Re: new /dev/random

From: Bill Unruh (unruh_at_string.physics.ubc.ca)
Date: 09/25/04 

Date: 25 Sep 2004 01:14:00 GMT

"Joris Dobbelsteen" <joris.dobbelsteen@mail.com> writes:

]"Bill Unruh" <unruh@string.physics.ubc.ca> wrote in message
]news:cj0djh$dhi$1@nntp.itservices.ubc.ca...
]>
]> Yes, everyone failed to convince you that what was wanted for
]> /dev/random was an RNG not another PRNG.

]Indeed, /dev/random should be the RNG integrated into some Intel chipsets or
]the on-chip RNG on the current Via C3 processor.
]It will (most likely) be more widely accepted. And finally some hardware
]might be put to good use.

The current /dev/random uses timing info from the mouse, the disks, the
keyboard, etc to provide the the true random information for the RNG. And
it blocks (stops feeding out data) if it does not have enough randomness.
urandom is a PRNG fed from /dev/random for seeding and occasional injection
of true randomness. Our friend objects, and wants to convert /dev/random
into a PRNG for reasons he has failed to convince anyone but himself of.
He has a very weak understanding of the difference between an RNG and a
PRNG.

Relevant Pages

  • Re: new /dev/random
    ... ]> PRNG and an RNG and what the definition of infeasible was. ... ] You are talking to cryptographers. ... true RNG while /dev/urandom is supposed to be a PRNG seeded by a limited ...
    (sci.crypt)
  • Re: new /dev/random
    ... > PRNG and an RNG and what the definition of infeasible was. ... > sufficient infeasibility. ... You are talking to cryptographers. ...
    (sci.crypt)
  • Re: new /dev/random
    ... For a proper PRNG, with the assumption that the algorithms are robust, ... is said to contain 40 bits of entropy if I could, ... If I want to attack a stream of 56 bits produced by a PRNG with a seed ... RNG resistance therefore relies on the same two classes of assumptions ...
    (sci.crypt)
  • Re: new /dev/random
    ... etc to provide the true random information for the RNG. ... >>into a PRNG for reasons he has failed to convince anyone but himself of. ... with input from "entropy sources" as it goes, ... In the current /dev/random implementation there is an entropy estimator ...
    (sci.crypt)
  • Re: new /dev/random
    ... PRNG and an RNG and what the definition of infeasible was. ... a PRNG always is differentiable with infinite resources. ... in cryptography, is too vague to be usefull. ...
    (sci.crypt)
Quantcast