Is there a difference between PRF and a hash function?
For example: Creation of a secret key is using PRF and creating a secret key is using hash function.
10
$\begingroup$
$\endgroup$
-
$\begingroup$ You should not be using a PRF to create secret keys. You should be using an RNG or CSPRNG, which may be built out of a PRG. $\endgroup$ – Stephen Touset May 1 '14 at 20:24
-
$\begingroup$ In the first phase of IKE protocol, the pseudo-random function is used to generate the secret keys. $\endgroup$ – user13342 May 1 '14 at 22:21
10
$\begingroup$
$\endgroup$
No, the two primitives are definitely not the same.
A pseudorandom function is a keyed function that is (computationally) indistinguishable from a function chosen at random from all functions with matching domain and range as long as the key remains secret.
On the other hand, a cryptographic hash function is a function with either a publicly known key (in theory) or no key at all (in practice) that is compressing and collision resistant. That means, it maps long inputs to outputs of a fixed length and it is hard to find two inputs that map to the same output.
-
$\begingroup$ i thought hash functions resemble to a random oracle (in theory), why would they have a public key in theory? $\endgroup$ – Subhayan May 2 '14 at 23:28
-
1$\begingroup$ No, hash functions are only modeled as random oracles for proofs in the random oracle model. If we work in the standard model, a hash function is only assumed to be collision resistant. The key of a hash function stems from a technicality of defining collision resistance. The problem is that it is (almost) impossible to define collision resistance for a single function. Therefore collision resistance is defined for a function family, and the key is required to select a member of that family. $\endgroup$ – Maeher May 3 '14 at 8:20
