hi,

i have to use RSA within .net 2.0 and i have to use a 256 bit key-
length.
RSACryptoServiceProvider requires a min. length of 384 bit.

my current solution-ideas are:
s1.) use the RSACryptoServiceProvider and bypass the min. length with
reflections are some similar hack.
s2.) use a third party lib for .net (like the lib from http://www.certicom.com/,
any others? )
s3.) use a lib written for another language (probably c) -> make a dll
and call that dll from .net or try to port the c code to managed c++

the solution should be reliable, fast (but no need for super-
performance) and idealy easy to implement (therefore i belive s2 seems
most reasonable).

thanks for any kind of hint, pingram

You are misunderstanding how RSA encryption works. RSA is only used to
encrypt a symmetric session key. Your requirement for a 256 bit key
probably refers to the symmetric session key. A 256 bit RSA key makes no
sense, as you should be using an RSA key that is at least 1024 bits.

Joe K.

--
Joe Kaplan-MS MVP Directory Services Programming
Co-author of "The .NET Developer's Guide to Directory Services Programming"
http://www.directoryprogramming.net
--
<wip (AT) pdi (DOT) at> wrote

joe,

thanks for your input.

basically i agree that a 256 bit key for encryption doesn't make much
sense.

unfortunatlly the key length really has to be 256 bit - i'd be happy
to change this specification but that's not an option (btw: in this
case we use rsa for digital signature (not encryption) and chances of
a collision actually are quite low).

any hints welcome!

br, pingram.


On 9 Mrz., 16:22, "Joe Kaplan"
<joseph.e.kap... (AT) removethis (DOT) accenture.com> wrote:

I don't think you understood what I said. When you are doing RSA
encryption, that is generally ALWAYS combined with an additional symmetric
encryption algorithm like 3DES or AES. RSA is just used for encrypting the
symmetric session key. The bulk encryption of data is done with the
symmetric algorithm and key. This is how protocols like SSL and SMIME email
encryption work.

A 256 bit session key makes perfect sense if you are using AES/Rjindael, so
that's probably what the spec is suggesting. If they are really saying the
RSA key needs to be 256 bit, then the spec itself doesn't understand how
crypto works and needs to go back to the author.

Joe K.

--
Joe Kaplan-MS MVP Directory Services Programming
Co-author of "The .NET Developer's Guide to Directory Services Programming"
http://www.directoryprogramming.net
--
<wip (AT) pdi (DOT) at> wrote

joe,

yes, symetric encryption will be done with 3DES; nevertheless the 256
bit RSA key length is mandatory and changing this spec is not
possible!

br, pingram.



On Mar 9, 6:41 pm, "Joe Kaplan"
<joseph.e.kap... (AT) removethis (DOT) accenture.com> wrote:

finally - valery jumps in

I was missing you Dude!

-----
Dominick Baier (http://www.leastprivilege.com)

Developing More Secure Microsoft ASP.NET 2.0 Applications (http://www.microsoft.com/mspress/books/9989.asp)

On Mar 12, 11:34 am, w... (AT) pdi (DOT) at wrote:
Do you know that it takes 20 seconds to factor 256 bit integer on a
low end tablet? And when I tried that om 4-way 64 bit server box it
took less than a second for that.
Even One Laptop Per Child will factor it in one minute. What could be
the reason to have a private key that everyone can find from your
public key in less than a minute?
Unless you are mistaken and you really talking about some other keys
(i.e. not RSA but for ex. ECC or AES)

-Valery
http://www.harper.no/valery

On Mar 12, 12:27 pm, Dominick Baier
<dbaier (AT) pleasepleasenospam_leastprivilege (DOT) com> wrote:
Thanks Dominick!


-Valery
http://www.harper.no/valery

Yes, definitely nice to see you back here again.

Joe K.

--
Joe Kaplan-MS MVP Directory Services Programming
Co-author of "The .NET Developer's Guide to Directory Services Programming"
http://www.directoryprogramming.net
--
"Valery Pryamikov" <valery (AT) harper (DOT) no> wrote

It sounds like you are out of luck then. Like Valery said, this key is so
insecure as to be useless and is not supported by .NET cryptography. I'm
not sure what options you have. Perhaps an alternate crypto API for .NET
supports such weak keys.

I'd suggest pushing back on the spec or looking for different options.

Joe K.

--
Joe Kaplan-MS MVP Directory Services Programming
Co-author of "The .NET Developer's Guide to Directory Services Programming"
http://www.directoryprogramming.net
--
<wip (AT) pdi (DOT) at> wrote