Microsoft SEAL explained
Simple Encrypted Arithmetic Library or SEAL is a free and open-source cross platform software library developed by Microsoft Research that implements various forms of homomorphic encryption.[1]
History
Development originally came out of the Cryptonets paper,[2] demonstrating that artificial intelligence algorithms could be run on homomorphically encrypted data.[3]
It is open-source (under the MIT License) and written in standard C++ without external dependencies and so it can be compiled cross platform. An official .NET wrapper written in C# is available and makes it easier for .NET applications to interact with SEAL.
Features
Algorithms
Microsoft SEAL supports both asymmetric and symmetric (added in version 3.4) encryption algorithms.
Scheme types
Microsoft SEAL comes with two different homomorphic encryption schemes with very different properties:
- BFV:[4] The BFV scheme allows modular arithmetic to be performed on encrypted integers. For applications where exact values are necessary, the BFV scheme is the only choice.
- CKKS:[5] The CKKS scheme allows additions and multiplications on encrypted real or complex numbers, but yields only approximate results. In applications such as summing up encrypted real numbers, evaluating machine learning models on encrypted data, or computing distances of encrypted locations CKKS is going to be by far the best choice.
Compression
Data compression can be achieved by building SEAL with Zlib support. By default, data is compressed using the DEFLATE algorithm which achieves significant memory footprint savings when serializing objects such as encryption parameters, ciphertexts, plaintexts, and all available keys: Public, Secret, Relin (relinearization), and Galois. Compression can always be disabled.
Availability
There are several known ports of SEAL to other languages in active development:
C++
C#/F#
- NuGet (Microsoft's official package)
Python
JavaScript
TypeScript
Notes and References
- Book: Advances in Cryptology – EUROCRYPT 2017: 36th Annual International Conference on the Theory and Applications of Cryptographic Techniques, Paris, France, April 30 – May 4, 2017, Proceedings. Coron. Jean-Sébastien. Nielsen. Jesper Buus. 2017-04-10. Springer. 9783319566146. 104. en.
- Downlin. Nathan. Gilad-Bachrach. Ran. Laine. Kim. Lauter. Kirstin. Naehrig. Michael. Wernsing. John. 2016-05-25. CryptoNets: Applying Neural Networks to Encrypted Data with High Throughput and Accuracy. Proceedings of the 33rd International Conference on Machine Learning. https://web.archive.org/web/20180826211245/http://proceedings.mlr.press/v48/gilad-bachrach16.pdf. 2018-08-26.
- Web site: The Microsoft Simple Encrypted Arithmetic Library goes open source. 2018-12-03. Microsoft Research. en-US. live. https://web.archive.org/web/20191112165543/https://www.microsoft.com/en-us/research/blog/the-microsoft-simple-encrypted-arithmetic-library-goes-open-source/. 2019-11-12. 2019-11-20.
- Fan. Junfeng. Vercauteren. Frederik. 2012. Somewhat Practical Fully Homomorphic Encryption.
- Book: Cheon. Jung Hee. Kim. Andrey. Kim. Miran. Song. Yongsoo. Advances in Cryptology – ASIACRYPT 2017 . Homomorphic Encryption for Arithmetic of Approximate Numbers . 2017. Takagi. Tsuyoshi. Peyrin. Thomas. https://link.springer.com/chapter/10.1007/978-3-319-70694-8_15. Lecture Notes in Computer Science. 10624. en. Cham. Springer International Publishing. 409–437. 10.1007/978-3-319-70694-8_15. 978-3-319-70694-8. 3164123 .
