[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

RE: Authentication Mechanism Matrix (was L2TP vs IPSEC)



On Thu, 22 Jun 2000, David Jablon wrote:

> Some banter on what kind of authentication method can handle a
> small legacy-style shared secret password ...
> 
> At 04:44 PM 6/21/00 -0700, CHINNA N.R. PELLACURU wrote:
> > [IKE (?)] probably wasn't designed, but to make it secure in this scenario is
> > just a matter of implementation (IE. no change to bits on the wire).
> 
> Guess again.  My secret is "HRW#54".  That's all I can remember.
> I dare you to show me how to use IKE securely with this as the secret key.


HOW?! It is simple, however any good password based system would protect
such a low entropy secret, to the maximum possible extent. And however
that is done, can be done without any change to the IKE protocol. If you
still feel that this is not possible, then I can give you the gory
details.


> 
> ... then a diversion into smart-card-enabled kiosks ...
> 
> CHINNA:
> >Let's assume that, before I stick my smartcard into the machine, I make
> >sure that all the software running on the machine is cryptographically
> >singed by the vendor [...] and I 
> >can verify the signature, and thus make sure that there are no trojan
> >software in there. ... 
> 
> David:
> Amusing, but still irrelevant.
> 
> How do you get the SHA-1 hash of the software?
> Were you thinking of just asking the software to hand it over?
> What if it hands you the signed hash that was computed before the BIG
> EVIL PATCH?  As you should guess, I'm not looking for an answer.

You can't achive prefect security in anything without sacrificing an awful
amount of functionality, and resources. You can build a system that is
extremely restricted, and perfectly secure in this scenario. The way it is
dealt with in real life is by compensating reasonable security that could
be tampered with(with very less propability), by grave consequenses.
Everybody knows that the encryption used in cell phones is not very
secure, but how many acutally exploit it.

I guess you are looking for perfect security. None of the security
protocols/primitives that are widely used, have provable security, and so,
do you say that they are not perfect, and hence useless?!

> 
> ... and finally, a "definition" ...
> 
> David:
> >> Public/private keys and IKE shared secret authentication work just fine
> >> with large secret keys.  But for passwords, you need to use either:
> >> 
> >> 	(1) an authenticated tunnel secured by *something else*,
> >> or
> >> 	(2) a zero-knowledge password method which tolerates small keys directly.
> >> 
> >> Or better still, use both. :-)
> 
> CHINNA:
> >And so what is your cryptographic definition of "user authentication" Vs
> >"machine authentication". I don't think you have provided any.
> 
> Fine.  Here's my formal definition:
> 
> 	User authentication needs a secret.
> 	Machines can remember big ones,
> 	but people can't.
> 	Deal with it.
> 
> OK, so it's not formal, and not exactly Haiku, but it gets to the point.
> 

And this is the formal cryptographic definition?! I hope it is your own,
not accepted academically. Do you have any reference even vaguely hinting
in this direction.

    chinna


> ---------------------------------------------------
> David P. Jablon
> Integrity Sciences, Inc.
> dpj@xxxxxxxxxxxxxxxxxxxxx
> www.IntegritySciences.com
> 
> 

chinna narasimha reddy pellacuru
s/w engineer