CryptGenRandom is a random number generator function in CryptoAPI in Windows. How much entropy has that random number generator ? I have already looked a lot, but I couldn't find it.
1 Answers
The exact algorithm of Windows CryptGenRandom was never published, therefore, some security experts suggest not to use it at all.
Some reverse-engineering and cryptanalysis was made. A published research (Cryptanalysis of the Windows Random Number Generator - Leo Dorrendorf, 2007) examined Windows 2000 RNG and found a number of weaknesses in the design and implementation.
The document also describes the entropy collector mechanism of the algorithm (section 4.2.3). The Entropy sources are:
Source Bytes requested
CircularHash 256
KSecDD 256
GetCurrentProcessID() 8
GetCurrentThreadID() 8
GetTickCount() 8
GetLocalTime() 16
QueryPerformanceCounter() 24
GlobalMemoryStatus() 16
GetDiskFreeSpace() 40
GetComputerName() 16
GetUserName() 257
GetCursorPos() 8
GetMessageTime() 16
NTQuerySystemInformation calls:
ProcessorTimes 48
Performance 312
Exception 16
Lookaside 32
ProcessorStatistics up to the remaining length (3584 bytes buffer)
ProcessesAndThreads up to the remaining length
Since we are talking about a pseudo-number-generator and not a real-number-generator, you may say that there is no real entropy at all, or count only the sources that you may consider as "entropy" (pseudo-entropy).
I was unable to find information regarding newer versions on Windows.

- 19,919
- 6
- 53
- 85
-
1
-
-
A presentation is also available: http://www.wisdom.weizmann.ac.il/~naor/COURSE/PRIVACY/pinkas_prg_insecurity.ppt – Lior Kogan Jan 29 '14 at 19:33
-
This is a very good answer! I really like your systematic approach. How can you determine the *bytes requested* for each algorithm? The *performance*'s inspection has a very high proportion in the overall entropy (312 bits). Can you explain why? - - It would love to get your systematic approach also for the image analysis generally because I cannot find any quality preview about any algorithm, etc here http://cs.stackexchange.com/q/4935/10350 – Léo Léopold Hertz 준영 Aug 09 '16 at 13:53