@Immutable
public interface Prf
- It is deterministic: PRF.compute(input, length) will always return the same output if the same key is used. PRF.compute(input, length1) will be a prefix of PRF.compute(input, length2) if length1 < length2 and the same key is used. - It is indistinguishable from a random function: Given the evaluation of n different inputs, an attacker cannot distinguish between the PRF and random bytes on an input different from the n that are known.
Use cases for PRF are deterministic redaction of PII, keyed hash functions, creating sub IDs that do not allow joining with the original dataset without knowing the key. While PRFs can be used in order to prove authenticity of a message, using the MAC interface is recommended for that use case, as it has support for verification, avoiding the security problems that often happen during verification. It also allows for non-deterministic MAC algorithms.
| Modifier and Type | Method and Description |
|---|---|
byte[] |
compute(byte[] input,
int outputLength)
Computes the PRF selected by the underlying key on input and returns the first outputLength
bytes.
|
byte[] compute(byte[] input,
int outputLength)
throws GeneralSecurityException
input - the input to compute the PRF on.outputLength - the desired length of the output in bytes. When choosing this parameter
keep the birthday paradox in mind. If you have 2^n different inputs that your system has to
handle set the output length to ceil(n/4 + 4) This corresponds to 2*n + 32 bits, meaning a
collision will occur with a probability less than 1:2^32. When in doubt, request a security
review.GeneralSecurityException - if the algorithm fails or if the output of algorithm is less
than outputLength.