PySEAL: A Python wrapper implementation of the SEAL homomorphic encryption library
Alexander J. Titus, Shashwat Kishore, Todd Stavish , Stephanie M.Rogers , and Karl Ni
Abstract
Motivation: The ability to perform operations on encrypted data has a growing number of applications in bioinformatics, with implications for data privacy in health care and biocurity. The SEAL library is a popular implementation of fully homomorphic encryption developed in C++ by Microsoft Rearch.
写给老婆的一封信
Despite the advantages of C++, Python is a flflexible and dominant programming language that enables rapid prototyping of bioinformatics pipelines.
Results: In an effort to make homomorphic encryption accessible to a broader range of bioinformatics scientists and applications, we prent a Python binding implementation of th
e popular homomorphic encryption library, SEAL, using 凉拌水萝卜的做法pybind11. The software contains a Docker image to facilitate easy installation and execution of the SEAL build process.
Availability: All code is publicly available at github：Lab41/PySEAL
Introduction
The volume of available genomic data is growing rapidly as the cost of quencing continues to decline. Biomedical rearch advances can greatly benefifit from this wealth of data, but the nsitive nature of such information demands that privacy and curity concerns be given full attention. For example, recent work has demonstrated an ability to re-identify individuals using their genomic data (Gymrek et al.王士真, 2013; Lippert et al., 2017). Furthermore, many analys, such as the Genome Wide Association Studies (GWAS), require large sample sizes that are often diffificult to collect at a single clinical site, which highlights the challenge of maintaining the curity of nsitive data being collected and shared among multiple platforms at multiple geographic locations.
In 2010, the Center for Integrating Data for Analysis, Anonymization and SHaring (iDASH) was founded at the University of California, San Diego to address privacy challenges to rearch progress. At the annual iDASH Privacy & Security Workshop, teams from around the world develop novel methods for private and cure computation on health data, and extend the applications of homomorphic encryption into the areas of biological data storage, transport, and computation (Kim and Lauter,
2015).
Homomorphic Encryption offers a promising approach to addressing privacy and curity concerns in the handling of nsitive genomic data, becau it allows data to remain encrypted even during u, and it offers a curity method to enable comparisons across disparate hardware and geographic distances. Under a fully homomorphic encryption cryptoscheme (Gentry, 2009), any arbitrary t of operations that take place on the ciphertext (encrypted space) are mirrored in unencrypted space (Figure 1).
Here, we prent a Python binding of the Simple Encrypted Arithmetic Library (SEAL简单的读书手抄报) ho
momorphic encryption library (Chen and Laine, 2017).
SEAL is maintained by the Cryptography Rearch Group at Microsoft Rearch and implements a fully homomorphic encryption scheme in C++.
The software has strong applications in bioinformatics (Dowlin et al.吆喝, 2017). While C++ is appropriate for software engineering and developed deployments on established architectures, biologists and bioinformatics scientists, especially tho with less-than-formal computer science training, often prefer rapid prototyping scripting languages like Python.
Fig. 1. In a fully homomorphic encryption scheme, operations on ciphertext space are mirrored in the plaintext space after decryption
Therefore, we prent PySEAL, a Python wrapper implementation of the SEAL library using 错误651pybind11. The PySEAL package implements SEALv2.3 and the package is under active maintenance.
Methods
PySEAL is a Python wrapper for the Simple Encrypted Arithmetic Library (SEAL) homomorphic encryption library. The 秋游的作文PySEAL software is implemented using pybind11 and is relead as a Docker container, a C++ SEAL example build, and a Python wrapper build. The software also includes bash scripts to build the Docker container and execute a ries of examples using PySEAL functions.
The PySEAL code is relead on the Lab41 GitHub page and is under active maintenance. The README includes an overview description of the steps to instantiate and u the fully homomorphic encryption scheme.
Brieflfly, the fifirst step to using the library for basic encryption tasks is to instantiate a ne
w EncryptionParameters object and t its modulus attributes. The parameters are stored in a SEALContext object which checks the validity of the parameters. Once validated, the parameters are ud to create encryption keys and Encryptor and Decryptor objects.
