review of RNGs
ianG
iang at iang.org
Mon Jul 6 10:03:59 BST 2015
http://www.cacert.at/random/
Random Number Generator Analysis
This is a project to make statistics about all random number generators
on the market.
We want to collect information on all kinds of random number generators
out there, eg. Hardware RNGs, Crypto PRNGs, encoding software,
compression software, ... We are analysing all of them, and we are
providing a report about the results here.
FAQ: Why do you need this?
We want to get a market overview of the whole random number
generator market
If I need X Random numbers per second from a good Hardware Random
number generator, how much do I have to to pay, and where do I get it?
We want to compare hardware random number generators against crypto
pseudo random generators
We want to have a randomness classification system, where we can
analyse unknown random numbers, and be able to compare the quality with
known hardware and crypto generators
We want to run this as a service, so that everyone can easily have
their own random data tested and compared against the best, in an easy
and objective test
FAQ: How can I help?
If you have a random generator yourself, please extract minimum 12
MB of randomness (or much more, if possible, maximum 1GB), and submit it
here.
If you know someone else, who has a random number generator, please
tell him/her about the project.
FAQ: How are the random numbers being analyzed?
At the moment, we are using ENT and diehard for analysis. ENT is
used to sort all results, and diehard is used to raise red flags. We are
trying to switch to dieharder which promises better tests than diehard.
FAQ: Which format do I have to save the random numbers in?
Please save your binary (non-decimal!) random numbers into a raw
binary file, make sure that you do not add CRLF line-endings.
FAQ: What was the impact of this project?
We discovered predictable random number generators, contacted the
vendors and got them to fix their products.
If you want to contact me: pg @ futureware . at
More information about the Discuss
mailing list