Crates.io | cloudproof_fpe |
lib.rs | cloudproof_fpe |
version | 0.2.2 |
source | src |
created_at | 2023-04-26 09:20:37.124951 |
updated_at | 2023-12-08 13:06:06.683297 |
description | Cosmian Cloudproof FPE library |
homepage | |
repository | https://github.com/cosmian/cloudproof_rust/ |
max_upload_size | |
id | 849205 |
size | 381,572 |
This library provides Format Preserving Encryption
(FPE) techniques for use in a zero-trust environment. These techniques are based on FPE-FF1 which is described in NIST:800-38G.
FPE aims to encrypt plaintext while retaining its format (alphabet). FPE-FF1 is a normalized algorithm that uses symmetric encryption, but it's not as fast or secure as standardized symmetric (or public key) encryption methods like AES or ChaCha. It should only be used where the format of the ciphertext container is constrained (e.g., a fixed database schema that cannot be changed).
The FPE implementation follows NIST specifications for FF1 (found in the NIST SP 800-38G specification).
The code is based on the cosmian_fpe
directory found on GitHub, which is based on str4d/fpe
. The number of Feistel rounds has been increased to 18 following the recommendations of this cryptanalysis paper.
The implementation also enforces the requirement that radix^min_len > 1_000_000
. For the Alphabet
and Integer
FPE facilities, this requirement is met with the following parameters:
radix | example alphabet | min text len |
---|---|---|
2 | "01" | 20 |
10 | "01234567890" | 6 |
16 | "01234567890abcdef" | 5 |
Cosmian FPE proposes 3 structures:
fpe::Alphabet
to encrypt textfpe::Integer
to encrypt integers with various radixesfpe::Float
to encrypt floating numbersThe fpe::Alphabet
structure provides the ability to encrypt a plaintext using an alphabet
.
Characters of the plaintext that belong to the alphabet are encrypted while the others are left unchanged at their original location in the ciphertext.
An alphabet can be instantiated using the Alphabet::instantiate()
method:
let hexadecimal_alphabet = Alphabet::instantiate("01234567890abcdef").unwrap();
There are multiple pre-defined alphabets available:
Alphabet::alpha()
Alphabet::alpha_lower()
Alphabet::alpha_upper()
Alphabet::numeric()
Alphabet::hexa_decimal()
Alphabet::alpha_numeric()
Alphabet::chinese()
Alphabet::latin1sup()
Alphabet::latin1sup_alphanum()
These alphabets can easily be extended using the extend_with
method
//0-9a-zA-Z
let mut alphabet = Alphabet::alphanumeric();
// add the space character
alphabet.extend_with(" ");
let key = [0_u8; 32];
let tweak = b"unique tweak";
let alphabet = Alphabet::alpha_numeric(); //0-9a-zA-Z
let ciphertext = alphabet.encrypt(&key, tweak, "alphanumeric").unwrap();
assert_eq!("jraqSuFWZmdH", ciphertext);
let plaintext = alphabet.decrypt(&key, tweak, &ciphertext).unwrap();
assert_eq!("alphanumeric", plaintext);
let key = [0_u8; 32];
let tweak = b"unique tweak";
let alphabet = Alphabet::numeric(); //0-9
let ciphertext = alphabet
.encrypt(&key, tweak, "1234-1234-1234-1234")
.unwrap();
assert_eq!("1415-4650-5562-7272", ciphertext);
let plaintext = alphabet.decrypt(&key, tweak, &ciphertext).unwrap();
assert_eq!("1234-1234-1234-1234", plaintext);
Note: since the -
character is not part of the alphabet it is preserved during encryption and decryption.
let key = [0_u8; 32];
let tweak = b"unique tweak";
let mut alphabet = Alphabet::chinese();
// add the space character to the alphabet
alphabet.extend_with(" ");
let ciphertext = alphabet.encrypt(&key, tweak, "天地玄黄 宇宙洪荒").unwrap();
assert_eq!("儖濣鈍媺惐墷礿截媃", ciphertext);
let plaintext = alphabet.decrypt(&key, tweak, &ciphertext).unwrap();
assert_eq!("天地玄黄 宇宙洪荒", plaintext);
Note: since the space character was added to the alphabet, it is also encrypted.
The fpe::Integer
structure offers the ability to encrypt integers with a radix between 2 (binary) and 16 (hexadecimal) and up to a maximum power of this radix.
To encrypt decimal integers up to u64::MAX, use:
let key = [0_u8; 32];
let tweak = b"unique tweak";
// decimal number with digits 0-9
let radix = 10_u32;
// the number of digits of the biggest number = radix^digits -1
// In this case 6 decimal digits -> 999_999
let digits = 6;
let itg = Integer::instantiate(radix, digits).unwrap();
let ciphertext = itg.encrypt(&key, tweak, 123_456_u64).unwrap();
assert_eq!(110_655_u64, ciphertext);
let plaintext = itg.decrypt(&key, tweak, ciphertext).unwrap();
assert_eq!(123_456_u64, plaintext);
There is also support for Big Unsigned Integers
let key = [0_u8; 32];
let tweak = b"unique tweak";
// decimal number with digits 0-9
let radix = 10_u32;
// the number of digits of the greatest number = radix^digits -1 = 10^20-1
let digits = 20;
// the value to encrypt: 10^17
let value = BigUint::from_str_radix("100000000000000000", radix).unwrap();
let itg = Integer::instantiate(radix, digits).unwrap();
let ciphertext = itg.encrypt_big(&key, tweak, &value).unwrap();
assert_eq!(
BigUint::from_str_radix("65348521845006160218", radix).unwrap(),
ciphertext
);
let plaintext = itg.decrypt_big(&key, tweak, &ciphertext).unwrap();
assert_eq!(
BigUint::from_str_radix("100000000000000000", radix).unwrap(),
plaintext
);
The fpe::Float
structure provides support for encrypting floats of type f64
:
let key = [0_u8; 32];
let tweak = b"unique tweak";
let flt = Float::instantiate().unwrap();
let ciphertext = flt.encrypt(&key, tweak, 123_456.789_f64).unwrap();
assert_eq!(1.170438892319619e91_f64, ciphertext);
let plaintext = flt.decrypt(&key, tweak, ciphertext).unwrap();
assert_eq!(123_456.789_f64, plaintext);
Tweaks
are public parameters that should vary with each instance of the encryption whenever possible. Tweaks
are described in NIST:800-38G: Appendix C. There is no size limit for the tweak
.
Run cargo criterion
from the current directory (./crates/fpe).
Install criterion and criterion-table
cargo install cargo-criterion
cargo install criterion-table
From the root of the project, run
bash ./benches/benches.sh
The benchmarks are then available in ./benches/BENCHMARKS.md