At last I said someting that is not totaly stupid after all ;-)
Good night Ben...
Nicolas
-Message d'origine-
De : Ben Laurie <[EMAIL PROTECTED]>
À : [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date : mardi 16 novembre 1999 18:30
Objet : Re: Compiling OpenSSL withou
Nicolas Roumiantzeff wrote:
> But this 2 key tripple DES (112 bit key length) would be the best solution
> regarding the French regulation:
> It would be not stronger than 128 bit and still currently secure (where as
> simple DES is not).
> It would aslo be as resistant as the 3 key tripple DES to
Hi Bruno,
>I've got another question about 3DES and SSL: isn't the SSL protocol
limited
>to a 128 bit keylength ? If this is true, how is 3DES handled ? Is the 3rd
>key only partially used ? Or is the "key1, key2, key1" scheme used ?
No, the "key1, key2, key1" scheme is not used for the 3DES cip
Nicolas Roumiantzeff wrote:
>
> But for the French authorities you would have to prove taht DES is a group.
>
> Beside, the 3-DES implementation used in SSL is not exactly a composition of
> 3 DES function (the initial and final seps of the DES algorithm are done
> only once instead of 3 times).
> But then, can't you just compile everything as normal, and
> change the allowed ciphersuites in the configuration..? It surely can't
> be illegal to compile the 3DES in, but simply not use it.
In fact, that's what I'll plead for, if there is no other simple solution.
It may even get acce
Nicolas Roumiantzeff wrote:
>
> Could you describe this "meet-in-the-middle" attack on the 3-DES?
OK, well, it's a known-plaintext attack. You encrypt the known plaintext
with all 2^56 possible keys for the first step, and store the results.
You then decrypt the ciphertext with all 2^112 possibl
Could you describe this "meet-in-the-middle" attack on the 3-DES?
Nicolas Roumiantzeff.
-Message d'origine-
De : Ben Laurie <[EMAIL PROTECTED]>
À : [EMAIL PROTECTED] <[EMAIL PROTECTED]>
Date : vendredi 12 novembre 1999 20:13
Objet : Re: Compiling Open
>The point is that _the French law_ says: "Thou shalt not use a keylength
>greater than 128 bits". There's no room for interpretation, here. Even if
>I invented my own cipher, no matter how rotten it might be (why not use
>XOR ? ;-)) ), I simply wouldn't be allowed to use a 129 bit key ! It's
Bruno Treguier wrote:
> I've got another question about 3DES and SSL: isn't the SSL protocol limited
> to a 128 bit keylength ? If this is true, how is 3DES handled ? Is the 3rd
> key only partially used ? Or is the "key1, key2, key1" scheme used ?
The simple answer is that SSL isn't limited to 1
But for the French authorities you would have to prove taht DES is a group.
Beside, the 3-DES implementation used in SSL is not exactly a composition of
3 DES function (the initial and final seps of the DES algorithm are done
only once instead of 3 times).
Nicolas Roumiantzeff.
>And whether 3-D
Ben:
> > Anyway, French laws aren't that specific. All they talk about is a
> > "key length", so even if you're right, Ben, I don't want to get into
> > trouble just because a pen pusher will have made the wrong assumption.
> > ;-)
>
> That's up to you, but I don't know _anyone_ who thinks that
Bruno Treguier wrote:
>
> Ben:
> > Is that true keylength or effective keylength? 3DES has an effective
> > keylength of 112 bits.
>
> Well, first of all I have to present my apologies to the list for my
> double posting the other day. Seems that I slipped on the "send" key
> before finalizing m
Date : jeudi 11 novembre 1999 16:08
Objet : Re: Compiling OpenSSL without 3DES
>Bruno Treguier wrote:
>>
>> Hello,
>>
>> I'm presently trying to get openssl (0.9.4) compiled without 3DES,
because
>> of French regulations which do not allow ciphers using a key
Ben:
> Is that true keylength or effective keylength? 3DES has an effective
> keylength of 112 bits.
Well, first of all I have to present my apologies to the list for my
double posting the other day. Seems that I slipped on the "send" key
before finalizing my message...
Now, about the effective
Bruno Treguier wrote:
>
> Hello,
>
> I'm presently trying to get openssl (0.9.4) compiled without 3DES, because
> of French regulations which do not allow ciphers using a key length of
> more than 128 bits. The simplest way seems to get rid of des, as it is
> supposed to be triggered by simple c
15 matches
Mail list logo
