New "Jitter Entropy" RNG Proposed For Linux

[phoronix]

Phoronix: New "Jitter Entropy" RNG Proposed For Linux

The Linux kernel RNG implementation for providing random numbers has worked quite well for its years in use. However, a new jitter entropy generator implementation has been proposed that is capable of providing 100 kB/s throughput of randomness...

http://www.phoronix.com/vr.php?view=MTI5NzY

[uid313]

Doesn't all new CPUs come with a built-in hardware (P)RNG?

[Ex-Cyber]

Doesn't all new CPUs come with a built-in hardware (P)RNG?

A PRNG doesn't actually provide entropy.

[LightBit]

Doesn't all new CPUs come with a built-in hardware (P)RNG?

http://en.wikipedia.org/wiki/RdRand (uses RNG to seed PRNG)
http://en.wikipedia.org/wiki/Hardwar...or#Clock_drift (VIA C3)

[RealNC]

A PRNG doesn't actually provide entropy.

Are you sure?

[Kristian Joensen]

A PRNG doesn't actually provide entropy.

For someone less technically inclined. What does this mean? What is entropy in this context and how does it differ from what a PRNG does? How does the proposed solution fix/handle that? Isn't what is being proposed here a PRNG, just a different one?

[przemoli]

More entropy the better when you really want Random numbers.

[LightBit]

Entropy is something that is unpredictable, hard to guess.
PRNGs are algorithms that need to be seeded (keyed) with some entropy before they can output anything. Simple example of PRNG: http://en.wikipedia.org/wiki/RC4
(T)RNGs get entropy from physical properties, like thermal noise, nuclear decay or in this case CPU timing jitter.
(T)RNG is usually used to seed PRNG.
Both types output statistical random data.

[Kristian Joensen]

Entropy is something that is unpredictable, hard to guess.
PRNGs are algorithms that need to be seeded (keyed) with some entropy before they can output anything. Simple example of PRNG: http://en.wikipedia.org/wiki/RC4
(T)RNGs get entropy from physical properties, like thermal noise, nuclear decay or in this case CPU timing jitter.
(T)RNG is usually used to seed PRNG.
Both types output statistical random data.

I was under the impression that true random number generators couldn't be implemented in software due to computers being 100% deterministic, was I wrong?

[LightBit]

I was under the impression that true random number generators couldn't be implemented in software due to computers being 100% deterministic, was I wrong?

No, you are not wrong. This proposal collects entropy from hardware.