Thanks a lot for helping me out. Im going back to reading now :).

On Aug 4, 9:30 pm, "Wojciech S. Czarnecki" <[email protected]> wrote:
> Dnia Wed, Aug 04, 2010 at 08:48:48AM -0700, Vikas patial napisa (a):
>
> What you try to do is named Key Escrow.
> It is built into MS Windows since version 2000.
>
> > Thanks a lot guys for helping me ...
>
> > @ Wojciech
> > I went though the document and will use PBKDF1 from PKCS #5 for key
> > derivation,its also supported in cryptopp so its great.
>
>   Good.
>
> > I would seek if it is possible to do what i am trying to do with crypto
> > ++ in a better way,all ideas are welcome.
>
> > My task is to securely store some keys which my software uses on the
> > pc, so in case the Client forgets these keys ( he types them to access
> > data and can change them ) then a administrator can recover them.But
> > these keys should not be recoverable by anyone other than the admin.
>
>   Your admin has keypair generated in GnuPG. Public key of named admin
>   is um - public and is given to all clients. Your app should
>   encrypt to this public key data you want to escrow (your symmetric
>   key). Thats all. And all your client data depends of admin's honesty.
>
>   GnuPG for MS Win:
>  http://www.gpg4win.org
>
>   Library.
>  http://www.gnupg.org/download/index.en.html#gpgme
>
> > Symmetric Cryptography requires the key to be stored somewhere so i
> > choose Asymmetric cryptography ( RSA 2048 ) , where in the
> > administrator has the private key while all clients have public key
> > for encryption.But as i do not store the private key anywhere i needed
> > a deterministic way of generating it incase it is required again for
> > recovering the public key encrypted keys.
>
>   Deriving, testing, attacking and proving each schema or protocol
>   like PKCS#5 above took many years and minds of highly skilled 
> cryptographers.
>   There is NO place for homebrew solutions. Aka any "good idea of not
>   storing deterministic randomness" is doomed even before born.
>
>   Plain _using_ ready libraries need understanding how crypto
>   works. So read books you were given and dig the net for articles
>   about crypto. For now use simpliest but tested solutions.
>
> Pozdrawiam, Ohir.
>
> --
>
> Wojciech S. Czarnecki
>  << ^oo^ >> OHIR-RIPE

-- 
You received this message because you are subscribed to the "Crypto++ Users" 
Google Group.
To unsubscribe, send an email to [email protected].
More information about Crypto++ and this group is available at 
http://www.cryptopp.com.

Reply via email to