itself.  True?  So, with
>the current implementation, if I run a Freenet node, I could end up with a
>bunch of "illegal" material on my hard drive, since the decision to encrypt
>is up to the inserter.  Is this the case?  Will this always be the case, or
>is there a plan to change this behavior?  Thanks.
>         Mark

There are two levels of encryption that are relevant here; You have the 
(optional) encryption of the datastore, where the data your node sees is 
encrypted when it is written to the hard drive so that it is possible to 
destroy any information stored there easily by destroying the key (how useful 
this is is debatable)

Secondly, data inserted into freenet is encrypted by the inserter and 
decrypted by the reciever.  Your node doesn't get to see the encryption key 
for that level of encryption, so it can't see the data.  While it may be 
possible to insert data which is not encrypted (which you could then read out 
of your datastore if you chose to have an uncrypted datastore), you wouldn't 
be able to retreive and view that data with a normal client because it would 
be expecting to get encrypted data--decrypting plaintext gives you garbage.

I suppose it might be possible to write specialized clients to insert and 
retrieve plaintext data over the existing freenet network, although there 
wouldn't be much of a point except to annoy people who look at their 
unencrypted datastores :)

--
Benjamin Coates


_______________________________________________
Tech mailing list
[EMAIL PROTECTED]
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/tech

Reply via email to