Re: Ultra-Fast Stateless Forward Signing

Luc The Perverse wrote:=>
With a name like "The Horny Rabbit" we didn't think so! j/k

I am curious what your background is though - because even though I didn't
know any answers, your "wording" seems intuitive in retrospect

:D

There is actually a movie called "Le Chaud Lapin" made in France
(http://www.imdb.com/title/tt0071305/) . I got this nickname the first
time I was in France and I found a dirty (literally) movie ticket stub
on the ground and read it in front of several lovely young french
girls. And so it was.

I am an electrical/software engineer who got the idea of rewriting
TCP/IP from scratch. At first it started off as a "What If?" and
before long I became obsessed. My dream is to finally settle an
argument started long ago, which is somewhat obvious now, that IPv6
will never achieve what its designers hoped because it was concocted by
committee. To do TCP/IP right, there needs to be structural integrity,
and the designers need to be brutally honest with themselves about the
virtue of whatever they concoct. After all, 1 billion people will want
to use what you make, and 1 billion people will demonstrate with the
crucible of life whether what you have is of virtue or not.

The protocol stack has security built in, so that programmers can twist
a few knobs on a C++ object to get the security characteristics they
want (independent signing or veiling). One thing I did not do that the
IPSec people did was whore my design to every conceivable scenario.
There are no options for selecting multiple symmetric ciphers or
asymmetric ciphers or key exchange schemes. Instead, I have tried to
define a hard framework, and what is there, so shall it be, which is
why I need experts like those that lurk in this group (eventually).
The drawback to this philosophy is that, if it turns out that one of
the primitives is breakable, everyone who uses your stack is
immediately at risk with no apparent recourse. I have determined that
this is not as big a problem as might first seem. I would rather have
a one-size-fits-all-but-change-it-if-its-broken model than a
here-are-19-ways-of-doing-it-mix-it-all-up-and-make-a-mess model.

The protocol stack is full-featured, meaning that, if you list all the
major problems that plagued IPv4 and was addressed by IPv6 (no pun
intended) , I have tried to include solutions to them. I saved the
problems of mobility and security for last. It turns out that security
is simpler than mobility, as the algorithms for security are entirely
deterministic, and if there is a bug somewhere, it is easily found.
The mobility problem, OTOH, is ...well, lets just say no one can say I
did not suffer for my art. :)

The good news is that the primitives that I am using now, SHA-256,
Rijndael, and RSA, are all implemented as C++ classes. As big as the
stack is, it really is a matter of taking out the, say, Rijndael class
and substituting another symmetric cipher, recompiling, and everything
throughout the stack will be modified to the new cipher accordingly.

For many reasons, I concluded that stateless forward signing was a
necessity, whether fast or slow. I am hoping there is some way to make
it fast. I still have to read the literature that Dave Wagner put
forth.

-Le Chaud Lapin-

.

Relevant Pages

  • [REVS] Writing Buffer Overflow Exploits - a Tutorial for Beginners
    ... The following security advisory is sent to the securiteam mailing list, and can be found at the SecuriTeam web site: http://www.securiteam.com ... Buffer overflows in user input dependent buffers have become one of the ... The bottom of the stack ... To keep it simple, shellcode is simply assembler commands, which we write ...
    (Securiteam)
  • Re: [Full-Disclosure] Re: Buffer overflow prevention
    ... his code cant run off your stack. ... It does provide some form security, but only a really tiny amount. ... You can download it from the PaX site ... I think there are several reasons why the major Linux ...
    (Full-Disclosure)
  • Re: Intermittent Net Access And Zonealarm
    ... Installing a personal firewall requires rigid ... Even one such change might open new security ... >>holes in the stack itself or in networking applications. ...
    (comp.security.firewalls)
  • Re: unspected shutdown. TCP attack?
    ... original debugging environment and see what is causing that problem. ... be your program has a bug that is bringing down the tcp/ip stack and causing ... >> I'll also hit one thing you may not have checked, in your local security ...
    (microsoft.public.win2000.networking)
  • Re: unspected shutdown. TCP attack?
    ... original debugging environment and see what is causing that problem. ... be your program has a bug that is bringing down the tcp/ip stack and causing ... >> I'll also hit one thing you may not have checked, in your local security ...
    (microsoft.public.win2000.general)