[Cryptography] crypto standards and principles

[Tom Mitchell]

On Tue, Feb 3, 2015 at 2:46 PM, John Denker <jsd at av8n.com> wrote:

> On 02/03/2015 12:05 PM, Robert L Wilson wrote:
>
> > I was told years ago that the cryptography group within NSA had as a
> > standard that no crypto system could be considered secure unless

...snip good stuff......

>
> Be that as it may ... the NSA still keeps secret the workings
> of its own most-advanced systems.  AFAICT there are only two
> possible explanations for this:
>  1) They think their best system might get broken if the
>   adversaries found out how it worked.
>

What little I know is that there are more than two reasons.

With a good cypher it is difficult or impossible to know how the data was
obscured without being told.  Given this it is also nearly impossible to
know
externally that a method or key or both change.

Many times we talk about a key or method that had 2^64 or 2^1024
or pick a number.   When two methods are possible an attacker must
attack 2^65 different permutations or 2^1025.  Sometimes doubling
the effective key length doubles a search space.   Expanding the method
list to one of any vastly complicates an attack.   The users may know
their key but the method may be unknown to the users yet known and
controlled to the management.    This is in effect a multi part key.

A common phrase associated with redacted documents and litigation of
crimes involving secrets is that "methods and techniques" are important
and a vast amount of energy is spent to keep methods secret.


-- 
  T o m    M i t c h e l l
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20150203/cb83c556/attachment.html>