Laszlo,

Are you worried about the fact that XML parsing is needed to export/import
the key? or bothered by the way the key material is encrypted?
For the latter, the P1619-D4 document suggests to encrypt the key using
standard XML encryption methods; this translates most likely to using AES,
which  will be an existing tool on a drive that uses LRW-AES.
For the former, I can envision that parsing will be done on a different
computer but the encrypted key material is transferred into/out of the
drive.
The original intention for the key export format was to allow it to be
stored elsewhere, not on the disk itself.

Dalit.




                                                                           
             [EMAIL PROTECTED]                                                
             Sent by:                                                      
             [EMAIL PROTECTED]                                          To 
             RG                        [EMAIL PROTECTED]                 
                                                                        cc 
                                                                           
             31/03/2006 01:54                                      Subject 
             AM                        P1619-disk: key export w/o          
                                       encrypting keys                     
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           
                                                                           




The last paragraph of P1619-D4 document says, that the "LRW-AES
transform should not be used to encrypt its own secret-key". (I am not
aware of this restriction for CBC, ECB or counter mode. Which other
encryption modes have this requirement?)

With the key export procedure, this represents a serious limitation. It
becomes forbidden to export/archive keys with ordinary tools in a
computer, with a single encrypting disk drive. When the XML document is
created, the OS could swap the memory to disk, or the editor could save
temporary copies of the document, which contains the keys in the clear.
On the disk these create forbidden encrypted copies of the key.

These necessitate the use of specially designed editors, special OS
versions (like DOS, which don't swap data to disk) or creating the XML
file on another computer (and securely transferring the key there). I
understand, that this issue is not to be covered by the standard, but
we have to know, that the problem can be solved with reasonable
complexity and costs, otherwise we will not be able to manufacture and
sell secure LRW storage for the mass market. What was the envisioned
procedure? A laptop user has to boot from a CD, specially crafted for
his computer (with all the necessary device drivers), and run an XML
editor from the CD, which only saves temporary files encrypted with an
independent key?

There must be something simpler, because the above procedure is too
complicated: the user has to burn a special CD with his particular set
of device drivers (all a huge set to cover all the cases), boot from
this CD, activate the key derivation/extraction procedure to get the
key into the document, figure out the scope of the key, etc., enter a
high entropy XML encryption key, and then save the encrypted file on
the disk. He also has to remember never to decrypt/view the encrypted
XML file in his computer under normal working mode, when the OS could
swap plaintext data to disk.

Laszlo

Reply via email to