sci.crypt
  Home FAQ Contact Sign in
sci.crypt only
 
Advanced search
December 2006
motuwethfrsasuw
    123 48
45678910 49
11121314151617 50
18192021222324 51
25262728293031 52
2006
 Jan   Feb   Mar   Apr 
 May   Jun   Jul   Aug 
 Sep   Oct   Nov   Dec 
2008 2007 2006  
total
sci.crypt Profile…
RELATED GROUPS

POPULAR GROUPS
 Up
  Dividing 512 bit number by 128 bit number in C program         


Author: vaidehikedlaya
Date: Dec 22, 2006 15:11

Hi all,
I have a huge problem finding " a mod m " using C program, when m is
128 bit number and a is 512 bit number. Can anybody suggest me a
solution.
Thanks, in advance.
5 Comments
  Re: RSA Performance         


Author: Le Chaud Lapin
Date: Dec 22, 2006 13:08

Carlos Moreno wrote:
> Le Chaud Lapin wrote:
>
>> So essentially the requirements are:
>>
>> 1. Alice signs every message. Bob verifies every signature of Alice.
>> 2. Message flow is strictly unidirectional, from Alice to Bob.
>> 3. Some messages may be lost in transit.
>> 4. Flow must be real-time (hence minimization of exponents on both
>> sides in RSA)
>
> Just a thought ... With so much *symmetry* requirements, how come
> you haven't considered *symmetric* encryption?
>
>
> If you use RSA only to arrange a shared secret between Alice and
> Bob (the shared secret would be nothing other than the encryption
> key to be used with the symmetric algorithm --- the so-called
> "session key"), then the authenticity of the rest of the packages
> is guaranteed by the fact that you can decrypt them and read ...
Show full article (1.21Kb)
no comments
  Re: RSA Performance         


Author: Le Chaud Lapin
Date: Dec 22, 2006 11:07

Mike Amling wrote:
> What patent dispute is that? The RSA patent expired years ago. DSA,
> to my knowledge, was never patented.
>> also need to learn more about DSA. I had read somewhere a while back
>> that signing and verification are asymmetric in performance. I need to
>> be able to induce symmetry to maximize throughput.
>
> How does symmetry affect throughput? If each signature is verified
> once, don't you want to minimize the sum of signature generation and
> verification?
Show full article (2.00Kb)
1 Comment
  Please Help: x.509 Certificates Library or non-certificate authentication         


Author: Jeff
Date: Dec 22, 2006 10:03

Hello

We are looking for a STANDALONE x.509 Certificates Library
written in ANSI-C (not C++) with BSD or similar license.
We need this for an embedded system - so size may also
be a consideration.

Alternatively, is there any industry-standard way to
authenticate an unknown 3rd party, if certificates
are not an option (and obvsiouly 3rd party's public
key is unknown to the embedded system)

Thanks,

Jeff
11 Comments
  Do Gap-CDH groups exist?         


Author: amitabh123
Date: Dec 22, 2006 09:07

Informally define a Gap-CDH group as a cyclic multiplicative group (of
some order n) such that the DL (Discrete Log) problem w.r.t. some
generator g is hard but the CDH (computational Diffie-Hellman) problem
w.r.t. g is easy.

Suppose we need a Gap-CDH group of composite order n with *large* prime
factors.
The literature seems to indicate a strong possibility of such groups.
For instance:

1) Maurer (Crypto 94) proved that for any *prime* n, if a smooth
elliptic curve of order n is known, then the CDH and DL problems in all
groups of order n are equivalent.

2) However, it is shown in Theorem 7, "lower bounds on generic
algorithms in groups" (Maurer-Wolf - Eurocrypt'98) that the DL problem
is *not* computationally equivalent to the CDH problem if the order of
the group contains large prime factors. In fact using a DH oracle to
solve DL is as time consuming as the baby-step-giant-step method.

Therefore, the above results still leave open the possibility of
Gap-CDH groups of composite order n with large prime factors, which are
exactly the ones desired. Do such exist?
14 Comments
  Re: Can we all talk about the elephant in the room?         


Author: David Eather
Date: Dec 22, 2006 04:08

vedaal wrote:
> David Eather wrote:
>
>> LTP got me thinking about the nature of the problem and how there is no
>> real fix.
>>
>> The problem is the posts of the criminal nature did not occur on
>> sci.crypt. So we could have someone who has ID'ed himself, plays by the
>> rules here, and then out of malice/spite/stupidity anonymously posts to
>> a few other news groups the sort of trash that has caused this problem.
>>
>> back to square one.
>
> but if it is known or suspected that one one the posters here did such
> a thing,
> and some sort of identity verification were required here,
>
> then that poster could be traced,
> (not, by you, me, or anyone here,)
> but by law enforcement personnel who have the authoriity to obtain such ...
Show full article (1.15Kb)
no comments