Re: [Sci.nanotech] Encrypting Nanotechnology


In article <12ld19cpdk1d93@xxxxxxxxxxxxxxxxxx>,
<URL:mailto:John.S.Novak@xxxxxxxxx> wrote:

In article <12latacigiqha0c@xxxxxxxxxxxxxxxxxx>, rory@xxxxxxxxxxxxxxxxxx
says...
[snip]

I in no way guarantee the systems that I have sketched will
work, there is always the possibility that something has been
overlooked, but from what I have seen so far, of it and similar
systems over the last 15+ years, I feel it is worth further
investigation.

There are two problems, Rory.

The first is that cryptography is a discipline that is, by its nature,
extremely well-grounded in very rigourous, very specialized
mathematics... and it seems to be one of those disciplines where non-
experts very consistently underestimate the difficulty of the tasks.
I'm certainly no expert myself, but I've caught extremely smart people
make elementary mistakes like rolling their own random number generators
and assuming that's "good enough." (I never understood why-- there are
cryptography-strength public libraries available for that.) It's not.

The second is that one of the basic rules is to assume a system is
unsecure unless proven, mathematically and if necessary, physically
proven otherwise. STarting with an insecure system and not giving
enough information to analyze it leaves outside observers with no better
assumption to make than unreliability for your system.

If you have details that you think will stand up to scrutiny, please
post them.

I was a bit reluctant to say too much because of wishing to stay
'on topic', but there is an introduction to encryption, below.
This is intended to explain some of the logic of why it would be
relevant, in this case for communication in a nanotech system,
where I think the issue of meta data is likely most important.

A brief sketch of the encryption scheme that I mentioned, for
which I would be quite happy to use a verified replacement, with
similar useful characteristics, follows at the end.

---

Encrypting information offers two benefits:

-- Unauthorised individuals or systems (i.e. those not given the
decryption key) cannot read the information. There are lots of
arguments about privacy, but at its extreme, the secret ballot
(preventing the voter being coerced), is regarded as a good
thing. Means of personal identification like PIN numbers need to
be kept secret. There may be a human need for privacy that is
balanced against a need for openness, for the public good.

-- Information cannot be changed except by the author. Arguably
if there is change this should be done by publishing a new
version. A major reason for this is to ensure that meta data
remains bound to its data. Meta data such as: who is the author,
who is the intended recipient(s), and how the data is agreed to
be used. Other meta data will be a name for the data, and some
means to place it in sequence with other data. If any of this
meta data is disconnected from its data, or is changed, then the
data cannot be trusted.

Encryption makes use of secrets, which are used to create a key
to encrypt, and to decrypt. Preferably there should be no way of
obtaining these secrets except by being the author or the
intended recipient(s).

The ideal secret is a one-time pad of truly random numbers, which
is used once, and only once, to encrypt, with the recipient
having a copy of the pad, so they can decrypt. Unfortunately the
problems of generating and distributing these means that a
variety of methods are used to generate pseudo-random numbers
from a secret seed. The methods are chosen to attempt to ensure
that it is not possible to work-out the seed by examining the
encrypted information.

Some encryption makes use of the current difficulty of factoring
the product of large prime numbers, the prime numbers chosen
being the secret. However, it is believed that quantum computers
may make solving these problems a lot easier, which is one reason
that some prefer to avoid these, or not trust them for future
use.

Quite a bit of the science and technology associated with
encryption is the development of new methods and discovering
whether there are any weaknesses in existing methods. There is
also the question of how encryption is used, so as to make it
sufficiently difficult for unauthorised persons to compromise or
break the encryption.

Stream, rather than block, encryption may be appropriate for the
exchange of a series of messages, rather than a single, stand-
alone, message. One advantage of stream encryption is that it is
possible to hide where individual messages begin and end. This
can make decryption attempts far more difficult, and hides the
messages from traffic analysis, by ensuring that information is
exchanged even when no message needs to be sent.

There are a number of methods that may be used to make breaking
the encryption more difficult. These include compressing the
information sent, and building a dictionary from previous
messages, so that the same text need not be sent more than once;
the second time a dictionary reference is sent instead. The
dictionary references may also change over time, as more messages
are received. The aim is to reduce the redundancy in the
information as much as possible - redundancy and repeated
information provide opportunities for cracking the encryption.

---

The encryption technique is based on having a number of 'rotors'
which have respectively co-prime circumferences, and which are
populated with pseudo-random numbers, by a cryptographic-grade
random number generator, all the numbers on a given rotor being
different.

Each value to be encrypted is XOR-d with all the values on the
current rotor positions, and all the rotor positions are
incremented by one. Decryption is done similarly.

The encryption key is the description of the rotors, and their
start positions. A given key will not be used for long enough to
give even the next anticipated generation of machines enough time
to break it.


This system is based on the work of Dr. Ian Newman, of
Loughborough University, UK, from the late 1980s onwards.

--
Rory McLean
rory@xxxxxxxxxxxxxxxxxx


.

Relevant Pages

  • Re: National Security Backdoor in telnetd - all versions.
    ... >>against the Secret Service for their violations of civil rights. ... encryption techniques to protect critical resources. ... plants have absolutely horrid security resulting from these stupidities. ... Of course I doubt you have an NDA with the government - so ...
    (comp.os.linux.security)
  • Re: Name this key exchange
    ... RSA is faster than DH at encryption and slower at decryption. ... Bob generates a secret by choosing a random number r and setting ...
    (sci.crypt)
  • Re: UserNameTokens + Password.SendNone Implimentation
    ... That is the one weakness in this system is that once the secret ... > AES step. ... The idea using the AES encryption of the data is that if they decrypt the ...
    (microsoft.public.dotnet.framework.webservices.enhancements)
  • Re: Securing data to a process principal
    ... encryption key first time for the user - and use it later). ... secret. ... I need the decryption to ... You MAY think that instead of a filter driver you can simply ...
    (microsoft.public.platformsdk.security)
  • Re: Securing data to a process principal
    ... that did the encryption. ... pOptionalEntropy (or more broadly, "secret") in the apps. ... You MAY think that instead of a filter driver you can simply ...
    (microsoft.public.platformsdk.security)
Quantcast