Implementation of symmetric order-preserving encryption scheme
Project description
This is an implementation of Boldyreva symmetric order-preserving encryption scheme (Boldyreva’s paper).
Disclaimer 1 This is a work in progress, which should be reviewed and evaluated before using in production and/or sensitive applications. If you have any concerns about used cryptographic primitives or specific implementation details, feel free to open a Github issue and we’ll discuss everything there.
Disclaimer 2 The Boldyreva scheme is not a standardized algorithm, so there are no test vectors and fixed plaintext-ciphertext mapping for a given key. It means that, generally speaking, a plaintext encrypted by two different versions of the package with the same key might not be equal to each other.
Quick examples
Quick start
from pyope.ope import OPE cipher = OPE(b'key goes here') assert cipher.encrypt(1000) < cipher.encrypt(2000) < cipher.encrypt(3000) assert cipher.decrypt(cipher.encrypt(1337)) == 1337
You can specify input and output ranges. Otherwise, default input (0..2^15-1) and output (0..2^31-1) ranges are used.
from pyope.ope import OPE, ValueRange cipher = OPE(b'long key' * 2, in_range=ValueRange(-100, 100), out_range=ValueRange(0, 9999)) assert 0 < cipher.encrypt(10) < cipher.encrypt(42) < 9999
About order-preserving encryption
Order-preserving encryption (OPE) allows to compare ciphertext values in order to learn the corresponding relation between the underlying plaintexts. The algorithm implemented in this package (Boldyreva’s method) is less secure than any deterministic encryption schemes or modes (such as ECB), because this scheme is deterministic by design (i.e., for a certain key equal plaintext are always mapped to a single ciphertext value).
How can OPE be useful? For example, some systems may need OPE to perform a certain set of queries (such as range SQL queries) over encrypted data. These systems include CryptDB and Monomi to name a few.
Security
As mentioned above, security guarantees for Boldyreva’s schema are weaker than those of deterministic encryption schemes, but security can still be improved if the encryption keys are long enough. It is advised to use randomly generated keys at least 128 bits long, with the optimal size being equal to 256 bits. Keys can be longer, but it won’t improve the overall security.
Running tests
PyTest is used as a test framework. Run all tests:
$ py.test tests/
TODO
More tests
Optimize speed
Security guarantees
Test on x86
History
0.1.0 (2017-01-01)
Use ‘cryptography’ instead of ‘pycrypto’
Project details
Download files
Download the file for your platform. If you're not sure which to choose, learn more about installing packages.
Source Distribution
Built Distribution
Hashes for pyope-0.1.1-py2.py3-none-any.whl
Algorithm | Hash digest | |
---|---|---|
SHA256 | b0cec5fc45063beac52bfb0b3b23f17b7450b6898c7870e479ff3e1d607af2b9 |
|
MD5 | 0583a02180f37f0ab402116fbee9efbf |
|
BLAKE2b-256 | bcb0271a020f2e88de41d256698cda9c8ca31752595240f558299fc37fac788b |