0

On Windows, CryptGenRandom is the standard random number generator to use. It is called by many packages like Python’s Random and Secrets modules, which both use os.urandom, which in turns calls CryptGenRandom.

For the algorithm of CryptGenRandom, I found out the following:

In Windows Vista with Service Pack 1 (SP1) and later, an implementation of the AES counter-mode based PRNG specified in NIST Special Publication 800-90 is used. In Windows Vista, Windows Storage Server 2003, and Windows XP, the PRNG specified in Federal Information Processing Standard (FIPS) 186-2 is used.

However, the NIST publication does not specify which entropy source is used.

In the case of my laptop, I have an Ideapad Gaming laptop by Lenovo, with an Intel(R) Core(TM) i5-10300H processor. On this laptop I have Windows 10 installed. The processor contains a RNG called Secure Key Technology. Is this used as entropy source by CryptGenRandom?

Peter Cordes
  • 328,167
  • 45
  • 605
  • 847
Riemann
  • 158
  • 10
  • The system RNG is probably the best to be relied upon either as random or to seed a CSPRNG within the application. If you use VM's or similar virtualization tech it makes sense to verify that the system RNG of the host is used - one of the reasons to always install the client utilities - although the presence of RDRAND and seeds like system time may be able to help there somewhat. – Maarten Bodewes Jul 19 '23 at 08:56

1 Answers1

2

The initial entropy sources include:

  • Seed file
  • External entropy
  • TPM randomness
  • RDRAND randomness (You mentioned Secure Key Technology-related instruction)
  • ACPI-OEM0 table
  • Output from the UEFI entropy provider
  • The current time

Windows 10 has many entropy sources; these work together to ensure that the OS has good entropy. Different entropy sources guarantee good entropy in different situations; by using them all the best coverage is attained.

From : The Windows 10 random number generation infrastructure

This whitepaper explores details about the Windows 10 pseudo-random number generator (PRNG) infrastructure and lists the primary RNG APIs. The whitepaper also explains how the entropy system works, what the entropy sources are, and how initial seeding works.

renzhexigua
  • 46
  • 4
  • 1
    Do remember that the working of any RNG is often system dependent. As the output is basically random bytes, the algorithms and entropy sources are generally not set in stone. There have been pretty big changes in the structure of Windows as well as Linux when it comes to implementations of `/dev/urandom`... – Maarten Bodewes Jul 19 '23 at 08:59